Around Christmas each year, I produce printed PDF monthly calendars for my wife, incorporating our family’s anniversaries and other important dates.

Originally this was done in a spreadsheet, but that was always a headache, especially when multiple events shared the same date. One year, I gave up and replaced the spreadsheet with a Perl script using PDF::API2 and this has worked well for a number of years. But, being CLI-based, this remained on my annual to-do list.

The idea to create a web front-end through Vibe Coding came from an episode of the Linux Matters Podcast, in which the hosts discussed maintaining a living specification document and using AI to generate an application directly from that specification. The process they described was disciplined: refine the specification, regenerate the system, repeat - until the specification itself becomes the definitive source of truth.

The process

I started by asking ChatGPT whether my choice of Perl + Dancer would be suitable, or whether a different language and web framework would be likely to produce better results. Sadly, ChatGPT felt that Codex would be much more effective using a more popular combination and, of the choices suggested, I opted for Ruby with Sinatra. I had previously used Ruby on Rails, but had found it to be too difficult to maintain. Sinatra is a more lightweight web framework that keeps structure simple and avoids unnecessary abstraction. For the same reason, I chose Sequel for the database layer instead of ActiveRecord.

The project was broken into sections - creating the base website, adding authentication, etc. At each major step, we worked on a structured specification, which I fed into Codex. I tested the results and gave feedback to Codex. Codex automatically adds unit tests, and uses them to give itself feedback - it is interesting to watch it attempting to fix its own mistakes, unprompted.

When problems occurred, I would often discuss these with ChatGPT before asking Codex to make changes. For minor steps, I worked directly with Codex, with more of a simple trial and error approach.

The underlying PDF generation remains largely the original Perl script. The surrounding system - authentication, subscription handling, templating, validation and deployment - was almost entirely developed conversationally.

The frustrations

There have been some frustrations.

At one point, what I thought was a simple change, resulted in a convoluted and entirely unjustified code change. This happened simply because I was not close enough to the JSON data being accessed, and Codex was trying to achieve something which normally I would have known was not possible. I reverted the change, but not before it had used up my remaining Codex tokens.

Another issue was that, whilst complex changes were often completed quickly and efficiently, sometimes unbelievably trivial changes had us looping around in circles until I was forced to take control. Minor CSS positioning tweaks were a constant frustration.

We experienced a lot of ‘mission creep’ - it was just too tempting to extend the project, given that it wasn’t me doing the hard work. We should support Internet Calendars! We should have a weekly calendar format! We should have a daily planner format! Quick and easy to type the commands into Codex, but the ensuing trial and error, added much to the length of the project.

Conclusions

After many iterations and ‘back to the drawing board’ moments, we finally reached the point where it received spouse-approval, especially when she realised that next year she will be able to simply login and print.

The result is MakeMyYear, a small subscription service that allows users to generate personalised downloadable calendars for £4 per year, which hopefully will offset some of the Codex costs, as well as giving me invaluable experience with working with Stripe.

I do feel it will be very difficult to go back to traditional coding. The line by line coding which I had always found so engaging, now seems like far too much like hard work!

Have I just spent months of free time and hard-earned money, simply making it possible for my wife to print her own calendars? Why yes, yes I have.

Having upgraded a DokuWiki server from 16.04 to 18.04, I found that SVG images were no longer displaying in the browser. As I was unable to find any applicable answers on-line, I thought I should break my radio silence by detailing my solution.

Inspecting the file using browser tools, Network and refreshing the page showed that the file was being downloaded as octet-stream. Sure enough using curl showed the same.

curl -Ik https://example.com/file.svg

All the advice on-line is to ensure that /etc/nginx/mime-types includes the line:

image/svg+xml   svg svgz;

But that was already in place.

I decided to try uploading the SVG file again, in case the Inkscape format was causing breakage. Yes, a long-shot indeed.

The upload was rejected by DokuWiki, as SVG was not in the list of allowed file extensions; so I added the following line to /var/www/dokuwiki/conf/mime.local.conf:

svg   image/svg_xml

Whereon the images started working again. Presumably Dokuwiki was seeing the mime-type as image/svg instead of image/svg+xml and this mismatch was preventing nginx serving up the correct content-type.

Hopefully this will help others, do let me know if it has helped you.

Following my purchase of a SONOS system, I wanted to be able to play my music collection through it. Unfortunately many of the album covers were missing. After considerable googling I eventually determined that the missing covers were missing because I had embedded the album cover with a higher resolution than SONOS can cope with.

To fix this problem I needed to extract the embedded images, change the resolution, remove the current images, and re-embed the new lower resolution image. I needed to do this for each album that was missing its cover.

Adding your music library to SONOS

Firstly, a word about adding my music library to SONOS. This involved transferring my MP3 music collection onto an external hard drive and plugging it into my router and sharing it as a Samba share. The final step is to add Music Library as a service within the SONOS app, and specify the location of the Samba share created above.

If your router doesn’t offer a Samba option, then you might consider a Raspberry Pi with a simple samba share, or using NextCloud or a dedicated plug-in device like the Tonido Plug, or pretty much any NAS device.

Important Notice

Unfortunately, whilst I made some key notes at the time, a couple of years has passed before I finally found the time to type up this post. As a result there may be some errors, or you might need need to interpret / adjust some of these commands.

I do not wish to be responsible for you losing your music collection, please do take a backup of your music before you start.

Please do let me know if you spot any mistakes, by commenting below.

Preparation

These instructions only require eyeD3 and either grapicsmagick or imagemagick. These instructions were written for Ubuntu GNU Linux, you will need to adapt these instructions if you use a different distribution.

$ sudo apt install eyed3 graphicsmagick

If you already have Imagemagick installed, then you may use that instead, just remove the “gm “ from the relevant commands below.

Step 1: Extract image from music

Please change directory into the first album that requires fixing, and stay in that directory until you have finished all steps.

As the album cover is embedded in each track in the album, you only need to extract this once per album:

$ eyeD3 --write-images=./ "01 - Track 1 name.mp3"

This will extract the embedded image from the file as:

FRONT_COVER.JPG

Step 2: Identify current resolution

The next step is to identify the size of the image, to do this I use graphicsmagick, but imagemagick will work the same (just omit the gm):

$ gm identify FRONT_COVER.JPG

If the size of the image is already smaller than 300x300 pixels, then it should already be working, otherwise we need to reduce its size.

Step 3: Convert

To reduce the size of the image, do a simple convert:

$ gm convert FRONT_COVER.JPG -geometry 300x300 folder.jpg

If you would like to check that this has worked just run:

$ gm identify folder.jpg

Step 4: Remove images from MP3 files

Before we can add the newly sized image, we need to remove all the embedded images in the tracks.

Please note that this command alters your music, please do take a backup to ensure that you can restore the tracks if necessary.

$ eyeD3 --remove-images *.mp3

Step 5: Add new image to MP3 files

And finally we can now embed the smaller image:

$ eyeD3 --add-image=./folder.jpg:FRONT_COVER *.mp3

Step 6: Update music index

I cannot remember whether or not I needed to update the music index, but if your album covers haven’t appeared already - try this:

• Open the SONOS app
• Settings
• Manage Music Library
• Update Music Index Now

Conclusion

I was in two minds as to whether to post this, given the time elapsed, but hopefully it will prove useful to someone. Please do comment below if you spot any errors, or found it useful.

Thank you.

I logged into my Windows 10 Professional (1703) desktop from home yesterday, using Remmina on Ubuntu 16.04. I wasn’t surprised when my desktop wallpaper was black, I know it does this to save bandwidth, but when I returned to the office this morning my desktop was still black and, as it is set by the administrator via GPO, could not be changed.

Searching the Internet was not helpful on this occasion; so I have made this quick posts to help others.

It turns out that this is not some weird absence of wallpaper, but rather is a plain black wallpaper image, which has managed to get itself cached. The solution is consequentially simple - find and terminate said cached image.

1. Open File Explorer and navigate to %APPDATA% (you can type that into the top address/location field).
2. In the search box at the top right enter the text “Cache” (see image below).
3. Delete the cached version of the black wallpaper once it is found.
4. Sign out and then sign back in.

On my system the specific file location was:

%APPDATA%\Microsoft\Windows\Themes\CachedFiles\CachedImage_1920_1080_POS4.jpg

Please do comment below if this was helpful, or if you needed to alter these instructions at all.

Today I released the first semi-stable version of sluggle, a simple IRC searchbot. The name is a contraction of Google and SLUG (an informal and inaccurate acronym for the Surrey Linux User Group).

Sluggle enables IRC users to search the Internet using Bing, to search Wolfram Alpha, to look up and shorten web addresses and to check the safety of sites.

For more details please visit sluggle.

A while ago I was explaining the difference between QEMU, KVM and LibVirt, and I ended up by emailing this nonsense. I don’t claim it’s accurate, it certainly isn’t. It’s probably not even funny. Enjoy :).

In the beginning there was QEMU, but it was slow and the people grieved.

Then KVM was forked from QEMU with a kernel module to use the CPU’s virtualisation features to work much faster and there was much rejoicing. Linus also rejoiced and welcomed KVM’s kernel module into the mainline kernel.

But the people did not rejoice, as they were mostly using Sun’s VirtualBox (also forked from QEMU).

QEMU awoke from its slumber and joined with KVM and their union caused almost no rejoicing, in fact I am not convinced anyone really noticed.

But the System Administrators were still dissatisfied and complained that there should be standardisation of commands across different hypervisors. And thus LibVirt was born and the System Administrators rejoiced.

Yet still the people used VirtualBox. But lo! The evil Oracle slew the Sun and VirtualBox moved into darkness, and there was much gnashing of teeth and wearing of sackcloth; although this was generally considered a step forwards from the t-shirts that they usually wore.

But still the people could not use QEMU-KVM, without issuing complex incantations, and so Virt-Manager was born and finally the people rejoiced, with much clicking of mice.

The End.

I’ve been aware of Sonos as a premium wireless speaker solution for a long time, but the price always seemed excessive for what, on the face of it, offers little more than a simple Bluetooth speaker. But after Subsonic needed its database rebuilding for the third time and I was unable to play music for a dinner party, enough was enough. I was willing at last to pay the premium for something that was purported to work.

Background

My music collection is mostly comprised of purchased Audio CDs that I have ripped under Linux. Currently I have a Music folder on our MythTV system, and have installed Subsonic to share our music to our many tablets and phones, using the excellent Subsonic Android App. If I want to play from Subsonic to my music system then I have a Logitech Bluetooth Audio Receiver Adapter that receives the audio and plays it through my old-school Sony amplifier.

The main issues with this set-up is that the music only plays in the living room and not elsewhere in the house. We have bought an additional Creative D80 Bluetooth Wireless Speaker, but of course each can only play independent streams.

I also find Bluetooth a frustrating technology where you don’t have a simple 1:1 paradigm. In our case we have probably a dozen tablets and phones, each determined to pair with the Bluetooth receivers and then prevent other devices from connecting.

Choosing Sonos Speakers

The Sonos range comprises of the small Play:1 at £155, a medium-sized Play:3 at £229 and a larger Play:5 at £413. On the Goldilocks principle of the middle one being “just right”, I opted for two of the Play:3 at £229 each - one for the living room and one for the kitchen. The plan was to move those elsewhere at a later stage and hopefully upgrade the living room system to a pair of Play:5 speakers.

The important thing to understand is that neither the Play:1 nor the Play:3 speakers have a Line-In. This means that you can only play from on-line content. If you currently subscribe to one of the supported Sonos Services, then that is fine, but if you’re wanting to play content from a CD or other input source - then you can’t. The Play:5 does have a Line-In, as does the Sonos Connect at £264.

A word about the Sonos Connect. A simple way to imagine it is that it is basically a Play:5, but without the speaker. In other words it has the same Sonos interface with Line-In but no speaker. If you have an existing music system then this is potentially ideal and with hindsight I wish that what I had done was to purchase one Play:3 for the kitchen and one Sonos Connect for the living room. The opposing view is that a pair of Play:5 speakers complete obsoletes an existing music system - so why not do away with the legacy equipment.

Amazon Prime Music

One great disappointment was that, whilst Sonos supports Amazon Music, it does not support Amazon Prime Music. One of the main reasons we had bought Sonos was to play Amazon Prime Music, so this was a major problem. At the time of writing it is available in the US as a Beta service and has been for a few months. One can only hope that it will trickle across to the UK in due course.

Google Play Music

Hoping that the Amazon Prime issue would be resolved, we signed up to a 30 day free trial with Google Play Music. This worked extremely well, except for the recommended playlists which do not appear as a Sonos Queue. The main issue we found was that our children would choose a song and click “Play Next” and this would interrupt the playlist - very irritating if you were enjoying a particular song. We assumed that this was a feature of Sonos, but Spotify does not work like that (see below).

Spotify

We then subscribed to the 30 days free trial with Spotify. You only need the individual member subscription to work with Sonos, but the ongoing cost is the same as Google Play. The only advantage of Spotify is that the recommended playlists appear as a proper Sonos queue, enabling you to save it as a Sonos Playlist, or add a song into the queue.

Subsonic

One delight was that we were able to play our local music via Subsonic. This is a Beta service and I did have a small problem getting it working. Unfortunately I cannot remember the nature of the problem, other than an Internet search solved it.

Conclusions

Obviously we were disappointed at the lack of Amazon Prime Music. I was also a little disappointed at the abrupt handling of music changes - if you click “Play Now” the music stops instantly and the next track starts. I do feel that with a premium set-up like this that music transitions should be handled more smoothly.

We also have had issues with our children messing about with Sonos - as the interface is open to all. We have sufficient control of our children that this isn’t a significant problem, but knowing some families this could be a serious issue. I do feel there should be some security, to enable clients to be de-authorised, or limited only to a subset of features.

Will I continue to invest in Sonos? Undoubtedly yes, but I think the next purchase will be a Sonos Connect followed by a better set of audio speakers.

I first wrote about the difficulties of submitting a UK self-assessment tax return back in January 2009, with a follow up in January of 2010. In the subsequent years I have submitted my tax return under Windows with nary a thought about attempting it under Linux.

This year was different, I went to download TaxCalc as usual, only this time there was a Linux option. I barely noticed at first, downloading the Windows version before I thought “Hang on, did that say Linux?”. It did indeed.

Despite the fact that I do have a Windows PC now, as well as a Linux PC and a Linux laptop, it probably makes more sense to perform my tax return on Windows now, but having posted on this subject before I really ought to give it a go!

I won’t detail all the installation instructions, which are clearly described on the TaxCalc website.

Installation

Unfortunately the download is for a compiled .run file and not a distribution package, so you will have to decide whether you are happy to run the file and potentially dump unknown rubbish all over your system which might then be difficult to remove.

I decided to give it a go, downloading the .run file and marking it as executable. The instructions provided did not suggest running with sudo or as root, so I decided to follow the letter of the instructions and simply executed the .run file.

This entered a typical Wizard style installer, which resulted in the application being installed within my home directory, including a 127MB lib directory containing a large number of Linux libraries. This is very much the “Windows” way of doing things, bundling everything you need within the installation, rather than simply specifying dependencies. But the fact is that it works and gives TaxCalc control over the installation, in a repeatable and supportable way.

Import from previous year

The first step in using TaxCalc is always to import the previous year’s files. This process completed for all files without error.

Step by Step Process

Probably the main reason for using TaxCalc is the simple step process for completing your tax return, and I am pleased to report that this aspect was identical to Windows.

On-line Submission

The one aspect that I was concerned about was whether the on-line submission to HMRC would work okay, again I am pleased to say that it worked flawlessly, both for the test submission and the live submission.

Conclusions

I must confess that I assumed that this day would only arrive when TaxCalc moved to a web-based service, something that surely is inevitable. I am delighted to at last be able to perform my tax at home, rather than on a Windows PC at work. Good news for those wanting to eradicate Windows from their lives.

Thank you TaxCalc.

At home we have an HP OfficeJet J6410 printer, which has worked brilliantly for years. Unlike every previous printer we’ve owned, this printer sits happily on our network and seems perfectly content to be used only very occasionally, whereon it willingly responds with decent quality prints and scans.

I have set up the printer with a static IP address and so the scanning is managed via the web interface http://192.168.0.100/webScan, whereon the scanned image opens a new tab in the browser, enabling you to save to wherever.

This printing and scanning Nirvana hit a road bump today, when it just instantly went to “Scan done” and failed to actually do anything. Rebooting the laptop, the printer and clearing temporary internet files all failed to resolve the problem, as did changing from Chrome to Firefox.

In desperation I tried running Ubuntu Simple Scan, but predictably enough it did not find the printer. A quick Internet search took me to this excellent Ubunutu community page.

I opened a terminal and typed:

$ hp-makeuri 192.168.0.100

CUPS URI: hp:/net/Officejet_J6400_series?ip=192.168.0.100
SANE URI: hpaio:/net/Officejet_J6400_series?ip=192.168.0.100
HP Fax URI: hpfax:/net/Officejet_J6400_series?ip=192.168.0.100

Done.

I then typed (taking the SANE URI from above):

$ simple-scan hpaio:/net/Officejet_J6400_series?ip=192.168.0.100

And it worked perfectly! Flushed with success I read on to see how to make this a permanent fixture, and it turned out that all I needed to do was configure the CUPS URI for the printer itself (it was previously set to something like socket://192.168.0.100):

$ vim /etc/cups/printers.com

DeviceURI hp:/net/Officejet_J6400_series?ip=192.168.0.100

And now simply running Ubuntu Simple Scan works perfectly.

Thank you Ubuntu community!

I wrote recently regarding using Microsoft Skype for Business / Lync 2013 in Pidgin, this has worked really well for our Lubuntu 14.04 clients, running Pidgin. Having some time off work I decided to try and get this working also on my Ubuntu 14.04 laptop.

The problem is that Ubuntu uses Empathy by default, rather than Pidgin, but I found that the steps were virtually identical (cut and pasted from my earlier post):

Download and Dependencies

1. Install build tools if you don’t already have them:

   sudo apt-get install build-essential

2. Install checkinstall if you don’t already have it:

   sudo apt-get install checkinstall

3. Download source files.

4. Extract source:

   tar -xvvzf pidgin-sipe-1.20.1.tar.gz

5. Change into source directory:

   cd pidgin-sipe-1.20.1

6. Read carefully the README file in the source directory.

7. Install dependencies listed in the README:

   # apt-get install libpurple-dev libtool intltool pkg-config libglib2.0-dev libxml2-dev libnss3-dev libssl-dev libkrb5-dev libnice-dev libgstreamer0.10-dev

These dependencies may change over time, and your particular requirements may be different from mine, so please read the README and that information should take precedence.

Compile and Install

Lastly, as an ordinary user, you should now be able to compile. If it fails at any stage, simply read the error and install the missed dependency.

$ ./configure --prefix=/usr
$ make
$ sudo checkinstall

I found checkinstall was pre-populated with sensible settings, and I was able to continue without making any changes. Once complete a Debian package will have been created in the current directory, but it will have already been installed for you.

Add Account in Empathy

Still in a terminal type:

$ empathy-accounts

Then simply click the + button to add an account, select the Account Type “sipe”, enter your email address in the Account field and click Log-in. Then highlight the newly created account and Edit Connection Preferences, opening up the Advanced section. I completed the following:

• Login: enter your Skype for Business email address
• Password
• Server: leave blank
• Transport: auto
• User agent: UCCAPI/15.0.4420.1017 OC/15.0.4420.1017
• Authentication: auto

And that really was all that there was to it.

I was looking for a Wiki for managing the sharing of a boat, including booking usage, task lists and maintenance, as well as acting as a repository for any-and-all information available.

What it looked like was incidental, but it needed to be really simple for users to edit pages quickly. I wanted to squeeze it onto a work server, so it also needed to be really lightweight.

The Choices

I have experience of using both Wordpress and Drupal, but felt they were too complex and cumbersome for this use-case. I have also used MediaWiki in the past but, whilst easy to edit, I found it fairly cumbersome to use - it is after all designed for managing vast sites.

In the end I opted for DokuWiki, after stumbling across a site that recommended it as a really simple alternative to MediaWiki.

Installing DokuWiki

Dokuwiki is really easy to install, with delightfully simple instructions for Apache, but following my lightweight wishes I decided to try Nginx for the first time. Unfortunately I could not immediately find Nginx Instructions from DokuWiki, so instead followed these Nginx Instructions from Rose Hosting.

On first attempt the landing page would not load, but the nginx logs were very clear and helpful in pointing me to increasing a value of server_names_hash_bucket_size in /etc/nginx/nginx.conf. This I found was commented out (with a default of 32), so I uncommented it and increased it to 64 and restarted nginx. At last I could reach the install page and from that point it all worked perfectly.

Later I decided to implement Nginx Pretty URLs, which again worked perfectly.

First Impressions

The main thing I love about DokuWiki is the simplicity.

Creating users is trivial, creating and editing pages is a doddle, there is a reasonably sane way of managing media files, and the Wiki syntax is bearable (not Markdown sadly).

Despite this simplicity, each time I have come across something I’d like to tweak, then there has been a way to do it without resorting to nasty hacks. Where Plugins are required, these can be installed in seconds via the web interface.

Most of what I was trying to do with this boat site was simply a matter of creating pages, but I did come across a few exceptions:

Checklists

One of the key uses for the site is to host a number of checklists, but I wanted tickboxes against each item. Not essential clearly, but it turned out to be trivial, simply deciding how to type a tickbox, e.g. square brackets for incomplete [] and [x] for complete, and then adding DokuWiki Entities to convert these into the appropriate character.

Bookings

The main problem that I faced was how to enable boat users to book time on the boat and in the end I settled on using simple text tables, but to make them simpler to edit I added Andreas Gohr’s excellent Edittable Plugin. This plug-in is so good that I would always install it in any DokuWiki that will have tables.

Whilst this will certainly do for now, I am on the look out for some sort of site to manage bookings properly - if you know of anything suitable please do comment below, otherwise at some point I will probably write one in Perl Dancer!

Conclusions

Under the Use Cases on the DokuWiki website there is the suggestion of using it for a Private Notebook - having used DokuWiki I can totally see how that would make sense - it is that simple to create and edit.

DokuWiki has definitely filled a gap in my web toolbox, alongside Perl Dancer and Jekyll.

Further to my recent post Intermittent USB3 Drive Mount, sadly I have to report that, after ten days of perfect operation, this morning once again the drive mount failed.

# dmesg | tail

[ 7659.848351] usb 2-3: Product: USB3.0 Hub             
[ 7659.848352] usb 2-3: Manufacturer: VIA Labs, Inc.         
[ 7659.852937] usb 2-3: Disable of device-initiated U1 failed.
[ 7659.856435] usb 2-3: Disable of device-initiated U2 failed.
[ 7659.856473] hub 2-3:1.0: USB hub found
[ 7664.860293] hub 2-3:1.0: config failed, can't read hub descriptor (err -22)
[ 7664.863787] usb 2-3: Set SEL for device-initiated U1 failed.
[ 7664.867279] usb 2-3: Set SEL for device-initiated U2 failed.
[ 8137.312105] usb 2-3: reset SuperSpeed USB device number 3 using xhci_hcd
[ 8137.559139] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with disabled ep ffff8800bc7e8a20

Power Control

Following the tack from last time, I once again checked the USB power control status:

# grep . /sys/bus/usb/devices/*/power/control

As before these were mostly set to auto, so I switched on again using the same command as before:

# for F in /sys/bus/usb/devices/*/power/control; do echo on >"${F}"; done
# grep . /sys/bus/usb/devices/*/power/control

Clearly I need a more permanent fix to that command.

Auto-Suspend

I also discovered two more settings:

# grep . /sys/bus/usb/devices/*/power/autosuspend

Apparently to switch off auto-suspend you should set the delay seconds to -1:

# for F in /sys/bus/usb/devices/*/power/autosuspend; do echo -1 >"${F}"; done
# grep . /sys/bus/usb/devices/*/power/autosuspend

To set this permanently I created a file /etc/modprobe.d/autosuspend.conf:

options usbcore autosuspend=-1

I came across Ubuntu Bug 1371233, in which there was a suggestion to try:

# echo -1 >/sys/module/usbcore/parameters/autosuspend

And on the same lines the suggestion to edit /etc/default/grub file and append to the GRUB_CMDLINE_LINUX_DEFAULT line:

usbcore.autosuspend=-1

Followed of course by the obligatory:

# update-grub

Power Level

You can also check your USB power levels:

# grep . /sys/bus/usb/devices/*/power/level

Mine were already all switched on, but if not:

# for F in /sys/bus/usb/devices/*/power/level; do echo on >"${F}"; done
# grep . /sys/bus/usb/devices/*/power/level

Back to Basics

Having run through all the above, I still could not see my drive. Unplugging and plugging it back in seems to have no reaction in the logs, suggesting that my drive is not being seen at all.

Re-plugging the USB hub seems to respond okay:

Nov 27 11:51:47 asusi5 kernel: [11372.680605] usb 1-3: USB disconnect, device number 2
Nov 27 11:51:47 asusi5 kernel: [11372.775937] usb 2-3: USB disconnect, device number 3
Nov 27 11:51:54 asusi5 kernel: [11379.603869] usb 1-3: new high-speed USB device number 7 using xhci_hcd
Nov 27 11:51:54 asusi5 kernel: [11379.733470] usb 1-3: New USB device found, idVendor=2109, idProduct=2812
Nov 27 11:51:54 asusi5 kernel: [11379.733478] usb 1-3: New USB device strings: Mfr=1, Product=2, SerialNumber=0
Nov 27 11:51:54 asusi5 kernel: [11379.733482] usb 1-3: Product: USB2.0 Hub
Nov 27 11:51:54 asusi5 kernel: [11379.733486] usb 1-3: Manufacturer: VIA Labs, Inc.
Nov 27 11:51:54 asusi5 kernel: [11379.734284] hub 1-3:1.0: USB hub found
Nov 27 11:51:54 asusi5 kernel: [11379.734556] hub 1-3:1.0: 4 ports detected

But still plugging in the USB3 drive into that hub still has no effect.

Next I tried re-plugging the USB hub into a different USB3 port. This time when I plugged back in the USB drive - it was recognised and worked perfectly.

Conclusions

It could be that one of the above fixes only affects new devices (the auto-suspend probably), and that by choosing a different port it was recognised as a new device and therefore worked.

Equally there could be a hardware problem with the original USB port, perhaps the connector inside the case is a bit dodgy.

For now I will monitor the situation for another week or two and report back.

In my recent post on the Bac Posture Stand I mentioned that I had purchased the Varidesk Pro Plus 48. I promised to review this “in due course” and that time has now come!

The Varidesk Pro Plus 48 is “desk riser” that enables you to choose your working height. You can start the day standing up and then, after your executive lunch, you can spend your afternoon sitting down.

N.B. This post includes Amazon Associates links, I have never actually received anything from them, but I live in hope!!

Pricing

At £437 (including delivery) this is a very expensive piece of equipment indeed. The alternative is to buy an actual standing desk and in many cases this will be a better option. The beauty of the Varidesk is that you are able to retain your existing furniture.

You do also need the Varidesk Standing Mat, which is another £67.50 (including delivery).

There are cheaper solutions, such as the LIFT Standing Desk mentioned on Bad Voltage, but at the time I was unable to locate this in the UK. You can even build your own.

Amazon

In the end I purchased the Varidesk Pro Plus 48 from Amazon*. Not only was it almost the only option that I could find, it was a professional looking unit and the reviews were also very positive.

*Please note that the photo on the Amazon item is actually incorrect and the Pro Plus 48 actually looks like the photo above, visit the Varidesk website for details. I did contact Varidesk about this anomaly and they explained that they had informed Amazon on multiple occasions and were unable to get the photo corrected.

To enhance my new standing desk I also purchased a number of other items:

• Varidesk Pro Plus 48
• Varidesk Standing Mat - essential for standing all day
• Anker USB3 Hub - to provide USB ports on the standing desk
• Amazon 3m USB Extension Cable - to connect up the Anker hubs
• Amazon 3m HDMI Cable - required to reach from the PCs under the desk up to the standing desk
• Fellowes Cable Zip - to tidy up all the cables

Installation

Typically the Varidesk arrived when I was on a day off. It turned up on a pallet, resulting in much consternation at work. In fact, once the packaging was removed, the item was the desired size and all was well.

There is no assembly required, simply unpackage and lift onto your desk. Be warned though, the Varidesk does weigh a lot and will require two fairly strong people to manoeuvre it into place.

I was a little concerned as to how this Pro Plus 48 would fit on my corner desk. I did originally order the Varidesk Cube Corner 48, but I received an email from Varidesk explaining that this unit had been discontinued and cancelling and refunding my order.

In the event the Pro Plus 48 fitted my corner desk beautifully.

First Impressions

The Varidesk Pro Plus 48 appears to be extremely well constructed and solidly made. It has what I assume are bolt holes for those scared of the desk falling over, but in practice the weight of the metal base is sufficient.

The desk can be raised and lowered without physical effort, although sometimes cables will get under the desk requiring their removal before it will lower fully. I cannot blame Varidesk for that, but you need to give some thought about cable routing to minimise the likelihood of this happening.

The First Month

After a month of using my Varidesk I am very happy with it. I would love to be able to report a stunning transformation of my painful back, but sadly my back is still very sore. That said, I certainly feel healthier from standing up and being more active.

Standing up all day is not without its issues, my feet and legs get tired, so I tend to sit down at lunchtime and late afternoon. But I have also found that sitting down for too long has started feeling very uncomfortable. Perhaps it always did? But now I can do something about it, raising up my desk and enjoying standing up again, if only for a while.

In short, for me at least, being able to choose to sit or stand feels very natural and it would feel very strange now to be forced to spend the entire day sitting down.

I believe I now have the closest thing possible to the perfect desk.

Where we require a lightweight mail client, we tend to use Sylpheed (from which Claws Mail was later forked).

It seems unlikely that you would be able to add an ActiveDirectory Address Book into such a lightweight email client, and indeed the manual states:

### FIXME: write this part.

But in fact it was trivially easy:

Caveat

Whilst these instructions worked for us, do be aware that we are using Samba4 ActiveDirectory. In theory this is a drop-in replacement for Windows ActiveDirectory and these instructions should work unchanged.

Add LDAP Addressbook

Firstly run Sylpheed and go to Tools and Addressbook. Within the Sylpheed Addressbook go to File New LDAP Server You should now see a screen like this:

Having entered the Name, Hostname and Port you are able to “Check Server”, to ensure connectivity. Next either enter your Search Base, or click on the &ellipsis; button to select from the detected Search Bases.

*You should probably choose 636 when connecting via a public network, and you may need to open ports on your router.

Now select the Extended tab and you should see the following screen:

Now click on OK to finish.

Test

You should now have a Search field available, enter a colleague’s first name and Search and you should be faced with their email addresses.

Conclusion

As far as I can tell the addressbook lookup is not automatic and you have to click on the addressbook icon in the Compose Window and search for the person, in order to add them to the To: field. A bit clunky perhaps, but arguably not so very different from the need in Outlook to press Check Names to look up new addresses. Needless to say - once the address is in the recent address list, it is auto-completed in the future.

After much research I decided to purchase ESET Antivirus for our Windows clients. But rather than install ESET Antivirus on each client in turn, I decided to install ESET Remote Administrator on an Ubuntu VPS. ESET Remote Administrator is a server program for controlling ESET Antivirus (and other ESET programs) on clients, you can use it to deploy the ESET Agent, which in turn can then install and monitor the ESET Antivirus software.

Our Windows PCs are controlled by an ActiveDirectory server (actually a Samba4 ActiveDirectory server, although that should not make any difference to these instructions).

I found the ESET instructions for so doing decidedly sketchy, but I eventually managed to install it and took some notes as I went. I cannot promise these instructions are complete, but used in conjunction with the ESET instructions they may be of help.

Update Hosts

Edit /etc/hosts:

127.0.0.1       localhost.example.lan      localhost.localdomain        localhost
192.168.0.109   eset.example.lan           eset

Test by typing the following two commands and checking the output matches:

# hostname
eset

# hostname -f
eset.example.lan

Dependencies

# apt-get install mysql-server unixodbc libmyodbc cifs-utils libjcifs-java winbind libqtwebkit4 xvfb
# dpkg-reconfigure libmyodbc 

Install MySQL Server

Edit /etc/mysql/my.cnf:

max_allowed_packet=33M

Restart MySQL

# service mysql-server restart

Configure MySQL

# mysql -u root -p

Cannot remember whether I created database and/or username and password, suggest you try without and then with.

Install ESET Remote Administration Server

Replace ??? with actual:

# sh Server-Linux-x86_64.sh --license-key=??? \
--db-type="MySQL Server" --db-driver="MySQL" --db-hostname=127.0.0.1 --db-port=3306 \
--server-root-password="???" --db-name="era_db" --db-admin-username="root" --db-admin-password="???" \
--db-user-username="era" --db-user-password="???" \
--cert-hostname="eset" --cert-auth-common-name="eset.example.lan" --cert-organizational-unit="eset.example.lan" \
--cert-organization="example ltd" --cert-locality="UK" \
--ad-server="ads.example.lan" --ad-user-name="era" --ad-user-password="???"

In case of error, read the following carefully:

/var/log/eset/RemoteAdministrator/EraServerInstaller.log

Install Tomcat7

# apt-get install tomcat7 tomcat7-admin

Wait 5 minutes for Tomcat to start, then visit http://localhost:8080 to check it has worked.

Configure SSL

See SSL/TLS Configuration HOW-TO.

Step 1 - Generate Keystore

$JAVA_HOME/bin/keytool -genkey -alias tomcat -keyalg RSA -keystore /usr/share/tomcat7/.keystore

Use password ‘changeit’.

At the time of writing $JAVA_HOME is /usr/lib/jvm/java-7-openjdk-amd64/jre

Step 2 - Configure Tomcat

You should have two sections like this:

<Connector port="8080" protocol="HTTP/1.1"
           connectionTimeout="20000"
           URIEncoding="UTF-8"
           redirectPort="8443" />

<Connector port="8443" protocol="HTTP/1.1" SSLEnabled="true"
           maxThreads="150" scheme="https" secure="true"
           keystoreFile="/usr/share/tomcat7/.keystore" keystorePass="changeit"
           clientAuth="false" sslProtocol="TLS" />

These will already exist, but need uncommenting and adjusting with keystore details.

Test

You should now be able to go to http://localhost:8443.

Join Windows Domain

The ERA server must be joined to the domain

Install samba and stop any samba services that start.

# apt-get install samba krb5-user smbclient

Edit /etc/samba/smb.conf:

[global]
         workgroup = EXAMPLE
         security = ADS
         realm = EXAMPLE.COM
         dedicated keytab file = /etc/krb5.keytab
         kerberos method = secrets and keytab
         server string = Samba 4 Client %h
         winbind enum users = yes
         winbind enum groups = yes
         winbind use default domain = yes
         winbind expand groups = 4
         winbind nss info = rfc2307
         winbind refresh tickets = Yes
         winbind normalize names = Yes
         idmap config * : backend = tdb
         idmap config * : range = 2000-9999
         idmap config EXAMPLE : backend  = ad
         idmap config EXAMPLE : range = 10000-999999
         idmap config EXAMPLE:schema_mode = rfc2307
         printcap name = cups
         cups options = raw
         usershare allow guests = yes
         domain master = no
         local master = no
         preferred master = no
         os level = 20
         map to guest = bad user
         username map = /etc/samba/smbmap

Create /etc/samba/smbmap:

!root = EXAMPLE\Administrator Administrator admionistrator

Edit /etc/krb5.conf:

[libdefaults]
      default_realm = EXAMPLE.COM
      dns_lookup_realm = false
      dns_lookup_kdc = true
      ticket_lifetime = 24h
      forwardable = yes

Make sure that /etc/resolv.conf points to the AD DC, and dns is setup correctly.

Then run this command:

# net ads join -U Administrator@EXAMPLE.COM

Enter Administrators password when requested.

Edit /etc/nsswitch.conf and add ‘winbind’ to passwd & group lines.

Start samba services.

Update DNS

For some reason the join does not always create the DNS entry on Samba4, so you may need to add this manually.

References

• Samba
• ESET Remote Administrator Installation Instructions
• Troubleshoot Linux Installation

I have a problem with a Samsung M3 Portable USB3 external hard drive only working intermittently. I use a couple of these for off-site backups for work and so I need them working reliably. In fairness a reboot does cure the problem, but I hate that as a solution.

To troubleshoot this problem, the first step was to check the system devices:

# ls -al /dev/disk/by-id

But only my main two system drives were showing, and not the USB3 drive, which would have been sdc, being the third drive (a, b, c).

System Logs

The next step was to check the system logs:

# dmesg | tail

This showed no issues at all. Then I checked syslog:

# tail /var/log/syslog

This was completely empty, and had not been written to for almost a year. I can’t quite believe I haven’t needed to check the logs in all that time, but there you are.

I checked that rsyslogd was running and it was as user syslog. But the /var/log/syslog file was owned by root with group adm, but whilst syslog was a member of the adm group, the files all had user rw permissions only -rw-------.

This was easily fixed:

# chmod g+rw /var/log/*
# service rsyslog restart

Now the syslog file was being written to, but there was a problem writing to /dev/xconsole:

Nov 17 12:20:48 asusi5 rsyslogd: [origin software="rsyslogd" swVersion="7.4.4" x-pid="11006" x-info="http://www.rsyslog.com"] start
Nov 17 12:20:48 asusi5 rsyslogd: rsyslogd's groupid changed to 103
Nov 17 12:20:48 asusi5 rsyslogd: rsyslogd's userid changed to 101
Nov 17 12:20:48 asusi5 rsyslogd-2039: Could no open output pipe '/dev/xconsole': No such file or directory [try http://www.rsyslog.com/e/2039 ]
Nov 17 12:20:48 asusi5 rsyslogd: [origin software="rsyslogd" swVersion="7.4.4" x-pid="11006" x-info="http://www.rsyslog.com"] exiting on signal 15.

So I duly visited the link mentioned, which gave instructions for disabling /dev/xconsole, but this made me nervous and further research suggested that that was indeed the correct fix for headless servers, but possibly not for a desktop PC. Instead I used the following fix:

# mknod -m 640 /dev/xconsole c 1 3
# chown syslog:adm /dev/xconsole
# service rsyslog restart

USB Powersaving

Now at least it seems that my syslog is working correctly. Unfortunately unplugging and plugging in the USB drive still was not writing to the logs! When I plugged in the drive the blue light would flash repeatedly and then switch off. I would have believed that the drive had a fault, if it weren’t for the fact that rebooting the PC solves the problem.

Thinking that perhaps this problem was USB3 related, I decided to Google for “USB3 drive not recognised” which found this post. Except, that post was only relevant when operating on battery power, whereas I am using a plugged in desktop PC. Clearly that page could not be relevant? Except that in my notification area there is a battery icon, relating to my Uninterruptible Power Supply. But surely Ubuntu couldn’t be treating my UPS as if it were a battery? Could it?

In order to find out the power status of my USB devices I followed the suggestion of typing:

# grep . /sys/bus/usb/devices/*/power/control

Most were flagged as “on”, but a number of devices were flagged as “auto”. I decided to try switching them on to see if that made any difference:

# for F in /sys/bus/usb/devices/*/power/control; do echo on >"${F}"; done
# grep . /sys/bus/usb/devices/*/power/control

Now all the devices were showing as on. Time to try out the drive again - I unplugged it and plugged it back in again. This time the power light flashed repeatedly and then went solid.

Unfortunately the drive was still not mounted, but at least it was alive now. What next? Ah yes, I should check the logs to see if they have any new messages, now that the drive is powered and my logs are working:

# dmesg | tail
[15243.369812] usb 4-3: reset SuperSpeed USB device number 2 using xhci_hcd
[15243.616871] xhci_hcd 0000:00:14.0: xHCI xhci_drop_endpoint called with disabled ep ffff880213c71800

Faulty USB-SATA Bridge

At last I have something to go on. I searched for that error message, which took me to this Debian.net page, which suggested:

Reason is faulthy usb-sata bridge ASM1051 in USB3 drives. It has nothing to do with motherboards. See this. https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=a9c54caa456dccba938005f6479892b589975e6a Workaround was made to kernel 3.17-rc5. Now those disks works. Some perfectly and some works but data transfer is not smooth.

Could that be my problem? I checked my current Kernel:

# uname -r
3.13.0-68-generic

So yes, it could be. As I cannot read the device yet, I cannot check whether ASM1051 is present in my USB3 drive.

Having uninstalled 34 obsolete kernels and a further 8 newer kernels that weren’t in use, I was in a position to install the latest Kernel from Vivid with:

# apt install linux-generic-lts-vivid

The problem of course is that in order to boot onto the newer Kernel I must reboot, thereby fixing the problem anyway.

And finally

It would be easy to portray this as clear evidence of the difficulty of running Linux, certainly Samsung will have thoroughly tested their drive on Windows and Mac operating systems. That said such problems are not unheard of on the more popular operating systems and debugging such problems is far harder, I would argue, than the above logical steps.

Having rebooted, as expected, the drive worked. I tried running lshw to see if ASM1051 was present, but I could not find it. Of course upgrading the Kernel could have fixed the problem anyway.

For the present I will labour under the comforting belief that the problem is fixed and will update this post again when either the problem reoccurs on in a week’s time if all is well!

• Please see Intermittent USB3 Drive Mount Continued…

I assumed that, when I migrated from Wordpress to Jekyll, one of the casualties would be the old Wordpress comments. I had decided to use DISQUS for comments on Jekyll, partly because it is the “new thing” and partly because it is incredibly easy to achieve. But I never for a moment considered that migrating the old comments would be a possibility.

In fact, not only is it possible, but it is almost trivially easy.

I did encounter a few issues, as I had foolishly renamed some pages to more logical names, thereby both breaking all Google and other links, but also breaking the link with DISCUS comments. That was a very bad idea and as a consequence I spent quite a while renaming files back again, but other than that it all worked flawlessly. If you are going to rename pages - probably best to do it in Wordpress before migrating to Jekyll.

Well done [DISCUS] for making something that you would expect to be difficult easy. And thank you Wordpress for providing a decent export routine, without which my data would have been locked away.

I look forward to your comments!

Chris.

I was reading Phones need ‘bed mode’ to protect sleep on the BBC, that computers should have a blue light filter to help people to sleep properly at night. As I am frequently working on my laptop until late in the evening, I felt that I should investigate further. Currently I do turn down the brightness, but it seems that is not enough.

The article linked to a program called f.lux, which is free for Linux. I also came across this article on Linux Magazine Avoiding Eye Strain.

Installation

Getting this working on my laptop was trivial:

sudo add-apt-repository ppa:kilian/f.lux
sudo apt-get update
sudo apt-get install fluxgui

Finally I ran the program and set it to launch on start-up. I entered latitude as 51.2, which is close enough I believe.

Operation

The software feels fairly basic, whether this is the same for Windows and Mac I don’t know. That said the applet seems to run perfectly in the Ubuntu notification area.

It is very noticeable that the display is more muted and much more comfortable to view in the evening. I don’t yet know whether this will work well during the day, nor whether it will improve my sleeping. But the logic behind it seems sound and there is no reason why it shouldn’t help.

Following my recent announcement, I thought I would give some of my reasons for the move and some early impressions of using Jekyll.

What is Jekyll?

Jekyll is a Markdown based website and blogging platform, written in Ruby. The principle is simple - you write markown text files and they are automatically converted to static HTML webpages.

What is Markdown?

I am assuming that most of my audience have at least a passing knowledge of Markdown, but basically it is a very clean virtually syntax-free way of writing text files, so that they can be easily converted into different formats. The key to markdown is the conversion utility and I currently use Pandoc. I write the file once, and then I can convert into whatever format I want it in:

• PDF: pandoc -o sample.pdf sample.markdown
• Word: pandoc -o sample.docx sample.markdown
• HTML: pandoc -o sample.html sample.markdown

I would imagine most people start using Markdown so that can continue to use the favourite text editor - Vim or Emacs. At work I have found myself using it in preference to a word-processor, I have written a simple md2pdf perl script, so that in vim I can simply type :md2pdf % to have my document saved as a PDF. And the PDFs that Pandoc produces are beautiful and headings and sub-headings are automatically converted into PDF Bookmarks, giving your documents an effortless professionalism.

For complicated documents I sometimes start in Markdown and then move to LaTeX, but increasingly I am finding myself able to do virtually everything in Markdown, including simple images and hyperlinks. But you also have the option of extending plain markdown with HTML tags.

So in simplest terms Jekyll is just an automated way of creating Markdown files and converting them to HTML.

But why change from Wordpress?

Wordpress has been great for me, it’s relatively simple, has great statistical tools, a build in commenting system and much more besides. So why leave all that behind for something which is fundamentally so basic?

Benefits of Jekyll

1. Text Editor: Once again the desire to use Vim was probably the key motivation.
2. Github Pages: The fact that Jekyll could be used with the free Github Pages was another.
3. Command Line: The ability to use grep, sed and perl and all the other command line goodies makes for an incredibly versatile system.
4. Version Control: To have the whole site under version control.

I cannot tell you how wonderful it is to grep -l searchtext *.markdown | vim - and be able to edit each matching file in Vim.

Bootpolish Blog

There was another reason too, which was that I still had an old blog at bootpolish.net, which I wanted to close down. I could have merged it into my Wordpress blog, but I thought it would be easier to transfer it to Jekyll. To be honest I can’t say that it was particularly easy, but thankfully it is now done.

The Migration Process

I followed the GitHub Instructions for Jekyll. I used rsync to copy the underlying text files from Bootpolish.net into the _drafts folder, before using bash for loops to auto-process each into a Jekyll markdown post. I used the Wordpress Importer to transfer my Wordpress blog. The importer did not work particularly well, so I ended up editing each file in turn.

I found there was some customisation required:

1. Categories: By default Jekyll has no Category pages, for example: http://chrisjrob.com/category/technology/
2. Tags: By default Jekyll has no Tag pages, for example: http://chrisjrob.com/tag/3dmodel/
3. Wordpress RSS: I wanted to maintain the existing feed locations, which required creation of various additional feeds.
4. Tag Cloud: By default Jekyll has no tag cloud functionality, which I believe is crucial to a blog.
5. Site Search: By default Jekyll has no site search. There are plug-ins, but these are not compatible with GitHub pages. For now I have used Google Custom Search, but it has not yet indexed the entire site.

I have written a script to build all the tags and categories, which is working well. I would like to integrate this into the git commit command somehow, so that I don’t forget to run it!

Any new categories would require additional RSS feed files creating, by simply copying the feed template into the relevant category/feed folder.

Conclusions

This has been much more work than I would have liked. That said, I now have my Markdown files in a format under my control. I can move those files to any Jekyll provider, or indeed to any web provider and have them hosted.

In short, I am hoping that this is the last time I will move platforms!

Lastly, if you’re unfamiliar with Markdown, you can view the Markdown version of this page.

For the past few weeks I have migrating chrisjrob.com from Wordpress to Jekyll. I have also been merging in my previous blog at bootpolish.net.

This process has proved to be much more work than I expected, but this evening it all came together and I finally pressed the button to transfer the DNS over to the Jekyll site, hosted by GitHub.

I have tried to maintain the URL structure, along with tags, categories and RSS feeds, but it can’t be perfect and there will be breakage.

If you notice any problems please do comment below.

Thank you.

Chris Roberts

Being tall (6’2”), sedentary and over 40, I have succumbed to the inevitable back problems. Listening to Jono’s review of the LIFT Standing Desk on Bad Voltage reminded me that I’ve been meaning to try out working standing up, at least for part of the time. I haven’t gone for the LIFT, as I wasn’t able to find it on Amazon UK, but I have purchased the Bac Posture Stand, better known on Amazon as Portable Folding Notebook or Laptop Table - Desk - Tray - Stand - (Black).

Hmm catchy.

It appears to be well made, although it does take a little setting up. There is no fan cooling, but there are holes for natural venting. One pleasant surprise was that it does include a mouse stand, which is not shown in the Amazon picture. There was a cheaper alternative, but the reviews led me to believe that this one might work out better. As I am not planning to buy both - I will never know.

It is early days and I don’t yet have a standing mat for home, which I understand from that podcast is absolutely essential, but so far I am pleased with my purchase.

I have also invested in a VARIDESK Pro Plus 48 - Height Adjustable Standing Desk for work, along with a VARIDESK - Standing Desk Floor Mat.

I will review these in due course.

Following the demise of my Novatech n1410, I have now taken delivery of a shiny (actually “shiny” is one thing it isn’t!) new Lenovo ThinkPad X1 Carbon Gen 3. This is not a review of the X1 Carbon - there are plenty of those available already, but merely my experience of the first few days of ownership.

So why did I choose the X1 Carbon?

1. Quality of construction and life expectancy;
2. Ubuntu hardware certification;
3. Screen resolution 2560x1440 - perfect for remote support;
4. Keyboard and Trackpad quality - two of my dissatisfactions with the Novatech;
5. Weight: at 1.3kg - it is much lighter than the Novatech and much of the competition.

The high cost was a concern to me, but having paid £479 for the Novatech and it having only lasted 32 months, I was very conscious that price isn’t everything. I bought the X1 Carbon from BT Business Direct for a touch under £1,000.

When it arrived I was very surprised by the size of the packaging - this was a very small box for something so expensive! The one thing that I was not expecting was to be excited by the looks of this laptop - the words “business-like” best summed up my expectations. In the flesh I found the X1 Carbon absolutely stunning - the photo above is one of the few that I have found that do it justice.

By this point I was as happy as a Frenchman who’s invented self-removing trousers.

Ubuntu Installation

I downloaded Ubuntu 14.04-3 LTS 64-bit and copied it to my iodd 2531 Black USB 3.0 External Hard Drive Enclosure Virtual DVD Blu-ray ROM. The Ubuntu installation was flawless and fast. I opted for a whole drive encrypted LVM install and in no time was booting into Ubuntu. The screen was absolutely beautiful - for the first time I had a laptop with a screen to rival an Apple Macbook.

Display Issues

Ubuntu prompted me to run software update (I never update during the install process), which I duly did. Having rebooted - disaster struck - the boot process almost immediately dumped me to this screen. I was able to enter my drive encryption password blind and it then progressed to the beautiful LightDM login screen as before.

I logged this as a bug with Ubuntu and then installed an Ubuntu Mainline kernel and, whilst it still exhibited the screen corruption, it included a new Drive Encryption Password Prompt which did not suffer from the problem and gave me a functioning system once again (but read on for more).

Trackpad Buttons

I then noticed that the trackpad buttons were not working. Well that’s not entirely true - they were scrolling up and down a page, rather than behaving as normal trackpad buttons. This Ubuntu Question suggested that I was not alone and offered a solution - that of changing the trackpad from synaptic to imps driver. I tried this, but instantly the Trackpad felt less responsive, the cursor slowed down and whatever I tried to customise it made no odds.

But more importantly the Trackpad buttons still didn’t work.

It is worth noting that even without the Trackpad buttons, you can still left-click (single finger tap to Trackpad) and right-click (two finger tap to Trackpad). In reality this is what I shall use in day-to-day operation, so this is issue is not a show stopper - at least for me.

After much head scratching and trial and error I found that by rebooting into the original 3.19.0-25 kernel, holding the Shift key down to access Grub during boot process, not only fixed the Trackpad issue, but it also resolved the screen corruption issues during the boot process.

As a result, I decided to uninstall the newer Kernel 3.19.0-30, which contains the problem, along with the mainline kernel and generic LTS kernel packages. This will prevent any further updates to the Kernel and is clearly only a short-term fix, but the benefit is a perfectly working system.

Review of Reported Ubuntu X1 Carbon Gen 3 Issues

It is possibly worth reviewing that Ubuntu Question again and the issues raised therein:

1. Wi-Fi is disconnecting: I have not encountered this problem at all.
2. Click-buttons are not working at all: Resolved by returning to kernel 3.19.0-25.
3. Vertical Scrolling (with trackpoint and middle button): Not tested as yet.
4. Special Keys (like Brightness): these all work fine for me.
5. Screen Characters scrambled on 14.10: not experienced in 14.04 as yet.
6. Trackpoint Sensitivity: I think that was a by-product of the imps driver - which I have not needed to use.

In short, at least for the 3.19.0-25 kernel, this is a fully functioning Ubuntu laptop!

Fingerprint Recognition

I was expecting problems with Fingerprint recognition, but that was not the case. I simply installed a couple of packages, scanned my fingerprints and it all worked fine.

$ sudo apt-get install fprintd fingerprint-gui

There were two surprises: The first was that the fingerprint reader works for command line use - so no more typing the sudo password - simply swipe your finger! The second surprise was that it doesn’t appear to work for the initial LightDM login. The fingerprint scanner flashes a couple of times and then stops. Given that you only type this once in a session, this isn’t a huge deal, compared to the number of times you enter passwords for sudo and the lock screen.

There are reports of LightDM not working with encrypted home directories, but I am using whole Drive Encryption which I wouldn’t have thought would have been affected in this way, given that I have already entered my drive encryption password by this point.

If I manage to solve this problem I will of course update this page. If you have any solutions please do comment below.

WLAN - 4G Wireless

Lastly, an oddity of the BT Business Direct specification is that this laptop has a 4G WLAN slot. This was not something I particularly wanted, but of course now that I have it I have had to order a £6.25 per month Vodafone 4G Data Plan. Once the SIM card arrives I will provide feedback.

Conclusion

It is easy to be critical of some of the pitfalls in running Ubuntu on the X1 Carbon, but the reality, with a few caveats, is that it works flawlessly. The laptop is a joy to use, the keyboard feels wonderful and running your fingers across the trackpad is an almost sensuous experience - I may be getting carried away here, but suffice to say that this is a very good laptop and it is working beautifully on Ubuntu with the kernel 3.19.0-25.

With luck the bugs will be resolved and I will be able to update the kernel in due course!

I received a surprising amount of interest in my Novatech n1410 review and my subsequent nine-months on review.

Over the past 32 months I have continued to be fairly happy with my Novatech n1410. It has generally been fast enough, reliable enough, light enough and with a decent battery life. The downsides over the long term proved to be the keyboard and trackpad which have never been enjoyable to use and the lack of audio volume. Overall though I would say that it has lived up to my expectations for a budget Ultrabook.

But then last night - disaster struck, I was closing my laptop after working and the case opened up at the right hand hinge. This is not a heavily-abused laptop - it has never been dropped, I don’t commute, the laptop is stored in a padded case and treated at all times with respect. So why would the case fail in this way?

On closer inspection this would appear to be a natural failure point, the case after all is only plastic held together by a single screw at that point and the pressures on that hinge are immense. Indeed I am very surprised that it did not fail sooner. I think I probably contributed by being right-handed and probably putting a greater strain typically on the right hand hinge.

On the one hand, after 32 months I can’t really complain - it has certainly justified its cost. But I do feel a bit aggrieved that such a lightly used laptop should have failed within 3 years.

I can hear many of you thinking that 3 years is a long time for a laptop, but to put it in context my last laptop, a Toshiba Satellite Pro A300-1GO, was bought in 2008 and is still in daily use after 7 years.

I also feel embarrassed, I was encouraged to buy a ThinkPad, but the cost deterred me. Realistically a ThinkPad would probably have carried on working for at least another 2-3 years, and if you divide the cost by the life expectancy then perhaps the Novatech was a false economy. Well let’s face facts - it was a false economy. And I wouldn’t have had to be put up with a grotty keyboard for the past 32 months!

At the moment the Novatech is still working and I can close it just about - by the insertion of a ruler between the screen and the keyboard, and putting my weight on that as I lower the screen. I will live with it for the time being, but will start giving some thought as to my next laptop.

Possibly even a ThinkPad!

I was surprised to learn that Ubuntu 14.04 can talk to Skype for Business AKA Lync 2013 using the Pidgin Instant Messaging client. The general steps were:

# apt-get install pidgin pidgin-sipe

And then restart Pidgin and add a new Account. The Office Communicator is the relevant plugin, with the following parameters:

• Protocol: Office Communicator
• Username: Your Office 365 or Skype for Business username - probably your email address
• Password: Your password is obviously required - and will be stored unencrypted in the config file, so you may wish to leave this blank and enter at each login
• Server[:Port]: Leave empty if your set-up has autodiscovery
• Connection type: Auto
• User Agent: UCCAPI/15.0.4420.1017 OC/15.0.4420.1017
• Authentication scheme: TLS-DSK

I am unclear why the user agent is required, and whether that will need to change from time to time or not. So far it has worked fine here.

Unfortunately a few days ago the above set-up stopped working, with “Failed to authenticate with server”. It seems that you must now use version 1.20 of the Sipe plugin, which fixes “Office365 rejects RC4 in TLS-DSK”. As this version was only completed three days ago, it is not yet available in any of the Ubuntu repositories that I have been able to find, you will probably have to compile yourself.

Broadly speaking I followed these key stages:

1. Install build tools if you don’t already have them:

   sudo apt-get install build-essential

2. Install checkinstall if you don’t already have it:

   sudo apt-get install checkinstall

3. Download source files.

4. Extract source:

   tar -xvvzf pidgin-sipe-1.20.1.tar.gz

5. Change into source directory:

   cd pidgin-sipe-1.20.1

6. Read carefully the README file in the source directory.

7. Install dependencies listed in the README:

   # apt-get install libpurple-dev libtool intltool pkg-config libglib2.0-dev libxml2-dev libnss3-dev libssl-dev libkrb5-dev libnice-dev libgstreamer0.10-dev

These dependencies may change over time, and your particular requirements may be different from mine, so please read the README and that information should take precedence.

Lastly, as an ordinary user, you should now be able to compile. If it fails at any stage, simply read the error and install the missed dependency.

$ ./configure --prefix=/usr
$ make
$ sudo checkinstall

I found checkinstall was pre-populated with sensible settings, and I was able to continue without making any changes. Once complete a Debian package will have been created in the current directory, but it will have already been installed for you.

For some reason I found that at this stage Pidgin would no longer run, as it was now named /usr/bin/pidgin.orig instead of /usr/bin/pidgin, I tried removing and reinstalling pidgin but to no avail. In the end I created a symlink (ln -s /usr/bin/pidgin.orig /usr/bin/pidgin), but you should not do this unless you experience the same issue. If you know the reason for this I would be delighted to receive your feedback, as this isn’t a problem that I have come across before.

Restarting Pidgin and the Office Communicator sprung into life once more. Sadly I would imagine that this won’t be the last time this plugin will break, such are the vagaries of connecting to closed proprietary networks.

The first quote we received was three times what we expected, and we could have had a new kitchen for very little more. The second quote was more reasonable, but still 50% more than we were prepared to spend. Some desperate Googling came up with Affordable Granite.

There were other similar sites around, but some Googling revealed many positive customer reviews for Affordable Granite, and so we made contact. The quote was still over our initial budget, but far less than the other quotes received and we accepted their quote.

Well to cut a long story short, our new kitchen worktop looks fabulous and was not only less expensive than the other quotes that we received, had fewer compromises (decent thickness granite, included a new under-mounted sink and an additional breakfast bar), and was completed in a single week when other companies were suggesting two weeks. Better still our hob and sink were kept operational for all but two days, when one supplier told us that they would be disconnected at the beginning and reconnected at the end.

The actual installation did not go completely smoothly, thanks to our kitchen units being uneven and the adjustable feet having seized up, with the result that we had workmen on site for longer than expected. But at all times this remained their problem and not ours and they did everything possible to ensure that our kitchen was completed on-time.

In short we are delighted with our new kitchen and very happy with the service received from Affordable Granite.

Upgrading Ubuntu at work can make you rather unpopular, as the Internet bandwidth is fully utilised downloading all the updates to packages you have long since forgotten that you installed.

It also takes time, time that you should be working rather than upgrading your computer.

For these reasons I like to trickle download the upgrade over a day and only perform the actual upgrade once all the packages are ready, typically the following morning.

This is how I performed my low-bandwidth upgrade…

N.B. This is not the official or recommended way of upgrading between Ubuntu versions. Specifically my method involves manually disabling some repositories and updating others to the new release. This would normally be done by the do-release-upgrade program itself. It works for me, but please do be aware that you are deviating slightly from the official method.

Step 1: Disable 3rd Party Repositories

Launch the Ubuntu Software Centre and from the menu select Edit followed by Software Sources. Under the Other Software tab please untick all active repositories.

(This step should in any case be done automatically by step 4).

Step 2: Update Repositories

Edit /etc/apt/sources.list and replace all occurrences of  ”precise” with “trusty”. If you are of a brave disposition, the following command should do this for you:

$ sudo sed -i.bak 's/precise/trusty/g' sources.list

(This will create a copy of sources.list to sources.list.bak, in case you wish to reverse this.)

Step 3: Download Packages

Still in the terminal, type:

$ sudo apt-get updatesudo apt-get -o Acquire::http::Dl-Limit=64 -d dist-upgrade

The 64 will limit the bandwidth to 64 Kbps, please adjust to suit your available bandwidth. The “-d” will instruct apt-get to merely download the packages and not to install them.

I believe this stage can be aborted with Ctrl+C at any time and run again, until such time as all the required packages are downloaded.

Step 4: Upgrade

Still in the terminal, I tend to use GNU Screen for extra resilience, type:

$ sudo do-release-upgrade

Conclusion

I am typing this on my newly upgraded 14.04 installation, after a clean and trouble-free reboot and an entirely fault-free upgrade.

The truly astonishing aspect to an upgrade is the fact that the computer remains largely usable throughout. I lost my fonts briefly in one application during Step 4, but otherwise I was able to work normally. It didn’t even seem to be slowing my computer down greatly, although this is a fairly powerful workhorse, so your mileage may vary.

Please do comment, if you feel I’ve left anything out in the above, or indeed if you have found it useful.

Good luck with your upgrade.

I thought I loved markdown. I thought Dillinger (an HTML5 Markdown editor) looked remarkable. So I downloaded and installed it on an Ubuntu 14.04.1 LTS server:

$ git clone https://github.com/joemccann/dillinger.git dillinger
$ cd dillinger
$ npm i -d
$ mkdir -p public/files/{md,html,pdf}
$ sudo apt-get install nodejs
$ sudo ln -s /usr/bin/nodejs /usr/bin/node
$ npm install express
$ npm install phantomjs
$ cd dillinger
$ node app

It looks beautiful and works perfectly and even includes the ability to export to PDF. But it seems that, for me at least, markdown without Vim just isn’t the same.

Shame that.

This page explains how to set up spelling in Irssi on Debian Lenny. It assumes that you have already installed irssi.

Install packages

# apt-get install aspell-en libaspell-dev aspell libtext-aspell-perl

Don’t forget to change the localisation to match your own locale.

Create user’s scripts folder

$ mkdir -p */.irssi/scripts/autorun

Download the irssi script

$ cd */.irssi/scripts
$ wget http://www.mimir.ch/mimir/irssi/files/aspell_complete.pl
$ cd autorun
$ ln -s ../aspell_complete.pl

Restart irssi

Lastly, either restart IRSSI, or, in type:

/script load ~/.irssi/scripts/autorun/aspell_complete.pl

You should see aspell loading and your locale mentioned.

Testing

How the spelling functionality works is that you enter a word and then, before pressing spacebar, you hit the [TAB] key, whereon the preceding word will be spellchecked.

Troubleshooting

If you have not installed the correct aspell localisation, then you may find that irssi segfaults.

References

• http://www.mimir.ch/mimir/irssi/
• http://www.eckrall.co.uk/?page_id=55

At work I needed a cheap laptop for a computer-illiterate user. Giving them Windows, would have meant that they would have had to keep up-to-date with Windows Updates, with all the potential issues that would cause, along with the need for malware protection. It would also have pushed up the cost, a laptop capable of pushing Windows along reasonably decently, would have cost a few hundred pounds at least.

Generally I would just have purchased a low-end Lenovo laptop and installed Ubuntu onto it, but I was aware that Ebuyer had recently launched an HP255 G1 Laptop with Ubuntu pre-installed for £219.99 inc. vat (just £183 if you can reclaim the VAT).

Buying pre-installed with Ubuntu afforded me the comfort of knowing that everything would work. Whilst Ubuntu generally does install very easily, there are sometimes hassles in getting some of the function buttons working, for brightness, volume etc. Knowing that these issues would all be sorted, along with saving me the time in having to install Ubuntu, seemed an attractive proposition.

Unboxing

My first impressions were good, the laptop comes with a laptop case and the laptop itself looks smart enough for a budget machine. An Ubuntu sticker, instead of the usual Windows sticker, was welcome, although the two sticky marks where previous stickers had been removed were less so. Still, at least they had been removed.

Whilst we are on the subject of Windows’ remnants - the Getting Started leaflet was for Windows 8 rather than Ubuntu. Most Ubuntu users won’t care, but this is a poor attention to detail and, if this laptop is to appeal to the mass market, then it may cause confusion.

First Boot

Booting up the laptop for the first time gave me an “Essential First Boot Set-up is being performed. Please wait.” message. I did wait and for quite a considerable time - probably a not dissimilar time to installing Ubuntu from scratch; I couldn’t help but suspect that was precisely what was happening. Eventually I was presented with a EULA from HP, which I had no choice but to accept or choose to re-install from scratch. Finally I was presented with an Ubuntu introduction, which I confess I skipped; suffice to say the new user was welcomed to Ubuntu with spinny things.

The first thing to note is that this is Ubuntu 12.04, the previous LTS (Long Term Support release). This will be supported until 2017, but it is a shame that it didn’t have the latest LTS release - Ubuntu 14.04. Users may of course choose to upgrade.

Secondly, the wireless was slow to detect the wireless access points on the network. Eventually I decided to restart network-manager, but just as I was about to do so, it suddenly sprang into life and displayed all the local access points. Once connected, it will re-connect quickly enough, but it does seem to take a while to scan new networks. Or perhaps I am just too impatient.

Ubuntu then prompted to run some updates, but the updates failed, as “91.189.88.153” was said to be unreachable, even though it was ping-able. The address is owned by Canonical, but whether this was a momentary server error, or some misconfiguration on the laptop, I have no idea.

This would have been a major stumbling block for a new Ubuntu user. Running apt-get update and apt-get dist-upgrade worked fine, typing Ctrl+Alt+t to bring up the terminal and then typing:

$ sudo apt-get update
$ sudo apt-get dist-upgrade

I notice that this referenced an HP-specific repository doubtless equipped with hardware specific packages:

http://oem.archive.canonical.com/updates/ precise-oem.sp1 public
http://hp.archive.canonical.com/updates precise-stella-anaheim public

I assume that adding this latter repository would be a good idea if purchasing a Windows version of this laptop and installing Ubuntu.

Hardware

This is a typical chunky laptop.  But, if you were expecting a sleek Air-like laptop for £220, then you need to take a reality shower. What it is, is a good-looking, well-made, traditional laptop from a quality manufacturer. At this price, that really should be enough.

Ubuntu “System Details” reveals that this is running an “AMD E1-1500 APU with Radeon HD Graphics x 2”, running 64-bit with a 724GB drive and 3.5GiB RAM. That would appear to be a lower spec processor than is typically available on-line for an HP 255 G1 running Windows; which generally seem to have 1.75Ghz processors (albeit at twice the price).

The great news was that, as expected, all the buttons worked. So what? Well, it may seem like a trivial matter whether, for example, pressing Fn10 increases the volume or not, but I think many of us have the experience of spending inordinate amounts of time trying to get such things to work properly. And buttons that don’t work, continue to irritate until the day you say goodbye to that machine. The fact that everything works as it should is enormously important and is the primary reason why buying Ubuntu pre-installed is such a big deal.

The keyboard and trackpad seem perfectly good enough to me, certainly much better than on my Novatech ultrabook; although not everyone seems to like them. In particular, it is good to have a light on the caps lock key.

I have not tested battery life, but, as this is usually the first thing to suffer in an entry-level machine, I would not hope for much beyond a couple of hours.

For other details on hardware, please refer to the product information and read more reviews here.

Performance

Booting up takes around 45 seconds and a further 20 seconds to reach the desktop. That is quite a long time these days for Ubuntu, but fast enough I would imagine for most users and considerably faster than it takes Windows to reach a usable state, at least in my experience.

Being that bit slower to boot, Suspend becomes more important: Closing the lid suspended the laptop and opening it again brought up the lock screen password prompt almost immediately. Repeated use showed this to work reliably.

As to system performance, well frankly this is not a fast laptop. Click on Chromium, post boot, and it takes about 9 seconds to load; LibreOffice takes about 6 seconds to load. Even loading System Settings takes a second or two. Once you’ve run them once, after each boot, they will load again in less than half the time. Despite the slow performance, it is always perfectly usable, and is absolutely fine for email and web-browsing applications.

The other thing to remember is that this will be the performance you should be able to expect throughout its life - i.e. it will not slow down even more as it gets older. Windows users typically expect their computers to slow down over time, largely because of the different way in which system and application settings are stored in Windows. Ubuntu does not suffer from this problem, meaning that a 5-year-old Ubuntu installation should be working as fast as it did when it was first installed.

Conclusions

I struggle to think of what else you could buy that provides a full desktop experience for £220. And it isn’t even some cheap unbranded laptop from the developing world. Sure, it isn’t the fastest laptop around, but it is perfectly fast enough for web, email and office documents. And the fact that you can expect it to continue working, with few, if any, worries about viruses, makes it ideal for many users. It certainly deserves to be a success for HP, Ubuntu and Ebuyer.

But, whilst this low-price, low-power combination was ideal for me on this occasion, it is a shame that there are no other choices available pre-installed with Ubuntu. I wonder how many newcomers to Ubuntu will come with the belief that Ubuntu is slow, when in reality it is the low-end hardware that is to blame?

Please HP, Ubuntu and Ebuyer - give us more choice.

And Lenovo, please take note - you just lost a sale.

For more reviews please visit Reevo.

At a dinner party a few months ago, the host asked if I could repair their go-kart, using my RepRap 3D Printer. Apparently the steering wheel had broken off and the Chinese supplier refused to supply only the steering wheel. I accepted the challenge, but did warn that PLA was not the ideal plastic for such a repair, being somewhat brittle.

I decided to approach the problem by removing all the broken plastic from the back of the steering wheel, and designing a new part to slot over the steering column nut and bolt onto the remains of the steering wheel.

Having designed the part, I posted it to Google+ for advice. Wildseyed Cabrer suggested that I do away with the buttresses of v1 and instead design a solid cylinder. Of particular interest was his suggestion to use pinholes around the nut hole, to force the slicing program to add reinforcement. Andreas Thorn reminded me to use the $fn = 6 for the chamfering of the nut hole, as I had for the nut hole itself. The Google+ 3D printing community really is amazing - thank you.

Unfortunately printing the item was much delayed by the fact that I did not have the precise measurements. As always I had designed the item to be fully parametric; so that the exact measurements didn’t matter until I came to print. Unfortunately the owners of the go-kart did not feel able to accurately take the measurement; so instead the go-kart had to be delivered to our house.

Having measured and printed the final part, I was delighted that it all worked first time. The only issue was that it was very difficult to turn the wheel and I was concerned that the part would not indeed be strong enough. I noticed that all four tyres were completely flat, so much so that the tyre profile was concave rather than convex. This was clearly putting a huge strain on the steering and may well have been a contributory factor in the original breakage. Having pumped up all the tyres the steering was very much easier.

The repair complete, the go-kart was delivered back to the owners. I have no idea whether the repair will last long enough to be worthwhile, but providing it lasts a reasonable time, it can of course be simply reprinted. We could also consider having it professionally printed in a stronger plastic.

• View design at FabFabbers

I’ve been looking into the best method for attaching sheets to the foresail. Metal shackles are dangerous and can rip holes in sails, bowlines are unreliable knots when not under pressure, and a continuous line looped in the middle through the clew is likely to compress the clew over time.

The ideal solution appears to be to splice soft eyes into the end of each sheet, and attach them to the clew with a “soft shackle”. Being a lover of knots, I thought I’d give it a go and this was my first attempt. The surprising thing was the cost - I used 5mm Marlow Excel Vectram 12, at a cost of £5.91 per metre, which is enough to make a single soft shackle. The reason is the strength 5mm Vectram has a similar strength to 10mm braid on braid, 2400kgs breaking strain.

I followed the instructions in this video, which resulted in a 5 inch loop, too short for the intended purpose. A second attempt with a 7 inch loop looks long enough; although I have not yet tried it on the boat. The next step will be to splice soft eyes into each sheet.

De-Dietrich’s service agent wanted £195 to come out and fix the oven - a ludicrous price for a 13 year old microwave. Other appliance repair centres would either not repair microwave ovens at all, or particularly excluded De-Dietrich.

Fortunately one such appliance repair person recommended that we contact Visual Eyes in Farnham, Surrey. We did so, and the next day a nice man came and fixed our oven for £40.

Great service, thank you Visual Eyes.

N.B. If you received this post in your RSS aggregator and would rather not have done, then please be assured that such posts will be rare. If you’d rather not take the risk, then I do also maintain separate RSS feeds for Technology and Personal and even sub-categories such as Sailing and Linux. Just follow the relevant menu option from my home page and see the RSS feed in the address bar.

I have long been a keen user of pdftk, the PDF Toolkit, but am frequently surprised when people have not heard of it. True, it is a command line tool, but it is easy to incorporate into service menus, scripts etc and doubtless there is a GUI front-end for it somewhere (in fact there is one linked to from the above page).

Clearly a blog post is called for, but, whilst you wait for a post that will never arrive, here is a link to some examples that should open your eyes to what is possible with pdftk.

To get started on a Debian-based system:

$ sudo apt-get install pdftk
$ man pdftk

Enjoy.

Anyone who has enjoyed the dubious benefits of working with IPSEC will find OpenVPN a delight, but what do you do with your client.ovpn file once you have it?

If you spend most of your time in a terminal anyway, then I would suggest just putting all your client.ovpn files into ~/.openvpn, renaming them in some appropriate way, and then using them simply by typing:

$ sudo openvpn client.ovpn

If, on the other hand, you live in a more graphically orientated world, then you might like to integrate them into Network Manager. Sadly, the Import feature in Ubuntu does not work, at least in the versions of Ubuntu that I have used, and you have to make a few changes first.

Firstly, I would always create a hidden directory into which to store your client files:

$ mkdir ~/.openvpn
$ cd ~/.openvpn
$ mv ~/Downloads/client.ovpn ./

Secondly, you need to ensure that you have installed openvpn for network-manager:

$ sudo apt-get install network-manager-openvpn-gnome

Thirdly, we need to extract some data out of the client.ovpn, and for this I followed these instructions, which I include below in case of link breakage:

1. Open client.ovpn in your favour text editor and copy the lines between the <ca> tags into a new file named client.ca.
2. Remove <ca> section including tags.
3. Now copy the lines between the <cert> tags into a new file named client.crt.
4. Remove <cert> section including tags.
5. Now copy the lines between <key> tags into a new file named client.key.
6. Remove <key> section including tags.
7. Now copy the lines between <tls-auth> tags into a new file named client.tls.
8. Remove <tls-auth> section including tags.
9. Remove the line “key-direction 1”.
10. Insert the following text above the line # —–BEGIN RSA SIGNATURE—– :

ca client.ca
cert client.crt
key client.key
tls-auth client.tls 1

Finally, save and close all the files and check that you now have all the above files sitting happily in your ~/.openvpn directory.

Go to Network Manager -> Edit Connections ->VPN and click Import, browse to the modified client.ovpn import that file.

Enter vpn username and password if prompted.

On the VPN page, select Advanced and on the General Tab, uncheck the first option, “Use custom gateway.

I understand that I must now log on,
and pasting your password can no longer be done,
but why does it have to take me all day,
to reach what could have been one click away?

The fact that the invoice is live generated,
is clever technology, though that could be debated,
but would it not be easier all round,
if the invoice was sitting ready to bring down

I would not mind if the system was quick,
but I click and I wait and I wait and I click.
So next time you think of a clever solution,
think, will it be quick and will it be easy?

Unable to succesfully[sic] process request.
ERR-010: Request timeout.
Contact the application system administrator or Help Desk.
Sigh.

Very early on in my Linux life, I came across this suggested header for crontab and I’ve used it ever since. So much so that I am always slightly thrown when I come across a crontab without it! No, you don’t need it, yes the standard commented header works just fine, but, if like me you prefer things neatly lined up, then this might suit you:

MAILTO=
#   _________________________ 2. Minute - Minutes after the hour (0-59)
#  |
#  |      ______________________ 2. Hour - 24-hour format (0-23).
#  |     |
#  |     |      ___________________ 3. Day - Day of the month (1-31)
#  |     |     |
#  |     |     |      ________________ 4. Month - Month of the year (1-12)
#  |     |     |     |
#  |     |     |     |     ______________5. Weekday - Day of the week. (0-6, 0 indicates Sunday)
#  |     |     |     |    |
#__|_____|_____|_____|____|___Command_____________________________________________________________

And if you recognise this as your’s, then thank you!

Occasionally users are unable to connect to our FreeNX server, they report an error “Startup Session Failed”. Clicking on “Detail” shows that it is unable to find the server session file.

Searching for solutions suggested a number of options, including removing the server /tmp/.X1\*\*\*-lock files, or simply removing FreeNX and installing NoMachine’s NXServer instead.

In the end the solution proved remarkably simple:

On the server run:

# nxserver --list
NX> 100 NXSERVER - Version 1.5.0-60 OS (GPL)
NX> 127 Sessions list:
Display Username Remote IP Session ID
------- --------------- --------------- --------------------------------
1001 chris 192.168.1.52  1BC6B4B9C3CF4C2B6BD7137AC7FDE5DA
1000 helen -             87443D16622EC0751551685A93DD023B
1002 michelle 192.168.1.102 DBAC430C8AA8B414A5E2228970E2BBDC
NX> 999 Bye

The session with no remote IP is for a session that has not ended properly. Terminate this session and users should be able to log in once again:

# nxserver --terminate helen

Update: This problem was little more complex than I at first thought, all I had really done is allow one more user to login before the issue re-occurred. You also need to check is that there are no old lock files in /tmp/.X11-unix:

# cd /tmp/.X11-unix
#ls -al
/tmp/.X11-unix# ls -al
total 316
drwxrwxrwt   2 root root      4096 2014-01-02 12:48 .
drwxrwxrwt 375 root root    315392 2014-01-02 12:48 ..
srwxrwxrwx   1 root root         0 2013-07-24 17:04 X0
srwxrwxrwx   1 helen helen       0 2014-01-02 10:32 X1000
srwxrwxrwx   1 chris chris       0 2014-01-02 11:42 X1001
srwxrwxrwx   1 michelle michelle 0 2014-01-02 08:52 X1002
srwxrwxrwx   1 terry  terry      0 2013-09-05 15:05 X1003

Notice that there is a .X11-unix file for terry X1003, but no corresponding user shown in nxserver –list above. Remove this spurious file and it should now work.

It would also make sense to ensure that the correct /tmp/.X1***-lock files are present:

# cd /tmp
# ls -al | grep -i X.*-lock
-rw------- 1 nx nx              0 2014-01-02 10:32 .nX1000-lock
-rw------- 1 nx nx              0 2014-01-02 11:42 .nX1001-lock
-rw------- 1 nx nx              0 2014-01-02 08:52 .nX1002-lock
-r--r--r-- 1 helen helen       11 2014-01-02 10:32 .X1000-lock
-r--r--r-- 1 chris chris       11 2014-01-02 11:42 .X1001-lock
-r--r--r-- 1 michelle michelle 11 2014-01-02 08:52 .X1002-lock

You should expect to see  the correct number of lock files for your user sessions,  in my case these required no changes, but if there had been spurious files, removing them would seem sensible.

If this has been helpful to you, please do consider rating this post or adding a comment!

Whilst the world seems to be moving email to “The Cloud”, rightly or wrongly I remain reluctant to give up the control of our own mailserver. For over ten years now we have been using a combination of the following open source applications:

• Postfix MTA with Amavis + Spamassassin
• Dovecot IMAP server
• OpenLDAP

Whilst these are super-stable and bulletproof solutions, the main issue with such a solution is the administration - setting up users, changing passwords, vacation notifications, sieve etc. It is also a lonely business administering a custom mailserver, where set-up is never going to be completely standard.

A couple of years ago I migrated to a new mailserver, on which I installed iRedMail. iRedMail is a pre-packaged mailserver solution, based on all the software that I was using already (Postfix, Dovecot, OpenLDAP), but with the benefit of a slick admin panel that pulls it all together and provides ongoing support. It also incorporates Roundcube for webmail, undoubtedly prettier than Squirrelmail, which we were using before. iRedMail even includes a pre-configure Fail2ban, essential to protect your internet-facing server from attack.

There are two web administration panels to choose from - the free and open source panel and the Pro panel. Both interfaces are attractive and functional, but predictably the Pro panel has more functionality at a price. Some would call this Crippleware, but bear in mind that the server itself has full functionality - this is merely the configuration of that server. If you are happy to configure LDAP yourself, then you don’t strictly need to use the Pro panel. See “Features and Comparison” half way down this page. Whilst the Pro panel does come with the source code, the developer explicitly restricts your right to distribute.

Installation is very fast - but must be on a fresh server - I tried to upgrade from an existing Postfix install and it ended badly. Updates are delivered as a web page of instructions, which I have come to prefer to running an upgrade script, as you can intelligently decide whether you are happy with each change proposed. The downside is that you could easily miss a step and leave your server only partially “upgraded”.

Whilst the forums are quiet, the developer is very responsive to posts and has dealt efficiently which each issue that I encountered. Roundcube has proven to be a very attractive webmail interface and includes Sieve administration for server-based filtering and vacation notices.

Overall we are very happy iRedMail users - if there is an easier way of running an Open Source mailserver, I have not discovered it.

For more information please visit the following link:

• HowTo: Make umount Work with sshfs | Unicom Systems Development

The power switch remains a momentary petty irritation, but the poor quality keyboard and intrusive trackpad are harder to live with. The wireless is pretty poor, but it generally connects fine and rarely actually causes me any problems.

But the Novatech n1410 remains a good-looking, lightweight laptop with a great battery life, and highly portable. And it was undeniably great value.

One new irritation, that I hadn’t noticed in my earlier review, is that the volume can be too quiet in some circumstances. If sound is important to you, and particularly if you require it to deliver the goods in noisy environments, then you will probably need to invest in some external speakers, for example a pair of XMI X-Mini MAX V1-1 Capsule Speakers.

Nine months on, and I decided to upgrade Ubuntu 12.10 to 13.04 and 13.10. The upgrades both went flawlessly and as usual I do wonder why Ubuntu users seem so tied to re-installing instead of upgrading.

But unfortunately it hasn’t been all good news - the upgrade to 13.04 saw the introduction of two bugs:

Brightness on Boot: On booting up, brightness is set to the maximum every time. The numerous suggested fixes didn’t seem to work and so, pending a better solution, I have installed “xbacklight” and added it to the start-up applications with the switches “-set 20” to set it to 20% brightness. And yes, I am aware that I could have just echoed the chosen brightness to the relevant device file:

# echo 1102 > /sys/class/backlight/intel_backlight/brightness

Volume FN Buttons: The volume buttons no longer work correctly, on the first press the volume shoots up to maximum or minimum and then the indicator keeps flashing, locking the keyboard and requiring a REISUB reboot. For the time being I have simply re-mapped the F10, F11 and F12 buttons to do the tasks that would normally be done by FN F10, FN F11 and FN F12 (see screenshot).

But I cannot fairly blame Novatech for these bugs, what I need to do is report bugs for them with Ubuntu. I have found an article about Backlight Debugging and another on Hotkey Debugging, both of which I will be following at some point.

For more information please visit:

• Linux Terminal Command Reference

Warning: The command '/usr/bin/unhide.rb' has been replaced by a script: /usr/bin/unhide.rb: Ruby script

Googling confirmed that this error was normal on Ubuntu systems, but I found no solution. Fortunately the solution was simple, simply editing /etc/rkhunter.conf and adding the following line at the appropriate place:

SCRIPTWHITELIST=/usr/bin/unhide.rb

My daughter wanted a small windmill for a school project, and this was my very quick’n’dirty attempt.

For more information please:

• View it on GitHub

My daughter needed a small bridge for a school project and I couldn’t find anything suitable on-line. This was my very simple design, intended to be fully parametric, but probably only within certain limits.

For more information please:

• View it on GitHub

Having been quoted £250 to replace the halyards and topping lift on my boat, I decided to do it myself, by purchasing the ropes from eBay for £100 and the fids to do the splicing for £33.

Having the fids would then enable me to do the rest of the running rigging: reefing pennants, kicking strap etc.

The only trouble was that I didn’t know how to splice braid-on-braid, and it turned out to be more tricky than I expected. In the end I found a YouTube video to help me.

Having bought an inflatable Avon Redcrest dinghy on eBay, I found that the pump no longer had a clip holding the two halves together. I knew that the older models used to have a piece of leather attached to one half with holes that clipped over screws on the other half, but on this pump only the screws remained.

Newer pumps instead came with a plastic clip that attached to the pump hose and clamped the two halves together and I felt that it would be quite simple to replicate and so it proved. This was a simple design exercise and worked perfectly first time. I have since re-sold the dinghy and hope that the pump clip is still serving its new owner.

For more information please:

• View it on GitHub

Finally got around to putting curtains up in our guest room and the curtain rings had been lost. They did eventually turn up, but these printed rings worked perfectly until we did. Yes they’re not pretty, but you really could not tell unless you got close.

For more information please:

• View it on GitHub

My Westerly sailing boat has a rectangular-shaped shorepower socket and the faceplate was broken in half. Given its very exposed position in the cockpit, this is only to be expected, but purchasing a replacement proved very difficult. Fortunately the broken half had been kept by the previous owner, which enabled me to design an exact replica. And the beauty is that it will certainly be broken again, so I have printed a spare.

For more information please:

• View it on GitHub

My mother-in-law’s curtains were pulled down by a Great Dane, breaking many of the curtain hooks. They seemed to be an unusual design and she asked if I might be able to print them on my RepRap.

This was very satisfying design, which did not take long, posed no particular challenges and printed very quickly. If only all jobs were like that.

For more information please:

• View it on GitHub

A friend saw a plastic bag carrier for sale and wondered if I could produce a 3D model suitable for printing. Here was the result. A simple handle with hooks to carry multiple bags at once.

For more information please:

• View it on GitHub

My first ever design was a shock cord toggle, but I have only now had a reason to print it and I am pleased to say that they are working perfectly. Shown here on my dinghy

• used to hold the oars when not in use.

For more information please:

• View it on GitHub

On my boat there was a locker below the cooker that was not closing properly. The problem was that the retaining clip was broken. Normally the only solution would be to buy a new clip, but this is the perfect use for a 3D printer, as it enables you to re-create just the broken half. And, even better, you can design the screw holes to be in the same place, enabling a super quick replacement.

For more information please see it on GitHub.

• View it on GitHub

As discussed in my earlier The Novatech nFinity n1410 Review, I made a mistake in sticking with the default 128gb SSD. I promised to write up how I carried upgraded my Ultrabook by adding an mSATA drive, but time has passed and I failed to take any notes.

The drive I ordered was a Crucial CT256M4SSD3 256GB m4 mSATA 6Gb/s Internal SSD.

The installation was not difficult, and to the best of my recollection followed these steps:

1. Undo screws on bottom;
2. Use knife to gently release the catches around the keyboard;
3. Carefully remove keyboard;
4. I seem to recall that there were additional screws behind the keyboard, which need to be removed;
5. Fit mSATA drive, fixing with screw supplied with the drive;
6. Close case.

Apologies that these instructions are so poor, but hopefully the photographs in the Picasaweb album will help.

This is how I installed x11vnc on our Debian Lenny LTSP5 server, these are my notes, nothing more. Use at your peril!

Installing x11vnc

Enter the chroot

$ sudo chroot /opt/ltsp/i386

And update the repositories that the ltsp environment uses for apt-get:

# apt-get update
# apt-get install x11vnc

Create start-up script

Still in the chroot, edit /etc/init.d/x11vnc and copy in the following script:

#!/bin/sh

### BEGIN INIT INFO
# Provides:x11vnc
# Required-Start:$remote_fs $syslog
# Required-Stop:$remote_fs $syslog
# Default-Start:2 3 4 5
# Default-Stop:0 1 6
# Short-Description:Start X11VNC
# Description:Start VNC server X11VNC at boot
### END INIT INFO

case "$1" in
        start) 
                sleep 6
                XAUTH=`find /var/run/ldm-xauth* -type f`
                logger -f /var/log/x11vnc "Starting with $XAUTH"
                start-stop-daemon --start --oknodo --pidfile /var/run/x11vnc.pid --background --nicelevel 15 --make-pidfile --exec /usr/bin/x11vnc -- -display :7 -loop -passwdfile /etc/x11vncpassword -nossl -logfile /var/log/x11vnc -auth $XAUTH
        ;;
        stop)  
                logger -f /var/log/x11vnc "Stopping"
                start-stop-daemon --stop --oknodo --pidfile /var/run/x11vnc.pid
        ;;
        restart)
                logger -f /var/log/x11vnc "Restarting"
                $0 stop
                $0 start
        ;;
        condrestart)
                PID=`cat /var/run/x11vnc.pid`
                RUNNING=`ps h --ppid $PID`
                if [ "$RUNNING" == "" ]; then
                        logger -f /var/log/x11vnc "No process matching /var/run/x11vnc.pid"
                        echo "No process matching /var/run/x11vnc.pid"
                        $0 restart
                else   
                        logger -f /var/log/x11vnc "Process matching /var/run/x11vnc.pid exists"
                        echo "Process matching /var/run/x11vnc.pid exists - no action taken"
                fi
        ;;
        *)
                echo "Usage: $0 start|stop|restart|condrestart"
                exit 1
        ;;
esac

exit 0

Finalise start-up script

Still in the chroot, make the script executable and link it into /etc/rc2.d:

# chmod 755 /etc/init.d/x11vnc
# update-rc.d x11vnc defaults
# ls -al /etc/rc2.d/*x11vnc

Set x11vnc password

Finally, create the /etc/x11vncpassword file with the password you want to use to connect to your thin clients:

# echo "thepassword" > /etc/x11vncpassword
# chmod 400 /etc/x11vncpassword
# chown root:root /etc/x11vncpassword

Replace “thepassword” with your chosen password.

Alternative Method

The issue with the above method is that if you user does not reboot their client after logging off, then x11vnc does not restart. An alternative, remove the above symlink from /etc/rc2.d/S99x11vnc and instead use an ldm start up script by adding the following two scripts:

# /opt/ltsp/i386/usr/share/ldm/rc.d/I99-x11vnc
# LDM Script to start x11vnc

XAUTH=`find /var/run/ldm-xauth* -type f`
start-stop-daemon --start --oknodo --pidfile /var/run/x11vnc.pid --background --nicelevel 15 --make-pidfile --exec /usr/bin/x11vnc -- -display :7 -loop -passwdfile /etc/x11vncpassword -nossl -logfile /var/log/x11vnc -auth $XAUTH

The “I” in I99-x11vnc denotes that this will be run after X and before LDM, which is perfect for VNC.

# /opt/ltsp/i386/usr/share/ldm/rc.d/X99-x11vnc
# LDM Script to stop x11vnc

start-stop-daemon --stop --oknodo --pidfile /var/run/x11vnc.pid

The “X” denotes that this script will be run upon logout, so this is closing down the x11vnc server, ready to restart with LDM.

This method will still not cope with X being terminated with ctrl+alt+backspace, in which scenario you can ssh onto the client and run /etc/init.d/x11vnc condrestart - you could also add this in as a local app to be accessible from the server desktop.

References

• https://wiki.edubuntu.org/InstallX11VncOnLtspClients

XPS is Microsoft’s attempt to replace PDF, the only difference is that everyone can read PDFs, and not everyone can read XPS. I understand that KDE 4 versions of Okular will support XPS, which may make these instructions unnecessary, although having a tool for conversion readily at hand is always useful!

These instructions were tested in Debian Lenny. These instructions worked for our specific systems YMMV.

Building from source

$ sudo apt-get install libxext-dev libxt-dev
$ wget http://ghostscript.com/releases/ghostpdl-8.71.tar.bz2
$ tar xvvjf ghostpdl-8.71.tar.bz2
$ cd ghostpdl-8.71
$ make xps

Testing

After the build you will find gxps in xps/obj

To test, you will need a test document in XPS format.

$ cd xps/obj
$ ./gxps -sDEVICE=pdfwrite -sOutputFile=test.pdf -dNOPAUSE test.xps

Move to bin

You probably want to move the gxps executable into a convenient location within your PATH. /usr/local/bin may be a good destination. Once there you ought to be able to run the command from anywhere and it just work. Not sure what your PATH is? Type “echo $PATH” in your terminal.

$ echo $PATH
$ sudo cp xps/obj/gxps /usr/local/bin/
$ sudo chown root:root /usr/local/bin/gxps

Creating file type

XPS probably does not exist on your Linux system as a file type, you can either create yourself using KDE Control Centre, or in KDE:

Single user:
~/.kde/share/mimelnk/application/xps.desktop

All users:
/usr/share/mimelnk/application/xps.desktop

[Desktop Entry]
Comment=XPS Document
Hidden=false
Icon=application-xps
MimeType=application/xps
Patterns=*.xps;*.XPS
Type=MimeType
X-KDE-AutoEmbed=false

Adding to ServiceMenu

If you are using Konqueror, you can add a service menu (to enable right-click / action menu).

Single user:
~/.kde/share/apps/konqueror/servicemenus/xpstopdf.desktop

All users:
/usr/share/apps/konqueror/servicemenus/xpstopdf.desktop

[Desktop Entry]
Version=1.0
Encoding=UTF-8
Name=xpstopdf service menu
ServiceTypes=application/xps
Icon=acroread
Actions=xpstopdf

[Desktop Action xpstopdf]
Icon=acroread
Name=Convert XPS to PDF
Exec=cd "%d"; gxps -sDEVICE=pdfwrite -sOutputFile="`echo "%f" | cut -d . -f 1`.pdf" -dNOPAUSE "%f"; mv "%f" ~/.local/share/Trash/files; kdialog --title "Convert XPS to PDF" --passivepopup "Done" 3; echo;

Testing ServiceMenu

You should now be able to right-click on the file and “Convert XPS to PDF”. This will create a PDF of the same name and move the XPS into trash.

References

• http://www.ghostscript.com/GhostPCL.html
• http://obscured.info/2010/03/01/converting-xps-to-pdf-on-ubuntu-9-10/
• http://blog.rubypdf.com/2009/04/14/convert-xps-to-pdf-in-two-ways/

I realise that I have made a number of posts on Google+, without posting first on my blog. This is the first of several posts that I will be making to correct that oversight…

Sailing boats typically have plastic or rubber cylindrical covers for the lower part of the shrouds and stays (the wires that hold up the mast). These covers are apparently called “Turnbuckle Boots”.

Turnbuckle boots should be topped off with caps, but it is quite common to see the caps broken or missing entirely.

Turnbuckle Boot Cap is a replacement for those caps. It is fully parametric and printed perfectly without support:

• View on GitHub

# update-alternatives --config x-www-browser

Unfortunately of course many of the applications that I am using are native to KDE or Gnome and probably are still respecting their environment’s settings. In the end it was simply a case of editing:

~/.local/share/applications/defaults.list

And adding the following lines:

x-scheme-handler/http=chromium.desktop
x-scheme-handler/https=chromium.desktop

Now opening links from my terminal is correctly opening a new tab in Chromium, or running Chromium if it isn’t already.

Joy.

On entering Currys there was plenty of evidence of the new Windows 8, but I noticed that all the copies were upgrades from Windows 7 or Windows XP.  Currys explained that they do not stock full copies of Windows, only the upgrades, and stated that this was not a Curry’s issue, but that it was a Microsoft policy to only sell full versions of Windows via their website.

So I found myself searching online for “buy windows 8” and ended up on the Microsoft site, but, as with Currys, the only versions available were upgrades.

Starting to feel like I’d entered the Twilight Zone, I searched at BT Business Direct, this seemed to be much more successful and I found 4 choices available, but all the versions were OEM copies which I assumed that I was not legally permitted to install on a VM. It appeared that the only choices were OEM licences or Retail upgrade licences, on the face of it - leaving users like myself unable to legally buy Windows at all.

I thought perhaps that this was a short-term anomaly post-launch, but it seems not. Apparently the OEM version is all things to all people, being both a Retail copy for non-system builders, and an OEM copy for system builders (read more). If this is correct then this means that those 4 choices at BT Business Direct may be okay for me afterall.

And here is the word from Microsoft on the matter: “If you are building a computer for your personal use or installing an additional operating system in a virtual machine, you can now purchase OEM System Builder software using the Personal Use Licence.” After pouring over the text of EULAs, this is actually easy and unequivocal. Well done Microsoft.

Knowing my preference for all-things GNU/Linux and FLOSS, some of you may be wondering why I need Windows at all. The reasons are very few and I don’t use Windows from one month to the next, but I would find it difficult to eradicate completely:

1. Tax return: I know that it can be done under Linux, but I prefer to use TaxCalc.
2. Road Angel: I have found no way of updating my Dad’s Road Angel without Windows.
3. Inforad: Similarly I have found no way of updating my Inforad without Windows.
4. Leappad2: I have found no way of updating my childrens’ Leappad2’s without Windows.

So there we have it, we can now officially buy the OEM version for our VMs.

At least I think so.

After spending much time trying to choose the perfect Ultrabook, I came to the conclusion that it is either not made, or is prohibitively expensive. In the meantime I ordered a Novatech nFinity n1410 (14” Intel Core i5 3317 Mobile Processor - 8GB DDR3 Memory) for just £450 (plus VAT).

My expectations were fairly low, given the price, but I expected something that would be adequate and I was pleased to be able to buy a laptop without a Microsoft operating system pre-installed.

Purchase and delivery were quick and painless, thanks to Novatech.

My first impressions were very agreeable, it looked much better than expected with a metal top. The dimensions were just as I had expected - it would fit in my briefcase and be light and portable, but still have a decent sized screen.

Then I went to switch the n1410 on, oh dear, the power switch is terrible! Cheap, nasty, with a horrible unsatisfactory movement. It still irritates me every time I use it. It glows blue when powered and red with disk activity. Once I went to put the laptop in my briefcase and the flashing blue button reminded me that it was still in standby - so yes the button is clearly functional - but still I hate it.

Installing Ubuntu 12.10 was very simple, with no issues whatsoever. With Ubuntu installed and working well, it was time to reboot. I made the mistake of glancing away from the screen for few seconds, and it was already sitting at the login prompt. It boots in less than 15s, it takes longer to shutdown, but this is a great benefit for a portable laptop.

Logging in and things just got better - the buttons all work, including the FN buttons like brightness and media playback. This is seriously impressive and the Ubuntu community deserve congratulations, as this is no mean feat. Given how painless the installation is, you can’t but wonder why Novatech don’t offer Ubuntu as an option.

This all seemed to good to be true, and it was, suddenly the wireless signal dropped out, despite my sitting less than a metre away from the wireless access point. And this kept happening. Googling for an answer I ended up adding the following line to the end of /etc/modprobe.d/iwlwifi.conf:

options iwlwifi 11n_disable=1 bt_coex_active=N

This seemed to make the laptop usable, but it is not a long-term solution. I was advised to try a newer kernel, and I am now on 3.6.3, but it seems no different. The problem seems to be a bug with the iwlwifi module,  and it does seem to be known about, so hopefully a fix will wander down to Ubuntu soon. Maybe upgrading to 13.04 will fix all, but that is one for another day.

I tried closing the laptop, whilst still on, and it promptly went into standby mode. I opened the lid again and the laptop sprung back into life, although the wireless was not connected. Restarting network-manager revived the wireless connection. I can live with that. In fact I do find that I frequently need to restart network manager when first powering up, which may well be more evidence of the flaws in iwlwifi.

The screen is glossy, which is never a good thing, but other than that I cannot fault it. The speakers are a little tinny, but I suspect that that is par for the course with an Ultrabook. The keyboard feels a bit cheap and I keep missing letters, or getting letters twice. I hope I get used to this, but it has to be said that the keyboard is not a pleasure to use.

I am also struggling with the trackpad. My previous laptops have had smaller trackpads, which I would occasionally catch when typing. This laptop has a large trackpad, which I am constantly catching, and it is driving me slightly mad. I am not sure that this is the fault of Novatech though, in theory Ubuntu should disable the trackpad when typing, but in my experience it could work better. Maybe I can improve this is some way and I will do further research.

Another trackpad irritation is that the right hand side of the pad seems to be the right-mouse-click, and the left hand side - the left mouse click. Whilst flawlessly logical, it means that right handed users have to travel a long way for left click. I suspect that this is a trait of new Xorg versions, but I have done no research on the matter. Maybe I will get used to it.

The trackpad does have buttons below it, but they are simply horrible to use, requiring a considerable pressure to work, for that reason I tend to only use the trackpad itself.

Sticking with the default 128gb SSD was a mistake though, as I could not even transfer my Pictures folder (blame my young family and camera touting wife!). I knew that popey had added an mSATA drive to his Lenovo X220, and I telephoned Novatech to find out if this would be possible with the n1410. The answer was that yes it had an mSATA port, but that it was limited to 32gb and would only be used to improve the boot speed. I was not convinced that they were correct and Googling the subject showed that Dell had said the same about their laptops, and it was not true. With some concern I ordered a Crucial CT256M4SSD3 256GB m4 mSATA 6Gb/s Internal SSD. To cut a long story short- this proved successful and I intend writing up the experience in a separate post.

Battery life seems excellent - with past laptops I have generally used them connected to the mains, but the battery life on the n1410 is good enough that I am finding myself using it more like a tablet, in leaving it on most of the time. I believe 5 hours should be possible.

Overall this is a very good value Ultrabook. Clearly it is not perfect, but I never expected it to be, I expected it to be adequate and functional, and that it certainly is. That it is also attractive, with a reasonable screen, battery life and all working with Ubuntu is just fantastic. All in all I am very pleased with my new Ultrabook.

So, with a heavy heart I reached for my dog-eared copy of the Internet, to see what this would cost me. I knew I needed Microsoft Office Professional, as Microsoft Access is occasionally needed. I found that a full retail licence from BT Business Direct cost £205 plus VAT, not so bad after all.

I considered buying 12 of these full retail licences, but decided in the end to contact our BT account manager to request a quote. I was informed by BT that the full retail licences could not be used on a server; no explanation was given for this, but I was assured that this was the case.  The result was that the cost would in fact be £288 each, an additional £1000. A rather strange reversal of the usual - the more you buy, the lower the unit price.

As I was going to be installing on a virtual machine, I was also interested in knowing that I would be able to reinstall on a new virtual machine, if for any reason I needed to rebuild. For some reason this seemed far less clear than I would have liked, but ultimately I did receive that assurance; albeit in a way that left me wondering if that would indeed ultimately prove to be the case.

One additional confusion is that 2013 is just out, and so there was a choice of 2010 or 2013, we had been told to buy 2010 or later, in order to work with Business Intelligence, but then I read an article on ZDNet Can Microsoft bring BI to the masses if the Excel 2013 masses can’t get BI?. The upshot seems to be that Microsoft Office Professional 2013 may not be enough - I might need Microsoft Office Professional Plus 2013, which as far as I can see is not even listed on BT Business Direct.

I visited the Microsoft page on Microsoft Office Professional Plus 2013 but this left me none the wiser. Following the link to Licensing Options looked promising, but was not. Following the link to View Licensing Options - seemed to suggest that the only option was a three year Enterprise Agreement, but there was no pricing shown, not even under the “Volume Pricing” heading.

So now I need to re-contact BT to find out if they offer Professional Plus 2013 and try and find out if this is what I need and whether I can in fact reinstall on different hardware.

To seasoned purchasers of proprietary software this might all seem par for the course, or perhaps there is an easier way that I have not yet found? I suspect that the truth is that we are too large a company for buying single licences, but too small a company for an enterprise agreement. Neither fish nor fowl, as the saying goes.

But, for the past 5 years or so, I have not had to think about licensing once. If I need a copy of office, then I download OpenOffice (or more recently LibreOffice); if I need a desktop publishing program, then I download Scribus; a graphics editor, the GIMP; an illustrator, Inkscape. And for each of those programs I can install it for all staff without even a thought about how many licences I might have available.

There is no way out for me, sadly, and I will need to navigate these difficult waters whether I like it or not, but there is a better way and that way is Free Libre Open Source Software (FLOSS).

Surrey Shore-based RYA Training

I was delighted to find this local RYA training centre. Operated from Phil’s back garden, you expect the worst, but actually the facilities are excellent, with a proper classroom with tables, projector, VHFs, tea and coffee etc.

I recently had my Marine VHF training there and I thought the course was excellent and I plan to go on to do my Day Skipper training at a later date.

Following an earlier post on using GPS under Ubuntu, I have been trying to get PolarNavy working under Ubuntu 12.10. Polar Navy is the only Linux chart navigation software available for Linux. I know people will correct me by quoting OpenCPN, but there are no charts currently legally available for the UK, at least that I could find.

Polar Navy has two components - PolarCOM which communicates with the GPS receiver, and PolarView which displays the actual charts. The two applications can connect together, to show the vessel’s current position.

I had Polar Navy working fine under Ubuntu 12.04 i386, but for some reason I had problems running under Ubuntu 12.10 amd64. The problem was that PolarCOM would simply not display the position, i.e. the Lat and Long. I tried running from the command line, and there were a large number of errors when running PolarCOM, for example:\

menu_proxy_module_load': ./PolarCOM.bin: undefined symbol: menu_proxy_module_load (PolarCOM.bin:4487): Gtk-WARNING **: Failed to load type module: (null)

I eventually fixed that by adding the following to the /opt/polarcom/bin/PolarCOM script:

export UBUNTU_MENUPROXY=0

Sadly PolarCOM was still not working. The GPS Receiver that I use is the BU-353, a terrific piece of hardware that “just works” under Linux. This is the recommended GPS for Polar Navy, which was a happy coincidence. When I plugged in the receiver it was detected as /dev/ttyUSB0, which I noticed had a group “dialout”. I added myself to that group:

$ sudo adduser dialout

I posted a request for help on the Polar Navy forums and they suggested I connect directly to the receiver using screen. I was unsure how to terminate that session, but ended up killing the screen session (Ctrl+Alt+a followed by k to kill). There is probably a better way - feel free to comment below.:

$ screen /dev/ttyUSB0 4800

But doing so just output a load of binary garbage, which I realised (with help from Surrey LUG IRC #surrey on irc.lug.org.uk) meant that the receiver was no longer in NMEA mode. Apparently GPSD automatically reconfigures the receiver to SIRC III binary. I have no idea why this was not a problem in Ubuntu 12.04.

To fix this problem I needed to ensure that GPSD was not running. Unfortunately stopping the service does not seem to terminate the process, so I also had to kill it off:

$ sudo service gpsd stop
$ ps aux | grep gps
$ killall PolarCOM.bin

Next I placed the GPS receiver into NMEA ASCII mode, instead of SIRCIII Binary mode:

$ gpsctl -f -n /dev/ttyUSB0

Lastly, you may also need to configure the serial port, although I did not need to do this:

$ stty -F /dev/ttyUSB0 ispeed 4800

A quick check to see that we are getting ASCII NMEA sentences, showed that it was working perfectly, with easy to read text output from the receiver.

$ screen /dev/ttyUSB0 4800

To prevent GPSD from reconfiguring the receiver again, I then needed to reconfigure it to use read-only mode:

$ sudo dpkg-reconfigure gpsd

And at the point where it requests switches, I added “-b” (Broken Device Safety Mode), otherwise known as read-only mode.

With trepidation I launched PolarCOM, which immediately showed my current position.

I was a bit surprised to be asked if my prop was left or right-handed. Turns out that my boat is unusual in being left-handed, as the prop spins anti-clockwise. This means that the prop-wash will make it easier to moor starboard side-to, which is important when choosing a berth.

This Christmas my wife bought me a Kindle Fire. Originally I wanted a Nexus 7, but was tempted by a Black Monday offer of £99 for the basic Fire (normally £129). There is also a more expensive HD version, which also has Dolby sound, for £159; but in my opinion you are getting dangerously close to Nexus 7 pricing at that point.

I am a great fan of the 7 inch form factor. Personally I would be embarrassed to carry around with me a 10” tablet, but the smaller book-sized 7” tablet is fine. Without a case, it will even fit into a suit jacket pocket.

My hope for the Fire was that it would be a quality device, being sold cheap as a loss leader for Amazon, but more to the point I hoped that it would still be usable for email and web browsing and that I would be able to find at least some of my favourite Android apps in the Amazon app store.

Opening it on Christmas day I was immediately pleased with the quality of the screen, which was bright with vibrant colours. The home screen was a disappointment, being a carousel style, which includes all the books bought from Amazon. Entries include your books and apps in what I assume is date last used. This may be a good layout for books, but the Fire is not an ebook reader and I would have preferred a traditional Android home screen. Also the web browser entry reflects the last website visited, which means that you don’t know what the entry looks like, nor where it is - as the sequence is in date last used. On the other hand pressing App on the menu seems to provide you with a list of the installed apps, so perhaps that’s not a major issue.

I knew that I was taking a risk with the Amazon app store, and certainly some of the apps that I was wanted were missing, for example Connectbot and Google+. That said it is possible to install from elsewhere so this may not be an insurmountable problem, and perhaps such apps will turn up eventually. Whilst on the subject of the app store, Amazon seem to offer some normally paid apps for free, and I was delighted to be able to download World of Goo on that basis.

One pleasant surprise is the keyboard, which is of the Swype variety, wherein you keep your finger against the glass sweeping through the letters of the word. This works brilliantly and is not only quicker,but also makes typing possible when typing on the move. It also seems to be very intelligent, you can safely skip repeated letters and it generally works out what you meant. I was very surprised that typing “go” in the previous paragraph correctly entered Goo, which suggests a knowledge of context - impressive in a keyboard. If you prefer a traditional touch keyboard then it can work like that as well.

One of the many free apps was BBC iPlayer and I was easily able to navigate to a program that I wanted to watch and it worked flawlessly, with smooth video offered in full screen.

Another huge positive is the speed of wireless acquisition, it has usually already found it by the time I unlock the screen. In contrast we have another Android tablet that takes forever to acquire our hidden SSID and my non-Fire Kindle has yet to acquire it without manually adding it afresh.

But the main reason for the tablet was my email and I was delighted that the stock email app was able to cope with both my work IMAP as well as my Gmail account.

I have not yet been able to test the battery life, for the simple reason that I have yet to charge it fully, not for lack of trying, but because it takes many hours to charge. This is an issue with USB charging. When buying my wife’s 10” tablet - a Zenithink ZTPAD ZT280 C91 - I was originally disappointed at the lack of USB charging, but now I see this as an advantage - as it will charge up fully in about an hour.

There’s no doubt that the Kindle Fire represents a compromise, but I have to say that this far I am very happy with the deal, the Fire is a very polished device for very little money and what it does it does extremely well. Would I have paid £129 though? I think I would have spent some more time looking at the Nexus before making that decision, but at £99 I am very pleased.

Update: I have managed to get Connectbot working thanks to this post.

It is commonly said, at least in Microsoft circles, that Linux is free if you do not value your time. But is that really true, given the ease of installing and updating software under Linux?

Today I needed to install a Windows-only budgeting program called Profitplanner on our Windows 2003 server. This process was straightforward enough, just a ZIP download which needed extracting and installing.

Whilst the install worked fine, it warned that I needed to install the Microsoft Visual J# 2.0 Redistributable package, and gave me a link so to do.

Unfortunately the Download button on the Microsoft website would not work in Internet Explorer 8, forcing me to use Chrome instead.

Having downloaded the executable I then tried to install it, but no good - it needed .NET installed first.

The link provided did not take me to the application download page, but instead took me to the main Microsoft download page, where there were 1042 results for “.NET framework”.

The instruction was to install .NET “2.0/3.0”, which I take to mean either 2.0 or 3.0. Even filtering for Windows Server 2003 R2 results in 78 results - 5 pages - including service packs, updates, patches, SDKs etc. There seems to be know way of filtering the results further, leaving me manually searching for Application entries.

I then realised that I should have checked that I don’t already have it installed already, even though it had told me that I hadn’t. Sure enough I did indeed already have .NET Framework 4 installed. Do I really need to install 2.0 or 3.0 alongside it?

I decided to attempt a fresh installation of the .NET Framework 4, but doing so merely told me what I already knew - that it was already installed. The conclusion had to be that that I needed an older version of .NET than that currently on my system, or perhaps a newer version of Visual J#? I searched for the latter first, but 2.0 does indeed appear to be the newest version.

Going down the tack of installing an older version of .NET, a Google search suggested that 3.5 might include versions 2.0 and 3.0. I would have assumed that 4.0 is just a development of that, but who knows - at this stage it is worth a try. I downloaded and attempted to install, but then realised that I have inadvertently downloaded the web installer, which just kept restarting, possibly our proxy server getting in the way. I search again, this time for the redistributable package, and downloaded that.

I then tried to install J#, but again it insisted on 2.0 of the .NET Framework. Only then did I realise that I had installed the .NET Compact Framework.

Yes this was completely my fault, but mistakes like this are inevitable when using the Microsoft Download Centre, as you are offered a bewildering array of similarly named downloads.

I uninstalled the .NET Compact Framework, and search again, eventually finding the 3.5 redistributable listed at the bottom of the .NET Framework 3.5 web installer page. The 197mb download (there are Linux distributions smaller than that) took a while to download and even longer to install. Eventually it completed successfully.

Installing J# now worked, so clearly 3.5 worked where 4.0 did not. With J# installed I was then able to easily install Profitplanner.

Success at last.

Having gone through the above craziness, it occurred to me that .NET is a supported CrossOver Linux application, and J# is probably easily installed as well… 10 minutes later I had a CrossOver bottle with a working copy of ProfitPlanner.

Perhaps unusually it was actually simpler to install a Windows application under Linux than it was under Windows.

If you support this letter, please share and re-post.

• Open Letter to E-Book Retailers

• Christopher Roberts, 4 October 2012

I have bought myself a GPS Receiver BU-353. Having plugged in the device into my Debian Wheezy workstation, I wanted to test that it was working.

A quick dmesg | tail showed me that the device has been found and installed correctly (no drivers required).

I then installed the GPS daemon:

$ sudo apt-get install gpsd gpsd-clients
$ sudo dpkg-reconfigure gpsd

This then started the GPS daemon. The next thing to do was get some example output, and the tool for this is gpspipe:

$ gpspipe -w -n 5

Lastly, I thought it would be fun to plot the output onto Google Maps and/or Openstreetmaps:

$ tpv=$(gpspipe -w -n 5 | grep -m 1 TPV | cut -d, -f4,6-8,13)
$ latitude=$(echo $tpv | cut -d, -f3 | cut -d: -f2)
$ longitude=$(echo $tpv | cut -d, -f4 | cut -d: -f2)
$ google_map_url="http://maps.google.com/?q=${longitude},${latitude}&z=${zoom}"
$ osm_map_url="http://www.openstreetmap.org/?mlat=${latitude}&mlon=${longitude}&zoom=${zoom}&layers=M"
$ xdg-open $google_map_url
$ xdg-open $osm_map_url

All worked beautifully.

I was delighted to read this morning that Wine 1.4 stable had been released, a quick dpkg -l wine revealed that my Debian Wheezy install was running version 1.0.1, several significant versions old.  There may be good stability reasons for that decision, but I can afford to be a little more adventurous on my desktop PC.

So the question was how to build, well some brief and helpful instructions are already available.

The only ways I deviated from these instructions was to use the standard Wine option for the build:

$ cd tmp/ $ ./tools/wineinstall

Taking the option for “no” to build without installation, and once complete I installed with:

$ sudo checkinstall

Checkinstall is a wonderful program to install from source via a Deb package, that can later be uninstalled or upgraded.

Many thanks to lindqvist for saving me the hassle of working out the dependencies!

Update Wine 1.5 has now been announced!

I was scraping the ice off my car yesterday and thought to myself that an ice-scraper was just the sort of thing that you ought to be able to print using a 3D printer. Surprisingly though, I could not find a single ice-scraper on Thingiverse.

This situation could not be allowed to continue a moment longer! 10 minutes later I had modelled a basic ice-scraper using OpenSCAD; although it was certainly no beauty.

60 minutes later and I had installed the MCAD library and was using “boxes.scad” to make the ice-scraper a little more rounded.

The resulting ice-scraper will certainly not win any design awards, but it was a fun project and I now know about the MCAD library, which will open a whole new world of functionality for me.

But the exciting thing for me is that a simple object like an ice-scraper can be designed in only a few minutes and, with a 3D printer, would be a reality only a short time later.

For more information please:

• Visit it on GitHub

Having mounted my LED television on the wall I have a 80mm hole behind the TV, through which all the cables arrive. No-one can see it of course, but I know it’s there and it needs finishing, otherwise I might start muttering incoherently.

So I decided to create a cable grommet to fit into the dry-lining wall. This is still very much work-in-progress, and in particular I need to find a way to cope with the necessary overhang on each side. It also needs a removable cap, so that you can easily feed additional cables, hopefully I will get these issues addressed very soon.

It was designed in openscad and fully parametric, it could doubtless be used for other purposes, including a desk grommet.

For more information please:

• View it on GitHub

I have just created my first OpenSCAD design - Bungee Cord Toggles. I don’t actually know if it will print, but will have to wait until my reprap is complete before trying.

For more information please:

• View it on GitHub

Phase 1 of the build of my RepRap Prusa is now complete and I have photographed it thoroughly, to assist others with their build and for general interest.

View Album on PicasaWeb

I have even taken a brief video, apologies for the lack of lighting:

The instructions I followed were the Prusa Mendel Assembly instructions, with modifications described in the Thames Valley Google Groups thread Frame bearing specifics?.

This Prusa is 16mm wider than standard, in order to allow for the heated bed. For further details on the design being followed, please visit:

http://tvrrug.org.uk/

Looking forward to Phase 2 of the build!

I have been following the RepRap project for some time, but it always seemed to me to require hardware hacking skills that were that little bit beyond me.

At Oggcamp 2011 I watched a presentation by Malcolm Napier of empirical on the Reprap project.  Whilst this sparked my interest anew, I was put off by Malcolm’s mention of builds taking months rather than days; as before this seemed to be a project for hardware hackers.

A few weeks later I received an email from Hugo Mills asking me to post an invitation to a new Reprap build event to the Surrey Linux User Group mailing list. The build event was being organised by Malcolm, and would help users to build their own Reprap in a supported way. This sounded perfect for me so, having forwarded the email, I booked a place on the build event.

To cut an unnecessarily long story short - I have now ordered and paid for my reprap parts and am now eagerly awaiting their arrival. There are 20 of us in the first build project and we are hoping to receive everything except the electronics by the end of November.  The expectation is that we will complete the build in January.

If you would like to read more, and perhaps consider participating in a second build project, then visit the very new TVRRUG website, and perhaps start off by reading the First Steps page.

I will try and blog about the experience here, in the meantime we have a fairly active IRC Channel #tvrrug on Freenode if you would like to pop in for a chat.

Before you start, you need to have sane installed, and you also need tesseract-ocr - both should be available in your distros repositories.

$ sudo apt-get install sane-utils tesseract-ocr

Next you need to find out what scanners you have available, and you do this with:

$ scanimage -L
device `v4l:/dev/video0' is a Noname Vimicro USB Camera (Altair) virtual device
device `plustek:libusb:004:002' is a Epson Perfection 1250/Photo flatbed scanner)

Obviously the latter is my scanner.

Assuming you have a working scanner, the following is a simple two liner to scan and OCR.

$ scanimage -d 'plustek:libusb:004:002' --mode Lineart \
--format tiff -x 215 -y 297 --resolution 200 > example.tif

And finally convert to text with tesseract:

$ tesseract /tmp/example.tif example

You should now have a file example.txt in your current directory, which you can open in any text editor.

Obviously this has limitations - it works for single-page A4 portrait typed documents - but it gives you the basics.

You could probably experiment with the resolution, 200 worked for me, so I didn’t bother trying anything else.  Traditionally the higher the resolution the better, but I seem to recall that tesseract works better on 300 and below.

On my Epson Perfection 1250 I found that I needed to add the sane switch --warmup-time 0 as otherwise it never finished warming up.

If you would prefer to OCR an existing PDF, which is another thing that I find myself doing from time to time, then first convert it to a tif:

$ convert -density 200 example.pdf -depth 8 /tmp/example.tif

And then run the above tesseract command.

It has been well known that the Android 3.x is not open source yet, but I had expected 3.x to be open sourced at any moment. Now that 4.x has been released and there is still no sign of the 3.x or 4.x source code, I think we have to ask - is Android really open source now?

Well Google has given some reassurances that 4.x will be open sourced “We plan to release the source for the recently-announced Ice Cream Sandwich soon, once it’s available on devices”, but this lacks credibility when they still haven’t released the code for 3.x.

This could all change in moment, all it takes is for Google to publish the source code, but this is truly not how an open source project is supposed to be run.

2011-11-03: Further to this article, Google has announced that Android 4.0 is to be open sourced in the coming weeks.

For more information please visit:

• OpenMediaVault 0.2 (Ix) has been released!

http://chrisjrob.com/

The old chrisjrob.wordpress.com will continue to work, but will redirect to chrisjrob.com. At least until the day I decide to leave Wordpress whereon I will disappear into oblivion.

Actually, I am very happy with Wordpress.com, so am unlikely to be leaving any time soon.

Until now.

In their infinite wisdom ICANN has enabled custom TLDs from $185,000 plus $25,000 per annum.

Needless to say, this limits their availability only to large companies. No longer is there a level playing field - large companies can pay for a TLD that is simply not available to smaller companies, organisations or individuals.

Does this matter? Well I don’t know for certain, but this is the first time that part of the Internet has been closed off to the rest of us, and it just doesn’t feel right to me.

An interesting blog post about ownCloud. ownCloud is a webdav based application allowing you to store files on your own personal server on the Internet and access them via http or webdav.

For more information please:

• Read more on BSDNexus
• Visit ownCloud.org

We all know that feeling when our disk fills up and you are left desperately scrabbling around to find out where your disk space has gone. In a previous blog post I discussed the use of the wonderful Konqueror File Size View, but this is no good for remote servers. Normally I would resort to “du” or the wonderful “find” utility to look for large files, but here is an interesting alternative that I had not come across before: ncdu (ncurses disk usage).

Its name tells you pretty much everything you need to know. It can be installed with a simple apt-get install ncdu and then the man page is a useful guide. In simple terms it can just be run with:

# ncdu /var

The lovely thing about ncdu is that once it completes its run (which can take a long time on a large disk or a nfs share), you can drill into the directory structure following the disk usage to determine where your space has gone.

It is a very simple program but one that I will find most useful.

The more that I work from the command line, the more of a nuisance I find it that all my passwords are in KeePassX - an excellent GUI application.

For some time I have intended moving to pwsafe, but am put off by the time it will take to migrate the data. An alternative is to export the KeePassX data into an encrypted text file and use the gnupg.vim Vim plugin to read and edit it. This post describes how to accomplish this in a very few steps.

I first ran KeePassX and deleted the backup group (as I did not wish to export all my previous generations of passwords). I then exported the contents to a text file named “passwords” (you could be more imaginative/secure if you wish!). Then I followed these instructions to create a GPG secured file:

$ gpg -c --cipher-algo AES256 passwords

And enter your secure passphrase twice. You should now have a new password file “passwords.gpg”. You can now remove the plain text version:

$ shred --remove passwords

In order to edit/view the new passwords.gpg file, we need the Vim gnupg.vim plugin, which is available in the debian vim-scripts package:

$ sudo apt-get install vim-scripts

Then edit /etc/vim/vimrc.local to add the following line:

set runtimepath+=/usr/share/vim-scripts/

This will import all of the vim-scripts plugins, which may not be what you want. Alternatively you could probably just copy or perhaps symlink the gnupg.vim plugin to your ~/.vim/plugin/ folder.

And that is it - now whenever you wish to view or edit your passwords you simply type:

$ vim passwords.gpg

This will prompt you for your password and open the file.

Initially I found this gave errors in relation to exuberant ctags, which I suspect is related to vim-scripts and not to gnupg.vim and is doubtless a feature of the way I have just pulled in all of the vim-scripts. One option (and what I did) is just to install exuberant-ctags as follows:

$ sudo apt-get install exuberant-ctags

Now I just have to format passwords.gpg into something a bit more structured, perhaps with Vim folding configured.

And if you are looking for a way to create passwords, the pwgen is one option - it offers a selection of fairly memorable passwords from which to choose.

$ sudo apt-get install pwgen

There are some concerns with this approach - anyone can read your passwords over your shoulder - if you work with others you might be better off with something like pwsafe. Also some users are concerned that the contents of the file might be written to memory, although I believe this is not the case, you should satisfy yourself as to the security of this approach, as I am certainly not qualified to advise you on this aspect.

I finally decided to upgrade my work desktop to Debian Squeeze:

Preparing for the Upgrade

I manually removed all but the official lenny repositories (and removed the Debian Volatile repository which is no longer used), and updated them to “squeeze”:

$ sudo vim /etc/apt/sources.list
deb http://91.121.125.139/debian/ squeeze main non-free contrib
deb-src http://91.121.125.139/debian/ squeeze main non-free contrib

deb http://security.debian.org/ squeeze/updates main non-free contrib
deb-src http://security.debian.org/ squeeze/updates main non-free contrib

Next, in order to minimise download time, I reviewed my installed packages:

$ sudo dpkg -l | grep "ii " | less

And I removed all the unnecessary packages. In particular I decided to remove all of KDE, as I am now using pekwm. When removing packages I found apt-get kept trying to upgrade large numbers of packages. Aptitude was much better, as instead of just trying to install a load of packages, it reported the problem (e.g. package XYZ will be broken by this change) and let me choose the solution. I just added the packages that would be broken to the list of packages to be removed.

Upgrade apt and dpkg

Then I installed apt and dpkg, to save these from having to be upgraded during the upgrade process:

$ sudo apt-get update
$ sudo apt-get install apt dpkg

Upgrade to Squeeze

Lastly, I ran the upgrade to squeeze with:

$ sudo apt-get dist-upgrade

During the upgrade I receive warnings about /lib/firmware/e100 missing. Mindful of the Debian’s much touted removal of non-free drivers, I did an apt-file search e100 which revealed that I needed to install the firmware-linux-nonfree package. This done, I rebooted.

First Boot

It booted fine to the command prompt, but of course without a login manager it got no further. I simply typed startx to get into pekwm and it all worked perfectly.

Installing KDE

I decided to install KDE again, although I would struggle to explain why, curiosity perhaps. I installed kde-core and it all just worked perfectly, with a neat first time wizard to handle the migration of configuration from KDE3.5 to KDE4, excellent work.

I used KDE4 for a few hours before returning to pekwm.

Conclusions

So far Squeeze has been flawless and yet another Debian dist-upgrade without a hitch. Well done Debian.

KDE looks great and seemed to work exactly as it ought. Over the few hours I encountered no problems whatsoever. But I do wish KDE would learn from some of the features of PekWM - I just love the left-mouse-click on the screen edge to move to the next window, I love the ability to group windows and I love the fact that you can resize windows to fill the available space.

I am not saying that all those features should be made the default in KDE, such matters are subjective, but they ought to at least be possible in KDE and as far as I can tell they are not. There is a screen edge option, but it works poorly, relying on you holding the mouse at the screen edge for a definable period of time.

Step 1: Sign up for Dynamic DNS

The first problem is that most people’s Internet service do not have a static IP address. In practice it is fairly static on modern broadband services, but, if you do not want to be fiddling about trying to find out what the current IP address is, then you need a dynamic DNS service.

Dynamic DNS gives you an unchanging address which is automatically mapped to the current IP address, whatever that might be. So for example you might have daddy.dyndns.xyz as your address. To make this work, you need to have a client machine on the network, that will keep updating daddy.dyndns.xyz with the latest IP address. Some home routers have this functionality built in.

I did some research and determined that the very popular http://dyndns.com service is reliable, and so I went ahead and signed up a new account for my father, and set-up a hostname for his router’s WAN address.

Step 2: Configure Client

If you’re lucky your router will support dynamic DNS. My father’s did, so I just logged onto the router and entered the dyndns account information, and it just worked perfectly.

If your router does not support dynamic DNS, all is not lost, simply install ddclient on a computer on the network (presumably, but not necessarily, the one that you want to be supporting):

$ sudo apt-get install ddclient`

Configuration of ddclient is simple, the ncurses configuration process takes you step-by-step through the necessary steps.

Step 3: Configure port forwarding

The next problem is that your router has a firewall that is designed explicitly to stop people getting into your network. We need to punch a small hole in that firewall. Needless to say this does have security implications, so we need to be careful.

So we need to open a port on the router and forward it to the machine that we wish to support. This machine does need to have a static IP address.

Log onto your router via its web management page. The main difficulty here is in finding the port forwarding option, when it is probably called something user friendly like Game Sharing or some other perversely unhelpful name. If you can’t find it, then you’re best of Googling for “port forwarding routername” where routername is the make and model of your router. Or just read the manual, if by some miracle you have one to hand.

But which port to forward? Your router will probably make it easy to forward the standard ssh port - port 22 - the problem is that everyone knows that is the ssh port and you may experience a high number of attacks on that port. In theory you should be safe enough with secure passwords, but personally I would not choose to forward the standard ssh port, but would instead add a custom service “ssh_obscure” on a different TCP port number and forward that to port 22 on PC that you wish to support:

Name: ssh_obscure
Protocol: TCP
Source port: 2121 (or whatever you choose)
Destination IP: (enter the client PC's IP)
Destination Port: 22

If it won’t let you specify a Destination Port, then it will forward to the same custom port on the PC. This will then require a change in the configuration of sshd on the PC, which I will explain in the next step.

Step 4: Install openssh-server

On the client PC, i.e. the PC you will be supporting, install openssh-server and its dependencies:

$ sudo apt-get install openssh-server

If you didn’t bother with the ssh_obscure suggestion above, then you are simply forwarding from port 22 on the router to port 22 on the client and no further configuration of openssh is required.

If you did set-up the ssh_obscure, but forwarding to port 22 on the client, then again no further configuration is required.

If you set up ssh_obscure, but were unable to set a Destination Port, then you are forwarding the same port to the client, so you will need to change the relevant setting in /etc/ssh/sshd_config. E.g. Port 2121. Then reload the ssh configuration with:

$ sudo /etc/init.d/ssh force-reload

Step 5: Install Fail2ban

As stated above, we have reduced the security of your network by taking these steps and we are now going to attempt to rectify this somewhat by installing fail2ban.

Some argue that this is unnecessary, if you have set a custom port for ssh and maintain secure passwords, but I see no reason for relying on security by obscurity. The main benefit of using a custom ssh port is to prevent the attempts in the first place, but hiding your front door is no substitute for locking it.

Fail2ban will monitor your /var/log/auth.log and block any hosts that have repeatedly failed to login correctly, by blocking them in iptables (the standard Linux firewall).

$ sudo apt-get install fail2ban

Next you should create a new file /etc/fail2ban/jail.local, with the following contents:\

[ssh]
enabled = true
port = ssh
filter = sshd
logpath = /var/log/auth.log
maxretry = 3
bantime = 600

Strictly speaking the above is already configured in /etc/fail2ban/jail.conf, but I think it makes sense to take ownership of those settings by redefining them in jail.local. If you have set a custom ssh port, then set it in jail.local as above, replacing port = ssh with port = 2121 or whatever. I have reduced maxretry to 3 from the default of 6, as I found it was giving nearer to 18 attempts, thanks to the peculiarities of the auth.log file.

Finally reload these new settings in fail2ban:

$ sudo /etc/init.d/fail2ban force-reload

In order to test fail2ban, try and log into the client from another machine on the same network. You will need attempt to log in with the wrong password a number of times. This may not be exactly 3 times, but if you get to 9 attempts and it still has not blocked you, then there is problem.

fail2ban will only block you for 10 minutes, after 3 failed login attempts (you can change these settings as above).

Step 6: Testing SSH

At this stage you should be able to log in remotely to your client PC.

$ ssh -p 2121 username@daddy.dyndns.xyz

Obviously change 2121 for your custom port, or leave out -p 2121 if you are using the standard ssh port 22.

Step 7: Installing VNC

Having ssh access is wonderful, but not much help when your Dad asks you where his OpenOffice toolbar has gone; so we need to add VNC on the PC you wish to support:

$ sudo apt-get install x11vnc

And on the PC from which you will be providing support you need to install vnc-viewer, or use your preferred vncviewer:

$ sudo apt-get install xvnc4viewer

Step 8: Testing VNC

Now, you could have set up x11vnc to be running permanently with a password, and then you could set your router to port forward a custom VNC port to the client PC. This approach does significantly degrade your security and is in any case unnecessary. Instead we can just run it when we need it by using ssh, by typing this command from the PC from which you will be providing support and changing as appropriate:

$ ssh -p 2121 -t -L 5900:localhost:5900 username@daddy.dyndns.xyz 'x11vnc -noxdamage -localhost -display :0'

This will connect by ssh to the client machine, run x11vnc bound only to localhost (so no-one else can connect to it) and build a tunnel between the remote VNC port to your local VNC port.

One caveat is that the username needs to be the currently logged in user, otherwise x11vnc will refuse to start.

And finally on the PC from which you are providing support we just run:

vncviewer localhost:0

And you should now be looking at your father’s (or whoever’s) PC.

Whilst this does all sound pretty involved, all but this final step is only required once, then a quick ssh command and vncviewer and you are connected.

apt-get

For apt-get you just need to create a new file:

/etc/apt/apt.conf.d/76download

with the following contents:

$ sudo vim /etc/apt/apt.conf.d/76download 
Acquire {
    Queue-mode "access";
    http {
        Dl-Limit "128";
    };
};
$

The above will limit your bandwidth to 128K, adjust this figure to suit your network.

Alternatively, if you don’t want this change to be set permanently, then you can specify it in the command line:

$ sudo apt-get -o Acquire::http::Dl-Limit=128 upgrade

wget

To rate-limit wget, simply edit /etc/wgetrc or your personal configuration at ~/.wgetrc and add or edit the following line:

limit-rate=128k

other

Other packages can be configured in different ways, but you could install trickle and then read its man page to determine how to use it. For example (from the man page):

$ trickle -u 128 -d 128 ncftp

To set this up, you need to install pdftotext. This application is available in either poppler-utils or xpdf-utils - most people seem to use the latter, but I suspect the former is probably a better option for a text-only server. I installed poppler-utils.

Next create (or edit) the ~/.mailcap file, so that it has the following line:

application/pdf;                pdftotext -layout %s -; copiousoutput

You could also add the line:

text/html;                      elinks %s; nametemplate=%s.html`

The latter line opens HTML attachments in elinks.

Then just hit “V” when viewing an email, and you will see the attachments, select the attachment and hit Enter to view it. Awesome :)

My folder layout is that I have a Spam folder, within which I have two folders - Spam/Ham and Spam/Verified. The Spam/Ham are good emails, particularly those that have been wrongly identified as Spam. The Spam/Verified are the Spam that I am confirming are definitely Spam. SpamAssassin places all the detected Spam in the Spam folder. I generally move all the Spam folder into Spam/Verified except the one or two false positives, which I move to Spam/Ham. Then overnight SpamAssassin learns from those two folders - a fairly standard SpamAssassin layout I think.

Occasionally we get Spam in our Inbox, and obviously these can be moved straight to the Spam/Verified folder.

The important factor here is that I want to be able to easily flag individual emails in my Inbox as Spam, but I also want to be able to tag the whole of my Spam folder and flag as Spam.

The manual gives a list of functions, which I used rather than the normal key bindings, as the manual advises that this is more portable and robust. I used the tag-prefix command to enable tagging. I was not using the keys H nor S, so I chose these as Ham and Spam respectively.

You need two macros for each function - one for use in the folder list (index) and one for use inside an email message (pager), so that you can flag it as Spam without moving back to the message list.

The biggest problem I encountered was the difficulty that removing the unread flag resulted in the selection moving down to the next email, and then the next email would be flagged as Spam - frustrating. The only solution seemed to be modify the muttrc configuration before and after, and almost unbelievably that is in fact possible using the function.

The final macros are shown below:

# Move message(s) to Spam/Verified by pressing "S"
macro index S "<tag-prefix><enter-command>unset resolve<enter><tag-prefix><clear-flag>N<tag-prefix><enter-command>set resolve<enter><tag-prefix><save-message>=INBOX/Spam/Verified<enter>" "file as Spam"
macro pager S "<save-message>=INBOX/Spam/Verified<enter>" "file as Spam"
# Move message(s) to Spam/Ham by pressing "H"
macro index H "<tag-prefix><enter-command>unset resolve<enter><tag-prefix><clear-flag>N<tag-prefix><enter-command>set resolve<enter><tag-prefix><save-message>=INBOX/Spam/Ham<enter>" "file as Ham"
macro pager H "<save-message>=INBOX/Spam/Ham<enter>" "file as Ham"
# Return to Inbox by pressing "."
macro index . "<change-folder>=INBOX<enter>" "Inbox"

Suddenly Mutt is the easiest way for me to handle by Spam.

I confess, I like XKCD. Even though some of them are unfunny and others beyond my understanding, sometimes, just sometimes, they strike a chord.

Once I moved to pekwm I found myself with a black plain desktop, what could be better, I thought, than the current XKCD comic in the centre, with the tooltip comment located underneath in an appropriate font.

After some trial and error, I ended up writing this script. It is far from perfect, and will probably end up being broken at some point, although it survived its first trial unscathed. I’m a little embarrassed by my tortuous curl/perl one-liner, ideally I would re-write this whole thing in perl, but it has worked reliably for a couple of months, so I dare say that I will never get around to it.

Currently it works for pekwm, and I dare say it will work for other lightweight window managers, like openbox, that can use feh to set the desktop wallpaper. If not, then it shouldn’t be too difficult to modify the script for your window manager or desktop environment. You can also run the script without the -w switch in order to just create the XKCD image in the directory for your own purposes.

You will obviously need perl, curl, wget, and fold, if they aren’t already installed on your system. Create a directory, say ~/.xkcd and save this script in there, along with the XKCD font, mark the script as executable. If you read the script, it is fairly self-explanatory I think.

You basically run it like this:

/path/to/.xkcd/xkcd -w -d '/path/to/.xkcd'

For pekwm, just add a line at the end of the ~/.pekwm/start file:

(sleep 30 && /path/to/.xkcd/xkcd -w -d '/path/to/.xkcd') &

I hope it works for you, feel free to comment any suggestions or criticisms. If you do start using it, it would be fantastic if you could add a comment to let me know.

So we took the drive out of the NAS and I used a USB to SATA cable to mount the drive on my Debian Squeeze laptop. All worked just fine, I could browse his music and pictures, but there wasn’t much I could do to recover his 200GB of music and pictures; I hadn’t brought an external USB drive with me (note to self: buy a second cable), and we were unable to share a Windows directory with my laptop, for the simple reason that my laptop did not have samba installed, and his Vodafone 3G broadband was not functioning. Yes, it was one of those days.

Simple enough, we thought, we would just plug the USB to SATA cable into his Windows 7 laptop instead. Unfortunately it seems that, despite working out of the box on Linux, it requires a driver installed under Windows 7. A driver which I did not have with me, and of course our 3G connection was still down at this point; so I could not download it (even if I could track it down).

Fortunately we had a spare SATA desktop PC, with a spare place for a hard drive. He did not wish to use Linux, so installed Windows 7. We then plugged in the SATA drive, enabled it in the BIOS, and rebooted. The drive and its partitions were detected, but not assigned a drive letter. We went into drive management, but the option to assign a drive letter was disabled. Clearly there was something adrift with the drive that was preventing it from being mounted in Windows 7.

What I would then have done was use a Linux live CD to boot onto this SATA PC, mount both drives and simply copy the data across, but time had beaten us - they needed to go to his parents, and we had to leave it at that. He decided to take the drive into work and get one of his technical people to recover his data onto a fresh drive.

What a frustrating few hours, and it would have been trivial to fix on a Linux network, firstly because the drive would have just mounted fine (which it did on my laptop), secondly because I could have trivially created an NFS share (okay unfair - as it wasn’t Windows fault that I didn’t have the samba installed), or thirdly I could just have used dd with ssh to transfer the drive image to another computer on the network.

Working on a Windows network is like having one arm tied behind your back. Working on a Windows network without functioning broadband is like having one arm tied behind your back whilst wearing a blindfold.

I need a beer.

This doesn’t give you the entire picture, as it only lists the dependencies that you don’t already have; which is usually all you care about, but there are occasions when you would like to list all of a package’s dependencies, for example when planning for a system that is not built yet, or not accessible at the current time. Or just for idle curiosity! Perhaps that’s just me.

Anyhow, apt-rdepends is the application for the job. It doesn’t just list the package’s dependencies, but it recursively goes through each dependency’s dependencies.

Install it with the usual $ sudo apt-get install apt-rdepends and then simply run with:

$ apt-rdepends packagename | less

Yes, it is quite verbose, hence the “| less” - leave it out if you prefer, or use “| more” which is more likely to be installed on your system (tip: install “less” - less is better than more, if that makes any sense).

For example, I had just installed “flite” and was amazed at how functional it was. I wondered to myself whether it was just a front-end to festival - but how to find out?

$ apt-rdepends flite

Which comes back with no other speech synthesis engine (e.g. festival), so clearly flite is a speech synthesis engine in its own right.

What I have decided is to cut down on the number of categories and treat these as “feeds”; so I have a technology feed, a Linux feed, a Debian feed, and only a couple of others. These are the feeds to which I am committing to maintain, so if you are only interested in my posts on Linux, you can feel confident about changing your subscription to my Linux category.

The remaining categories are:

• http://chrisjrob.com/category/personal/
• http://chrisjrob.com/category/technology/
• http://chrisjrob.com/category/technology/linux/
• http://chrisjrob.com/category/technology/linux/debian/
• http://chrisjrob.com/category/technology/linux/ltsp/

As you can see from the above layout, my technology feed will include Linux and Debian.

For the RSS feeds, just append “feed/” to the end of the above categories, or visit the above pages and subscribe to the feeds via your browser feed functionality.

This means that you can subscribe to the element of my blog that you find interesting, and save yourself the grief of reading about my dog waking me up at 5.00am vomiting. Or, if you are one of my non-geek friends, you could subscribe to my personal feed and avoid all my weird ramblings about the leenucks.

If you want a more fine-grained approach, then you can of course use the tags in the same way, but the difference is that I am not committing to maintain tags, the tags are by their nature more anarchic. But nevertheless if you wanted to watch my posts on the dreamplug (random example) you would just visit:

• Web: http://chrisjrob.com/tag/dreamplug/
• Feed: http://chrisjrob.com/tag/dreamplug/feed/

Just don’t be surprised if there are no further posts on dreamplug. There probably won’t be.

Be warned, part of the reason for this change is that I intend starting a more personal blog, which could be exceedingly boring for some of you; just subscribe to the technology or Linux feeds instead.

• Virtualization With KVM On A Debian Squeeze Server

I have used KVM, but never with libvirt.  I understand that libvirt is a standard api for managing various virtualisation technologies, e.g. KVM, QEMU, Xen and others, giving you a common way of managing your guests regardless of the technology in use.

For more information please visit:

• Virtualization With KVM On A Debian Squeeze Server

Mike sorted this out by setting up IP forwarding on his laptop, so that we had a wired network which used his wireless connection as a gateway. I knew this was technically possible, but I had not actually seen it done before.

So this morning I sought out suggestions from the Surrey LUG Mailing List, and the following was the end result.

Please note that this solution is as-yet untested by me and involves changing your existing iptables rules - albeit only until you reboot.

Assuming eth0 is your LAN port and wlan0 is your wireless WAN port, and that you have connected your client PC to your host PC with a crossover ethernet cable (or perhaps via a dedicated switch), then the following instructions should work:

On the Laptop

On the laptop, which is assumed to have a working wireless connection, and a wired connection to the Client PC:

Firstly, you will probably need to set some static IP details:

# ifdown eth0
# ifconfig eth0 192.168.10.1 netmask 255.255.255.0

Next we need to activate IP forwarding, and setup iptables to NAT:

# sysctl -w net.ipv4.ip_forward=1
# iptables -t nat -A POSTROUTING -o wlan0 -j MASQUERADE

The following may be worth trying if the client gets Host Prohibited responses:\

# iptables -F FORWARD 
# iptables -A FORWARD -j ACCEPT 
# iptables -nvL

DNS and DHCP

If you plan to do this regularly, then you might consider using dnsmasq to provide a DNS nameserver and DHCP to your client PC. If you do this, then you should not complete the next section entitled “On the Client PC”, which should just pick up its details using DHCP. This method is particularly handy with a Debian net install.

# apt-get install dnsmasq 
# vim /etc/dnsmasq.conf interface=eth0 dhcp-range=192.168.10.10,192.168.10.19,4h 
# /etc/init.d/dnsmasq restart

On the Client PC

Do not carry out this section, if you configured dnsmasq above.

On the client PC, which is assumed to have a wired connection to the laptop, all you need do is configure static IP networking and add a working nameserver to /etc/resolv.conf (normally populated via dhcp-client):\

$ sudo -i
# ifdown eth0
# ifconfig eth0 192.168.10.10 netmask 255.255.255.0 gateway 192.168.10.1
# echo "nameserver 8.8.8.8" > /etc/resolv.conf

Finishing up

When you’re finished, you should comment out the lines in /etc/dnsmasq.conf and reboot, and you should be back to normal again.

I am really looking forward to testing this set-up; so in the meantime if you get it working, please do let me know. Any suggestions for improvement - please do feel free to comment.

The thin clients were IGEL thin clients with 1GB RAM and 1GB transcend flash card. The RAM is plenty for Linux, so the main challenge was keeping the installation under 1GB, when modern Linux distros are probably 6gb or so.

I decided to try installing a minimal version of Debian, running just the PekWM Window Manager and not much else. This would me the advantage of the enormous Debian repositories, with all the flexibility that this provides.

Please note that this is not a howto, in fact I have typed it up a day later from memory, there are bound to be missing steps in the process, but hopefully enough to give you the jist of what I did.

The Installation

As usual with thin clients there was no optical drive, I tried using my fabulous IODD external hard drive to serve up a Debian net install image, but it did not seem to work, not sure why as the BIOS had a USB-CDROM option. So I had to mount the IODD on my laptop and transfer the image over to a memory key. Fortunately Debian has a trick up its sleeve - you can now dd ISOs straight onto a memory stick - how fantastic is that?! The command was something like this (note that you write to the drive not to a partition - e.g. /dev/sdb not /dev/sdb1):

$ sudo dd if=/mnt/_iso/debian-squeeze-netinst.iso of=/dev/sdX bs=1M

Having set the BIOS to boot first from USB-HDD and next from USB-ZIP (I can never remember which), I rebooted. Success, I was presented with the Debian installer, taking the standard installer (non-graphical), I set the language and keyboard settings and I choose to manually partition the drive. This was because the guided partitioning reserved 90mb for swap space, rightly or wrongly I thought this was pretty pointless, and reduced this to nil. It has to be said that I did have some memory problems later, and some swap would be advisable.

Towards the end of the Debian installation, it prompted me to select what type of system I wanted to install, I unchecked all the options, in order to leave me with a bare Debian system.

The installation finished, I rebooted and removed the USB key.

The Graphical Environment

Once booted I was faced with a text login prompt. Amazing, a full Debian system at my finger tips. But no-one was going to be very impressed with that - I needed a graphical environment. A quick “df” showed me that I had used 500mb of disk space, so I was already very tight for space.

I installed xserver-xorg and pekwm:

$ sudo apt-get install xserver-xorg pekwm

But when I tried to run “startx” to enter the graphical environment the command was not found. For this sort of problem you need “apt-file”. apt-file is a wonderful program that enables you to find which package you need for a known command (or other file).

$ sudo apt-get install apt-file
$ sudo apt-file update 
$ apt-file search bin/startx
xfce4-utils: /usr/bin/startxfce4 xinit: /usr/bin/startx`

In this instance I wish to know which package includes the file bin/startx. I could just search for “startx”, but I will get some pointless hits, most programs are located in a directory called “bin”, so searching for bin/startx will only match bin commands. The output tells me that the package that I am missing is xinit, so a quick apt-get:

$ sudo apt-get install xinit

Now I can run pekwm with startx.

Configuring pekwm

Once into pekwm I was faced with a black screen with a cursor - don’t panic - that is pekwm in its default state. Right-mouse-click and you will have the pekwm menu. Hmm, Terminal does not work and there are no applications listed in the menu. Returning to the terminal with Ctrl+Alt+F1, I typed Ctrl+c to terminate pekwm and installed xterm ($ sudo apt-get install xterm). Running startx again and this time selecting Terminal from the pekwm menu worked and I could continue my work from with pekwm.

Debian Menus

To get applications in the menu, we could edit the ~/.pekwm/menu file and add the relevant menus. A better solution is to get the standard Debian menus working. Firstly you should check that they are not already working for you, if you have an “Applications” item withing you pekwm menu, and if it is populated with the Debian applications list, then it is already working. If not, edit the ~/.pekwm/menu file and look for a line that looks like this:

INCLUDE = "debian-menu"

And edit it to read:

INCLUDE = "/etc/pekwm/debian-menu"

Now select PekWM from the menu and Reload and you should now have a properly populated Applications menu.

WWW Browser

Whilst there, add a Web Browser item, for example - see below:

RootMenu = "Pekwm" {
    Entry = "Terminal" { Actions = "Exec $TERM &" }
    Entry = "Web Browser" { Actions = "Exec x-www-browser &" }

This just tells it to run whatever the default X web browser is, which we can set later.

Now in a terminal window you should install your preferred browser. I ended up installing iceweasel, but there may be a better more lightweight browser that requires fewer dependencies. Feel free to comment with suggestions.

Once installed, the menu option Web Browser that we created earlier should be working. If not, then just type:

$ sudo update-alternatives --config x-www-browser

And select iceweasel (or your preferred browser).

Set desktop background

Now to make the desktop look a little prettier. For my experiment I downloaded a desktop wallpaper that I created a while ago. I installed feh ($ sudo apt-get install feh), and added the following to the ~/.pekwm/start file:

feh --bg-center your-wallpaper.jpg &

Install tint2 panel

Most people expect a panel, although strictly speaking they are not required. For this experiment though I think we should have one:

$ sudo apt-get install tint2

And then add the following to the ~/.pekwm/start file:

tint2 &

Conclusions

Well, I think that was about it. People seemed to be impressed that I had a complete Debian with a GUI environment. Though it has to be said that I was down to about 250mb of free space by the end, not even enough to install OpenOffice.

I forgot all about sound, and had to install alsa ($ sudo apt-get install alsa). As an experiment I installed Flash, and tried watching You Tube. Not good - without sound You Tube worked fine, but with sound the picture disappeared. I had no time to fix this before the end of the session.

I really enjoyed this challenge, but what about the alternatives? Mike installed Slitaz, which looked gorgeous, and used well under 200mb in total. I have been impressed with Slitaz before, and might be a good option for a project like this.

Thanks to Mike for organising the challenge.

To fix this I entered “about:plugins” in the address bar, and clicked “disable” against Flash. This results in Google Chrome falling back on the version of Flash installed in the operating system. If you do not have Flash installed, then in Debian you can install as follows:

Firstly ensure that the repositories in /etc/apt/sources.list include “non-free”, for example:

deb http://ftp.uk.debian.org/debian squeeze main contrib non-free

Then install Adobe Flash with:

$ sudo apt-get update $ sudo apt-get install flashplugin-nonfree

And you should now have a working Flash installation. I have tested You Tube and all seems to be well, but I will update this page if I find any issues.

By the way, I would obviously prefer to run Chromium, the open source version of Chrome, but at the moment the version in Squeeze is outdated, and there are no backports yet for Wheezy.

Window grouping: I love the fact that you can middle click on a window’s titlebar and drop it onto another window and the two windows join together - with the titlebar split in segments. You can switch between the applications by clicking on the relevant part of the titlebar, or use the Windows key with TAB - so just as Alt+Tab switches between windows, Windows+Tab switches between the applications grouped inside the current window.

Double-click to fill available space: I love the fact that you can open an application and just double-click on its titlebar to fill the available space. This feels so intuitive and just so damned right, that I struggle to understand why it is not the standard behaviour in all window managers and desktop environments.

Clicking on the desktop edge moves you to the next workspace: I’ve never really got on with multiple desktops/workspaces - it is difficult enough to keep track of multiple applications inside a single workspace, but pekwm has converted me. I find it just so intuitive that, by clicking on the edge of the desktop, that it it progresses you to the next desktop.

So what do you love about pekwm?

I have just been recommended to try the “tint2” panel. It looks lovely and minimalist, and has happily taken the kwalletmanager system tray icon; so hopefully will work on my laptop with battery and wireless monitors.

Whilst you’re looking at my screenshot, note the Window grouping between Google Chrome and Iceweasel and the two terminal windows grouped together. The window grouping feature really is very, very handy.

Installing tint2 is simple:

$ sudo apt-get install tint2

And then add tint2 to your ~/.pekwm/startup file:

tint2 &

If you need to configure it, then edit ~/.config/tint2/tint2rc; so far I have found the defaults to be okay.

Update: Battery monitor

To get the built in tint2 battery monitor, simply edit

~/.config/tint2/tint2rc:
battery = 1 battery_hide = never

Hunting the ‘net for anything to do with pekwm is a frustrating business. You can’t help thinking that there are only a handful of us that use it. I may post something about why I think pekwm is great, but that is a topic for another day.

As my laptop’s primary desktop environment is KDE, I wanted to use the KDE power management with pekwm, rather than install the Gnome Power Manager, which would bring in a load of Gnome dependencies. If you already have Gnome installed, then I suspect the Gnome Power Manager is for you.

There seem to be three parts to this:

1. Getting the power management daemon to run
2. Adding a power meter
3. Adding the configuration tool to the menu

Whilst these instructions are primarily for pekwm, I am sure that openbox, fluxbox and other *box window managers could use them with a little adjustment.

Power Management Daemon

In the ~/.pekwm/start file, add the lines:

kded4 & qdbus org.kde.kded /kded loadModule powerdevil &

If you do not want to restart pekwm to test, just run those two commands live.

Battery Meter

Update: Since writing this post I have discovered tint2 which includes a simple battery monitor.

What I really want is a battery level meter that will sit happily in Docker, but I failed in this quest. If you have any suggestions, please do let me know and I will update this post. In the meantime I have installed xbattbar, which gives a full thin width line at the bottom of your screen.

$ sudo apt-get install xbattbar

In the ~/.pekwm/start file, add the line:

xbattbar &

Power Management Configuration

It is quite possible that you will wish to configure your power management from within pekwm. To do this, add the following to your ~/.pekwm/menu file:

Entry = "Configure Powerdevil" { Actions = "Exec kcmshell4 powerdevilconfig & " }

And that appears to be a working solution for me, I hope it works for you too!

Apparently the ZX81 is 30 years’ old, wow.   The wedge-shaped box that started my interest in computers, and led inexorably to my love of Linux and tinkering with the command-line.

Popey clearly has a better memory than I, but I do remember rushing up to my bedroom with ZX81 in hand, plugging it all in and then at the prompt typing “hello”. I was very disappointed that the computer seemed to have no ability to talk back! But once I got passed that disappointment, I fell in love with the thing. I didn’t do much programming on it, but I used to get the computer magazines and type in the assembly code in order to play games. There was an awesome dinosaur game I remember.

I was fortunate enough to be able to have the ZX Spectrum when it came out and was blown away by how much better it was than the ZX81. I remember the games with such fondness, but sadly many of them are still available on the net – and they’re actually a bit crap – but at the time they seemed absolutely incredible. I spent many a happy hour with Jetpack. But the main memory I have is anxiously waiting whilst the games retrieved from the audio cassettes – would it load? Can you imagine a storage technology that was so slow and only worked 2 in every 3 times – laughable really.

Firstly, I tried to load the nvidia module:

NVIDIA: Failed to load the NVIDIA kernel module. Please check your
system's kernel log for additional error messages.
Failed to load module  "nvidia"

Clearly the correct nvidia driver for this kernel is missing.  If it was working and an update has messed it up for you, then possibly you have had a kernel update and don’t have the correct kernel headers installed.  The kernel headers are needed by dkms to rebuild the nvidia module.  If you’re lucky, you may just be able to install the Linux kernel headers and reboot:

$ sudo apt-get install linux-headers-`uname -r`
$ sudo shutdown -r now

With luck dkms will have rebuilt the nvidia module and all will be well.  If it hasn’t worked, then we need to take more dramatic action.  Let’s start with an nvidia-free zone:

$ sudo apt-get purge nvidia*

Next we should ensure that the nouveau module is blacklisted, so that we do not end up fighting the open source nouveau driver:

$ sudo vim /etc/modprobe.d/blacklist.conf
blacklist nouveau
blacklist vga16fb

Now we can install the latest nvidia modules:

$ sudo apt-get install nvidia-current nvidia-settings nvidia-kernel-common

And run the nvidia configurator:

$ sudo nvidia-xconfig

Finally reboot, and in theory the nvidia module should be rebuilt by dkms:

$ sudo shutdown -r now

After the reboot, it may be all working, but we can check that the nvidia module has been loaded:

$ lsmod | grep nvidia
nvidia 10832442 28

If this returns nothing, then it has not worked, if it returns “nvidia” as above, then you should have a working system.

I wanted to change my blog so that the home page only has an opening paragraph of each blog post, rather than having the full text.  In my opinion this makes the blog easier and more pleasant to navigate.  Unfortunately it seems that in trying to achieve this, I changed the RSS feed as well.

My apologies for this error, it wasn’t deliberate and I do appreciate that it is more convenient to read the whole text from within your chosen RSS aggregator.  Although it was nice having the additional page views - twice as many today as yesterday…

I think I have corrected the problem, so we should be back to full text for this and all subsequent posts.  If not, let me know!

Following my recent post regarding Silverlight, unfortunately I have hit disaster.  When the audio starts, it sounds choppy with a double-echo or reverb.  In short it is unusable.  All other audio works fine, even on the same webpage, but as soon as you try and listen to a Silverlight widget it sounds terrible.

Googling “silverlight rdp choppy” in desperation has not filled me with confidence, there are a number of posts including:

• Has anyone else ran into a problem where if you use a Silverlight app over Remote Desktop and it plays a WMA or MP3 file using MediaElement, the sound is choppy?

It is possible that upgrading to Windows Server 2008 would resolve this issue; although there are issues I believe in connecting to the latest RDP version using rdesktop.

I did find this advice from Microsoft, the Microsoft RDP Client was the only sound card available, but I followed the advice anyway to no avail.

What I simply do not understand is why all other audio works perfectly via RDP, except Silverlight, especially given the fact that Silverlight and RDP are Microsoft products.

Any suggestions gratefully received!

We have to be able to play Silverlight media on our two virtual Windows Server 2003 servers.  This is completely unavoidable, because you first need to authenticate via an Active-X website, so any option of using Moonlight on Linux was impossible.

The problem that we had was that when we clicked on the Play Media button, there were two pop ups stating “This page has an unspecified potential security risk” followed by another stating “A problem was encountered with the media file specified.”

I tried bypassing our proxy server;  I ensured that the Internet Enhanced Security was uninstalled; I tried adding the relevant sites to “Trusted sites”; I tried reducing Trusted Sites to “Low” security; I brought all the critical updates up-to-date.  Nothing seemed to make a difference.

Finally, I decided to review the optional Windows updates, just to see if there was anything related to Silverlight and noticed that there was a major update to .net available “.NET framework 4”. Knowing that Silverlight is a .net technology, I installed it and, after yet another reboot, the pop ups had gone and it was working.

My reason for writing about it is simply this:  I had googled endlessly about this problem, without any success, there was seemingly nothing relevant to this problem.  Hopefully this post will help others.

But you should be using Linux not Windows!  There, couldn’t resist ;)

So now this user’s PDFs open in Acrobat Reader on our Windows server, instead of using Acrobat Reader on Linux.  That would not in itself cause a problem, except that the Windows server cannot access the location of the temporary location of Kmail attachments.  I could probably fix that, but I don’t really want users opening email attachments using Windows applications.

This user cannot access File Associations to fix this problem, as the menu option is removed; so I need to find a way to fix this problem from the command line.

In solving this particular problem I discovered the existence of the users’ ~/.local/share/application directory, and was able to remove the application shortcut to the PDF and normal functionality was restored, but I thought it would be worth writing up what I have found to be the case.  Some of the following information may be wrong or misleading and please feel free to comment any issues, and I will endeavour to update this page accordingly.

KDE File Associations

I have to say that KDE file associations seem hellishly complicated, which is to say that there are so many different locations of files that it is very difficult to work out what is happening.  The three key factors in play seem to be:

MIME Types

Mime-types, appear to be stored in:

• ~/.kde/share/mimelnk directories (for each user)
• /usr/share/mimelnk
• /etc/kde-profile/standard/share/mimelnk (for kiosktool - a symlink back to /usr/share/mimelnk)

These look like this:

$ cat pdf.desktop
[Desktop Entry]
Comment=PDF
DocumentHidden=false
Icon=application-pdf
MimeType=application/pdf
Patterns=*.pdf;*.PDF
Type=MimeType
X-KDE-AutoEmbed=false

In other words they link the file extension, in this case *.pdf, with the MimeType, in this case application/pdf.  I believe I am right in saying that the users’ files take precedence over the global ones.

Application Shortcuts

Next up are the application shortcuts, and these also can be in diverse locations:

• ~/.kde/share/applnk
• ~/.local/share/applications
• /usr/share/applications
• /etc/kde-profile/standard/share/applications (for kiosk-tool)

As with the mimetypes, the users’ copy take precedence over the shared ones.  These application shortcuts define how to run an application, and look something like this:

$ cat "Adobe Reader 9 PDF Viewer.desktop"
[Desktop Entry]
Comment=
Exec=acroread
GenericName=PDF
ViewerIcon=AdobeReader9
MimeType=
Name=Adobe Reader 9
Path=
StartupNotify=false
Terminal=0
TerminalOptions=
Type=Application
X-KDE-SubstituteUID=false
X-KDE-Username=

As you can see, this tells the system how to run the application.

profilerc

Lastly there needs to be some way of tying the above two things together, something that tells the system that you use X application shortcut to open Y mime-type, and this appears to be the role of the profilerc files.  As with the other components, there are multiple copies of this file:

• ~/.kde/share/config/profilerc
• /etc/kde-profile/standard/share/config/profilerc

An extract of this file in respect of my PDF example, looks something like this:

[application/pdf - 1]
AllowAsDefault=true
Application=Adobe Reader 8 PDF Viewer.desktop
GenericServiceType=Application
Preference=2
ServiceType=application/pdf
[application/pdf - 2]
AllowAsDefault=true
Application=kpdf_part.desktop
GenericServiceType=KParts/ReadOnlyPart
Preference=1
ServiceType=application/pdf
[application/pdf - 3]
AllowAsDefault=true
Application=kde-kpdf.desktop
GenericServiceType=Application
Preference=2
ServiceType=application/pdf

This shows both the mime-type application/pdf and the application shortcut, e.g. “Adobe Reader 8 PDF Viewer.desktop”, and gives the different choices for opening the same mimetype.

I have found this exercise to be hugely frustrating, and I hope this information is of use to someone.  As I said earlier, my intention will be to correct this post as I find out more, so please forgive any errors, and if you have the time, please do let me know by commenting below.

I went to look for an article on fcheck, that I originally found on http://debaday.debian.net, only to find the site down, there is a related bug, but no news on where it has gone nor why.  The debaday service was fantastic for finding out about useful packages, in fact that was how I learned about fcheck.

Then I ended up at http://news.debian.net/, and what do I see but:

news.debian.net is closing. In the next days, the website will become a static copy.

Thank you for following all this 19 months!

No explanation, just that.  If it was coming back soon, one would have expected the announcement to state that, so the assumption has to be that these valuable services have both come to an end.

This makes me sad.  The answer is for the community to leap in, and I hope this time that the work can be distributed; so that it is not reliant on any one person.

I’ve just bought a new HP N36L Microserver, for £210 with £100 cashback.  It came with a 250gb SATA, so I added a second 250gb SATA; so that I could have a simple software RAID array.  I also added 2gb of Crucial RAM.

When the server arrived, I was surprised to find it had RAID built into it, but my understanding is that this is FakeRAID and thus should be avoided at all costs.  If you were particularly interested in that  option, with particular reference to Windows Home Server (yikes) then this page may be helpful.

I plugged in my wonderful IODD ISO drive and selected Debian Squeeze AMD64 netinstall, and off I went.  When it came to the partitioning, you need to carry this out manually, in the following order:

• Partition both disks as follows:
  • Create boot primary partition (about 500mb)
    • Set mount as /boot and make bootable
  • Create partition using most of the rest of the first drive (e.g. 247gb)
    • Set partition as Use as physical volume for RAID
• Configure software RAID
  • Create MD Device, RAID1, 2 devices, 0 spare devices
  • Select the RAID partition on each drive
• Create Physical Volumes for LVM
  • Create volume group
  • Create suitable volumes for your systems, e.g. swap (2 x RAM, root (10GB or so), home (to your taste), usr (2gb+), var (2gb+), tmp (1gb+).
    • it is a good idea not to use the whole space, as the whole benefit to LVM is being able to then add available space into any volume
  • I also created a large volume for /var/lib/vz/private, as this will be an openvz server.
• Partition disks
  • Set each LVM volume to use the appropriate file system type, e.g. ext3 or ext4, and set the appropriate mountpoint, e.g. home should be /home etc, swap should be allocated to swap space etc

I found this howto on LVM + RAID for debian-installer most useful but, providing you follow the logical progression, it really is not as difficult as it seems.

After the partitioning, I continued with the installation, removing all options including “Standard System” from the software profiles; so that I basically installed a base system, which is how I prefer it.

Next I installed: sudo, GNU Screen, openssh-server, vim, fcheck, locales, localepurge, apt-file,  apt-show-versions and nullmailer (ssmtp is  better if your mailserver requires secure authentication).  Doubtless you have your own favourite packages to install!

Next step will be to configure as a small openvz server, and migrate the virtual machines off my old server (which I will be commissioning as a storage server).

The mystery is that, at one of our branches, it reduces size to 0, but is never actually removed; after which time the file cannot be edited until that lock file is manually removed.

Both servers have the same davfs2.conf and I have done a diff between the two servers:

# for file in `find /usr/lib/openoffice/ -name *.xcs -type f`; \
  do echo "=== $file ==="; \
  ssh root@other-server "cat $file" | diff $file - ; \
  done

But there are no differences.  I have checked users’ davfs2.conf files, but these are all commented out entirely.

Previously I had “fixed” this problem, by setting UseDocumentOOoLockFile to false in:

/usr/lib/openoffice/basis3.2/share/registry/schema/org/openoffice/Office/Common.xcs

This doesn’t stop the lock files being created, but it does tell Openoffice not to worry about them.  Unfortunately this no longer seems to be working, and I really can’t understand why.

Currently I have “fixed” the problem by setting davfs2.conf to:

use_locks 0

This is now allowing the lock files to be removed properly.  I believe that actually this will work quite well, with the UseDocumentOOoLockFile set to true, given that 99.9% of our documents are Openoffice documents.  Always assuming I can persuade Openoffice to obey its own registry!

We’ve seen the various plug-in computers recently, but this one looks very good value:

TonidoPlug [TND003SP1A03] - ₤65 : Tonido Store

It was discussed on Floss Weekly in Episode 153.

Other than a vague feeling that it would be handy as an always on ssh host inside my home network, I’m just not sure I’d use it …

More about the TonidoPlug here, oh - and it comes with a selection of mobile apps for accessing your content.

Further to my recent post on the TonidoPlug the following has been sent to me:

• DreamPlug ARM box boosts plug computing thinq

The inclusion of two gigabit ports and e-sata is very exciting, but it is more than twice the price.

Still, choice is good!

Further to my recent post on the subject, I have now found more information on the Barclays Corporate website:

To access Business Internet Banking you need four things present at the same time:

1. A PC or laptop with the software that runs the smart card reader installed
2. A smart card reader. This is a hardware device, attached to your PC that the smart card is inserted into
3. A smart card unique to each individual user. The card will become “locked” if an invalid PIN is entered in excess of a pre defined number of times
4. A PIN unique to each individual user.

Unfortunately the Software Requirements link, at the bottom of that page, states only that you need Adobe Acrobat Reader, which I suspect is not the only software requirement!

The Barclays Personal On-line banking works with all the browsers, so if the same is true of the Corporate banking, then it is at least possible that I could install Firefox under Wine on Linux, providing only that I can get the smart card reader to work.

The difficulty would be if it needs Active-X support, which whilst theoretically supported under the latest versions of Wine, would make the proposition very much less likely to work.

Today I had a meeting with a telecoms expert to advise us on our options, which seem to be:

• LLU (Local Loop Unbundled) - this would boost download bandwidth to up to 24mb, but leave the upload bandwidth unchanged, there might also be a small quality improvement.  Maybe.
• SDSL - Synchronous DSL has the same upload and download bandwidth, unfortunately that bandwidth is just 2mb.  And is probably only available at one of our branches.  Cost is around £370 install and £275 per month.
• EFM (Ethernet First Mile) - based on ISDN30, this uses twin twisted copper cables, which means that the service quality degrades the further you are from the exchange.  We would hope to receive close to 10mb bandwidth upload and download.  Cost is around £2000 install and £500 per month.
• E1 Leased Line - a fibre product, offering 2mb upload and download bandwidth.  Install cost is around £3-5000 and £270pm.
• LES10 Leased Line - a fibre product, offering 10mb upload and download bandwidth.  Same day support.  Install cost is around £3-5000 and £500pm.
• LES100 & LES1000 Leased Lines - beyond our budget.

So the cheapest solution, SDSL, which actually would reduce our download bandwidth, would still cost three times what we are currently paying, and we would need to pay six times what we are currently paying, in order to get the service that we need.

Other options that we could consider are:

• Multiple ADSL lines - load balanced
• Multiple ADSL lines - bonded

For example, we could double our costs by purchasing two separate ADSL lines, and bonding such that the download bandwidth became 14mb with a 1.4mb upload bandwidth.  But is bonding reliable?  Our telecoms expert could not remember a successful example of ADSL bonding, i.e. where the customer didn’t end up ditching the solution after a short time.  Is that true do you think?

Ideally I would upgrade our connection for something that gave us 20mb+ download and 2mb upload.  Hmm, what about Virginmedia . . .

Virginmedia offers a 20mb connection for £50pm, but with an upload bandwidth of just 1mb, this wouldn’t actually address our main problem.

So it does seem as if there is no cost-effective solution, and the price gap between ADSL and Leased Lines seems very large - an additional £10,000pa for our business.

FreeRDP is a fork of the rdesktop project that intends to rapidly start moving forward and implement features that rdesktop lacks the most.

For more information please visit:

• FreeRDP.

Linux.com :: Automatically process new files with fsniper

Now I am wondering if I can use it to prompt an rsync to sync our shared documents to our remote site, and it seems I could.  This is a major headache for me, as we have two branches and a shared documents repository.

I have previously tried using Unison to synchronise between the branches, but this has created a lot of load on the master server, massively slowing down performance for all the users on the master server.

We are currently using apache2+webdav+svn but this is not working well for us.  Potentially fsniper+rsync could work very well.  It would clearly need to be running on both servers and I can see a potential clash if the two servers happen to try and update the same file at the same time.  More thought is required, any suggestions - do let me know.

The links to the fsniper site seem to be outdated in both of the above articles, but it may be found at:

http://freshmeat.net/projects/fsniper

It doesn’t appear to be packaged for Debian, but I will probably compile from source and try it soon.  More later!

We’ve all had occasions when we have needed to find where all our disk space has gone, well under KDE3.5 it was easy - you just ran Konqueror and then selected View » View Mode » File Size View.  But shock horror, under KDE4 that option seemed to be missing.  Fortunately it was only an apt-get away:

$ sudo apt-get install konq-plugins

Then run Konqueror and follow instructions above.

This is not just a static either, you can click on any directory and it will refresh for the usage of that directory.

Is it wrong to be in love with a browser?

• http://chrisjrob.wordpress.com/feed/

I did consider setting up my own Wordpress server, but rightly or wrongly I decided to save myself the effort, time will tell whether that was the right decision or not!

My gripe is this - that these companies are taking the produce of the work done by Linux developers to make new products and thus generate profits, and then not allowing those same Linux developers to access those newly developed devices.  Worse, some of these companies are now starting to support Apple Mac OS/X, but not bothering to support Linux, the very operating system that gave birth to their product.

And the daft thing is that in many cases they don’t even need to do the work themselves.

I have decided to set up a web for my wider family, as a place to chat and share information. I’ve had some experience with Drupal 6, but knowing how tough it can be to upgrade between versions of Drupal, I decided to opt for Drupal 7.

Drupal is great for one main reason - it is extendible to the n’th degree - a Drupal site can be a blog, a forum, a static site, a database application. In fact there is very little you can’t do with a Drupal site. There is a price to be paid for all the functionality, I do see Drupal as being something of a jack of all trades and yet a master of none, but, if you want flexibility without the need to host multiple services, then Drupal is it.

Installing Drupal was very simple, providing you know how to configure mysql (or another database). The instructions are clearly laid out in the tar.gz; although newer users would probably be best advised to find a howto.

Having googled for a decent link for you, I have found nothing, so here goes a brief howto - from memory (you have been warned):

Install prerequisites (hopefully I haven’t missed any that won’t be brought in anyway):

$ sudo apt-get install apache2 php5-common php5-mysql \
libapache2-mod-php5 php5-gd

Visit Drupal 7 page and copy the link to the  tar.gz file for Drupal 7 and download to a suitable directory on the server:

$ wget http://ftp.drupal.org/files/projects/drupal-7.0.tar.gz

Extract contents and change into the Drupal directory:

$ tar -xvvzf drupal-7.0.tar.gz
$ cd drupal-7.0

Read the INSTALL.txt file and follow the steps to move the installation to its proper home (under debian /var/www):

$ cd ..
$ sudo mkdir /var/www/drupal
$ sudo mv drupal-7.0/* drupal-7.0/.htaccess /var/www/drupal/

Set the ownership and group on the new directory:

$ sudo a2enmod rewrite
$ sudo /etc/init.d/apache restart

Once installed Drupal 7 looks very attractive and feels a very slick product. I was particularly impressed by the ability to paste a link to a Drupal module tar.gz file and click Install. Doubtless this is not the recommended way to install modules, but it worked fine.

One thing that most users will wish to install is a WYSIWYG editor, and after much trial and error I found that the CKEditor module (and not the WYSIWYG module + CKEditor) worked best.

On the downside, I am struggling to find all the modules that I need. As in previous versions of Drupal, there are a great many that have still not updated for Drupal 7. In particular, I would like an instant messaging module, but cannot find such a thing.  I would imagine this issue will get easier over time, and I would rather be on the current version now and be able to add features later, than be on the old version now with all features and have the worry about how I am ever going to upgrade!

Apparently I am not the only one that thinks that the Debian graphics are appalling:

Debian did have a design contest of sorts, well a brief mailing list thread that had the impossible task of voting the best out of just a handful of rubbish themes.  Given the choices, personally I would have opted for the upstream themes, or just to stick with the Debian Lenny theme which was perfectly okay.

I do wonder how easy or difficult it is to submit themes - do you have to be a Debian developer?  If so, then they should really think about lowering the bar.  I’m sure that I could produce something better, although my last attempt does not appear to be getting widespread approval.

Anyhow, we don’t love Debian for how pretty is looks - we can soon change the theme if it bothers us.

I have looked into moving to Barclays on-line business banking, but apparently it requires a hardware card reader that only works on Windows and Mac. This is a major problem as we run Linux everywhere and only access Windows via RDP to a virtual server.

Currently we use Barclays BusinessMaster that is installed on a shared Windows server that we can access via RDP or ICA. Thus any user with a BusinessMaster login can update our bank, create payments etc (subject to authority levels of course!).

Our only option now seems to be to install a standalone Windows PC somewhere and physically have to visit it in order to do our business banking.

Welcome to 2011.

I don’t have the HP Management suite installed on my HP Proliant servers, largely because they proved to be a huge resource hog, but I would like to be able to monitor stuff. On PCs I am used to being able to use lm-sensors, so thought I’d give that a go. Installing it prompted me to install ipmitool, which in turn needs openipmi.

# apt-get install lm-sensors openipmi ipmitool

Then I ran the standard sensors-detect command, accepting YES for all options (I was feeling quite brave):

# sensors-detect

This ended up writing the following into /etc/modules:

# Generated by sensors-detect on Thu Feb 17 18:04:48 2011
# You must also install and load the IPMI modules
ipmi-si
# Chip drivers
# Warning: the required module ipmisensors is not currently installed
# on your system. For status of 2.6 kernel ports check
# http://www.lm-sensors.org/wiki/Devices. If driver is built
# into the kernel, or unavailable, comment out the following line.
ipmisensors

# no driver for SMSC SCH4307 Super IO Fan Sensors yet
coretemp

To test this without rebooting I typed the commands in succession, but found that ipmisensors did not exist, some googling suggested the module to be ipmi_devintf, and that indeed seemed to work:

# modprobe ipmi-si
# modprobe ipmi_devintf
# modprobe coretemp

So the additions to my /etc/modules now looked like this:

# Generated by sensors-detect on Thu Feb 17 18:04:48 2011
# You must also install and load the IPMI modules
ipmi-si
# Chip drivers
# Warning: the required module ipmisensors is not currently installed
# on your system. For status of 2.6 kernel ports check
# http://www.lm-sensors.org/wiki/Devices. If driver is built
# into the kernel, or unavailable, comment out the following line.
# ipmisensors
ipmi_devintf

# no driver for SMSC SCH4307 Super IO Fan Sensors yet
coretemp

And then I tested the commands with:

# sensors
# ipmitool sensors

Both of which commands brought back sensible information. Now I just need to work out how to incorporate this into nagios with nrpe, but that can wait for another day.

Just installed Etherpad on our openvz server. Having created an openvz guest for the purpose, running Debian Squeeze, installing Etherpad proved no more difficult than adding the Etherpad repository to your /etc/apt/sources.list:

$ sudo su
# echo "deb http://apt.etherpad.org all ." >> /etc/apt/sources.list
# apt-get update
# apt-get install etherpad

All the requisite dependencies were brought in. Unfortunately the etherpad server did not start after installation, and I had to do the following commands:

# sed -i.orig 's/local/share/' /etc/init.d/etherpad

Then it started perfectly simply with /etc/init.d/etherpad start.

To make this happen automatically on boot:

# update-rc.d etherpad defaults

If you would prefer to use openjdk than sun-java (and you should), then google for “etherpad openjdk”. Sorry to say I was in too much of a hurry on this occasion.

References

• Etherpad
• http://sis.bthstudent.se/2011/525/how-to-install-etherpad-on-debian-lenny/

Skype is important, though, for new Linux users. It is no good telling them to use SIP, as their friends have Skype and most people do not want to make the sacrifices that some of us feel are necessary.

Feeling a little guilty, I decided to give it a go. I was expecting to download the 32-bit version of Skype, and the 32-bit libraries, but was surprised to see a 64-bit version for Ubuntu. Given that Ubuntu is based on Debian, it was worth a try:

• Download Skype for Linux

I installed it with:

$ sudo dpkg -i skype-ubuntu-intrepid_2.1.0.81-1_amd64.deb

And, erm, it seems to work. Well at least the test call worked perfectly.

So we seem to be able to tick the Skype 64-bit box as “job done”, now how about iTunes…

“For some reason it is not saving a bookmark when you stop watching a recorded program in the middle, so that you have to find where you were manually. This worked fine on our old machine, so just need to find out what is wrong. I can research this when I have time.” Read original post.

Well I am delighted to say that this was fixed by an option in the Playing TV settings for saving bookmarks.

I am a rank amateur at both tar and mt. This page constitutes no more than you could discover yourself by reading the manpages for tar and mt, or Googling.

You have been warned!

Simple instructions

Rewind

Use the rewind command before backup to ensure that you are overwriting previous backups.

# mt -f /dev/st0 rewind

Backup

E.g. directory /www and /home with tar command (z - compressed)

# tar -czf /dev/st0 /www /home

Use -v to receive verbose feedback.

Backup with Verify

If you do not compress your backup, then you can verify in the same process:

# tar -cWf /dev/st0 /www /home

Where

Find out what block you are at with mt command:

# mt -f /dev/st0 tell

This does not appear to work on my version of the software. In theory the Status option has a line for block number=, but surprising after completing a backup it seems to still return 0. If I ever work out why this is, I will update this entry.

List

Display list of files on tape drive:

# tar -tzf /dev/st0

Restore

E.g. /www directory

# cd /
# mt -f /dev/st0 rewind
# tar -xzf /dev/st0 www

E.g. /home/test directory

# cd /
# mt -f /dev/st0 rewind
# tar --checkpoint -xvvzpkf /dev/st0 home/test

--checkpoint : provide occasional checkpoint messages
-x : extract
-v : verbosely
-v : even more verbosly
-z : uncompress
-p : retaining permissions
-k : leaving existing files alone
-f /dev/st0 home/test

Unload

Unload the tape:

# mt -f /dev/st0 offline

Status

Display status information about the tape unit:

# mt -f /dev/st0 status

Erase

Erasing the tape may take hours and there is not normally any need to do this; simply rewind the tape before performing backup, or use the mt command to position at the beginning of the tape.

# mt -f /dev/st0 erase

Moving about the tape

You can go BACKWARD or FORWARD on tape with mt command itself

Go to end of data

# mt -f /dev/st0 eod

Goto previous record

# mt -f /dev/st0 bsfm 1

Forward record

# mt -f /dev/st0 fsf 1

Restore

This code has not been checked or tested:

• Check status of tape: mt -f /dev/st0 status
• Go to the directory where you want to restore your file(s).
• Go to the right file on the tape with the following commands:
  • Check file number and position in file: mt -f /dev/st0 status
  • Advance one file: mt -f /dev/st0 fsf 1
  • View contents of tar-file: tar -tvf /dev/st0
  • Go back one file: mt -f /dev/st0 bsf 1

  • If you are in the last block of a file and you should be at the beginning of the file, do the following:

    mt -f /dev/st0 bsf 1

    mt -f /dev/st0 fsf 1

  • And check with: mt -f /dev/st0 status
• Extract your file(s): tar -xvf /dev/st0 ~files~
• Rewind and eject tape: mt -f /dev/st0 offline

Hardware Compression

It may be possible to switch off and on the hardware compression on the:

# mt -f /dev/st0 compression 0
# mt -f /dev/st0 compression 1

Other people report replacing the 0 with “off” and 1 with “on”.

References

• man tar
• man mt
• Howto: Linux Tape Backup with MT and TAR commands
• How to use the DAT-tape with Linux
• Verify tar command tape backup under Linux or UNIX
• Tape drives naming convention under Linux
• Backup home directories in Linux
• Howto: Use tar command through network over ssh session
• Linux Set the Block Size for a SCSI Tape Device

I know nothing about Ruby on Rails, but this is how I ended up successfully (I think!) installing it. There may be better ways.

Download and Compile Ruby

At the time of writing, the current version is 1.9.2

• http://www.ruby-lang.org/en/downloads/

I tend to use checkinstall wherever possible, so that I can manage the package via apt-get and dpkg.

$ sudo apt-get install checkinstall

$ wget ftp://ftp.ruby-lang.org//pub/ruby/1.9/ruby-1.9.2-p0.tar.gz
$ tar -xvvzf ruby-1.9.2-p0.tar.gz
$ cd ruby-1.9.2-p0

$ ./configure
$ make
$ sudo checkinstall

Set the version number to 1.9.2

Install Rails

$ sudo gem install rails

Install Passenger

This assumes you are using apache2.

$ sudo gem install passenger
$ sudo passenger-install-apache2-module

Please read thoroughly the output from this the last command, as it will give you the information you require for the next section.

Enable Passenger

$ cd /etc/apache2/mods-available

Create or edit the following files, using the output from passenger-install-apache2-module above.

passenger.conf

<IfModule mod_passenger.c>
    PassengerRoot /usr/local/lib/ruby/gems/1.9.1/gems/passenger-2.2.15
    PassengerRuby /usr/local/bin/ruby
</IfModule>

passenger.load

LoadModule passenger_module /usr/local/lib/ruby/gems/1.9.1/gems/passenger-2.2.15/ext/apache2/mod_passenger.so

Enable in Apache

You should now be able to enable the apache module with:

$ sudo a2enmod passenger

Update Vhosts

This will vary depending on your system, but you are aiming to have a vhost configuration similar to the following:

<VirtualHost *:80>
    ServerName www.yourhost.com
    DocumentRoot /somewhere/public    # <-- be sure to point to 'public'!
    RailsEnv development              # <-- change to testing/production as appropriate (see note below)
    RackEnv development               # <-- change to testing/production as appropriate (see note below)
    <Directory /somewhere/public>
       AllowOverride all              # <-- relax Apache security settings
       Options -MultiViews            # <-- MultiViews must be turned off
    </Directory>
</VirtualHost>

By default, passenger uses the Ruby on Rails production database, which may or may not be appropriate, depending on where you are in the development process. If you are still developing your rails app, then you may want to set in the above:

RailsEnv development
RackEnv development

In theory you only need the former, but the presence of the file config.ru in your application directory makes passenger require RackEnv instead.

Otherwise, when making changes to your database, do remember to:

# db:migrate RAILS_ENV=production
# touch tmp/restart.txt

Otherwise your application will not work.

ISPConfig

If you are using ISPConfig, then I currently have this working by adding the following to the sites Apache Directives in the admin control panel:

DocumentRoot /var/www/web11/web/blog/public
<Directory /var/www/web11/web/blog/public>
    AllowOverride all
    Options -MultiViews
</Directory>

And then edit /etc/apache2/vhosts/Vhosts_ispconfig.conf and remove the old DocumentRoot line from the relevant vhost section. Unfortunately you will have to do this everytime you change your site configuration. This is very poor, there has to be a better way.

Installing JQuery

These are rather poor notes, you are probably better looking elsewhere, but the key thing is that the long and complex instructions for installing jquery that abound on the Internet should be avoided. The installation should be a mere couple of commands.

Either type:

$ gem install jquery-rails

Or as I preferred, edit your Gemfile (in the root of your rails project) and add:

gem 'jquery-rails'

My handwritten notes don’t say this, but I believe you would then need to run:

$ bundle install

And then (add “–ui” on the end of the following command to include the jquery UI):

$ rails generate jquery:install

You’ve inserted a disk or memory stick in your thin client, and nothing happened, what next?

Check lts.conf

Each user you wish to be able to access their local devices needs LOCALDEV = True in lts.conf, alternatively you may add to the ‘[default]’ section to make this global.

[192.168.1.198]
# Joe Bloggs Desktop
    LOCALDEV             = True

Check user is in fuse group

# groups userid
userid satff users audio fuse

If they are not in the fuse group then:

# adduser userid fuse

Check lts.conf settings are reaching client

A simple typo can mean that your carefully crafted lts.conf file is not reaching its target audience. To test, boot up the thin client and ctrl+alt+f1 to get to the console, login as root and type the following:

See LTSP Client Root Password if you are unable to log into the client.

# getltscfg -a

This will give you a list of all the settings that apply to this particular client, including any default settings. You need to check that you can see the above LOCALDEV=”True” is visible.

Check logs on client

# tail -f /var/log/syslog

Then try inserting the device, and you should see the detection - look for errors that might indicate what has gone wrong.

Check UDEV can see drive

Still on the client:

# udevadm info -q all -n sda1

Replace sda1 with the required device, e.g. sr0 for a cdrom etc.

P: /block/sda/sda1
N: sda1
S: block/8:1
S: disk/by-id/usb-USB_DISK_2.0_J68J5KQHN7TQ0H7G-0:0-part1
S: disk/by-path/pci-0000:00:1d.7-usb-0:2:1.0-scsi-0:0:0:0-part1
S: disk/by-uuid/99B8-E4E6
S: disk/by-label/PKBACK#\x20001
E: ID_VENDOR=USB
E: ID_MODEL=DISK_2.0
E: ID_REVISION=1219
E: ID_SERIAL=USB_DISK_2.0_J68J5KQHN7TQ0H7G-0:0
E: ID_SERIAL_SHORT=J68J5KQHN7TQ0H7G
E: ID_TYPE=disk
E: ID_INSTANCE=0:0
E: ID_BUS=usb
E: ID_PATH=pci-0000:00:1d.7-usb-0:2:1.0-scsi-0:0:0:0
E: ID_FS_USAGE=filesystem
E: ID_FS_TYPE=vfat
E: ID_FS_VERSION=FAT16
E: ID_FS_UUID=99B8-E4E6
E: ID_FS_UUID_ENC=99B8-E4E6
E: ID_FS_LABEL=PKBACK# 001
E: ID_FS_LABEL_ENC=PKBACK#\x20001
E: ID_FS_LABEL_SAFE=PKBACK#_001

An example of a CDROM:

P: /block/sr0
N: sr0
S: block/11:0
S: scd0
S: disk/by-path/pci-0000:00:1f.2-scsi-1:0:0:0
S: cdrom
S: cdrw
S: dvd
S: dvdrw
E: ID_CDROM=1
E: ID_CDROM_CD_R=1
E: ID_CDROM_CD_RW=1
E: ID_CDROM_DVD=1
E: ID_CDROM_DVD_R=1
E: ID_CDROM_DVD_RW=1
E: ID_CDROM_DVD_RAM=1
E: ID_CDROM_DVD_PLUS_R=1
E: ID_CDROM_DVD_PLUS_RW=1
E: ID_CDROM_DVD_PLUS_R_DL=1
E: ID_CDROM_MRW=1
E: ID_CDROM_MRW_W=1
E: ID_PATH=pci-0000:00:1f.2-scsi-1:0:0:0

Check ltspfs is installed on the server

# dpkg -l ltspfs | grep ^ii
ii  ltspfs                                       0.6-1~bpo50+1                        Fuse based remote filesystem for LTSP thin clients

Check fuse is running

# lsmod | grep fuse
fuse                   47124  3 

If fuse is not running, then:

# echo fuse >> /etc/modules
# modprobe fuse
# adduser USER fuse

KDE Desktop Icons

• Local Device Icons on KDE Desktop

References

For some additional troubleshooting, try: http://wiki.ubuntu.com/DebugLocalDev

Despite the cheap PCs, there are still some direct savings to be made, obviously Linux users can save the Microsoft licence, but there are other savings too. For example you can re-use components from your existing PC, this might be a short-term way of you getting the PC that you want, and then replacing the components over time.

Already have a keyboard, mouse and DVD/RW? Reuse them - it’s more environmentally friendly and, if they fail later, then you can always replace them and thus stagger the cost. Why not plan to re-use your existing hard-drive, perhaps as a second drive or as a back-up drive, it’ll make it easier to transfer data. If you’re trying to save money then you could build your PC using on-board graphics and then plan to buy a decent graphics card at a later stage.

Another advantage to building your own PC is that you get to choose the compromises that are made. The cheapest components are incredibly cheap, so for example if you were planning to keep the base unit behind the sofa, then you could specify a £15 case, instead of a £100 case.

You also get to specify things like RAM - we’ve all gone to upgrade PCs, only to find that both slots have already been used - now you can make sure that you buy a single 2GB RAM with the knowledge that you can buy a second when you have more money.

It is this type of flexibility which still makes building your own compelling.

It is true that it takes longer to build your own, but I find the bulk of the time taken is in choosing the components, the build itself is usually no more than an hour, and I probably waste three hours in agonising over the purchase of the components! The only thing that is tricky is in deciding which way around the different connectors go, but generally the booklets that come with the motherboard are very helpful, and failing that the Internet is a wonderful resource.

The ability to repair the device is I find not so much down to your experience of the build, but in your confidence that all the components are standard off-the-shelf items. If you suspect that the graphics card is shot, then, buying a replacement is easy, because you know exactly what you need.

I tend to choose the CPU, then knowing the CPU I purchase a compatible motherboard, from there it is pretty simple to choose the case to fit the motherboard, and the rest of the items should be straightforward. PSU obviously needs to be sensitive to the case, but if you have gone for a standard ATX case, that should not be an issue.

Alternatively Lambda-Tek have a PC Designer to do this for you, I have not used it myself yet, but it looks impressive.

If you want a small form factor device than visit Lin-ITX, who have knowledgeable people on the end of the phone to advise you.

First I removed OpenOffice:

$ sudo apt-get --purge remove ".*openoffice.*"

Then I visited the Document Foundation LibreOffice website and downloaded:

• LibO_3.3.0_beta2_Linux_x86_install-deb_en-US.tar.gz
• LibO_3.3.0_beta2_Linux_x86_langpack-deb_en-GB.tar.gz

Then, having checked that the md5sums matched, I extracted both with tar -xvvzf and changed to the DEBS directory in each and ran:

$ sudo dpkg -i *.deb

All very clumsy, but this was only my work PC, which isn’t mission critical, as I generally use NX client to log into my desktop on our LTSP Server.

So what are my first impressions? Sadly these will have to wait for another day, apart from the obvious branding change, I couldn’t see much different, but the good news is that it does all appear to be working.

I am not sure whether LibreOffice yet has the Go-OO Patches patches, which I believe were included by default in Debian and Ubuntu, but the fact that they are being combined into the core must be good news.

I am really hoping that LibreOffice improves quickly on OpenOffice, which still falls considerably short of Microsoft Office for power users. In particular Macros, Pivot Tables, Conditional Formatting, and Mailmerge all need a lot of work. Not to speak of the appalling performance, it really is a system hog. I can’t help feeling that it really needs to be re-written from scratch, but I appreciate that is unrealistic.

In the meantime I have hopes that KOffice will soon step up to the mark - KWord’s mailmerge feature is a lesson in simplicity that OpenOffice/LibreOffice would do well to learn.

Install Ruby on Rails3 on Debian Lenny

Warning

I know nothing about Ruby on Rails, but this is how I ended up successfully (I think!) installing it. There may be better ways.

Download and Compile Ruby

At the time of writing, the current version is 1.9.2

• http://www.ruby-lang.org/en/downloads/

I tend to use checkinstall wherever possible, so that I can manage the package via apt-get and dpkg.

$ sudo apt-get install checkinstall

$ wget ftp://ftp.ruby-lang.org//pub/ruby/1.9/ruby-1.9.2-p0.tar.gz
$ tar -xvvzf ruby-1.9.2-p0.tar.gz
$ cd ruby-1.9.2-p0

$ ./configure
$ make
$ sudo checkinstall

Set the version number to 1.9.2

Install Rails

$ sudo gem install rails

Install Passenger

N.B. This assumes you are using apache2.

$ sudo gem install passenger
$ sudo passenger-install-apache2-module

Please read thoroughly the output from this the last command, as it will give you the information you require for the next section.

Enable Passenger

$ cd /etc/apache2/mods-available

Create or edit the following files, using the output from passenger-install-apache2-module above.

passenger.conf

<IfModule mod_passenger.c>
PassengerRoot /usr/local/lib/ruby/gems/1.9.1/gems/passenger-2.2.15
PassengerRuby /usr/local/bin/ruby
</IfModule>

passenger.load

LoadModule passenger_module /usr/local/lib/ruby/gems/1.9.1/gems/passenger-2.2.15/ext/apache2/mod_passenger.so

Enable in Apache

You should now be able to enable the apache module with:

$ sudo a2enmod passenger

Update Vhosts

This will vary depending on your system, but you are aiming to have a vhost configuration similar to the following:

<VirtualHost *:80>
    ServerName www.yourhost.com
    DocumentRoot /somewhere/public    # <-- be sure to point to 'public'!
    RailsEnv development              # <-- change to testing/production as appropriate (see note below)
    RackEnv development               # <-- change to testing/production as appropriate (see note below)
    <Directory /somewhere/public>
        AllowOverride all             # <-- relax Apache security settings
        Options -MultiViews           # <-- MultiViews must be turned off
    </Directory>
</VirtualHost>

By default, passenger uses the Ruby on Rails production database, which may or may not be appropriate, depending on where you are in the development process. If you are still developing your rails app, then you may want to set in the above:

RailsEnv development
RackEnv development

In theory you only need the former, but the presence of the file config.ru in your application directory makes passenger require RackEnv instead.

Otherwise, when making changes to your database, do remember to:

# db:migrate RAILS_ENV=production
# touch tmp/restart.txt

Otherwise your application will not work.

ISPConfig

If you are using ISPConfig, then I currently have this working by adding the following to the sites Apache Directives in the admin control panel:

DocumentRoot /var/www/web11/web/blog/public
<Directory /var/www/web11/web/blog/public>
    AllowOverride all
    Options -MultiViews
</Directory>

And then edit /etc/apache2/vhosts/Vhosts_ispconfig.conf and remove the old DocumentRoot line from the relevant vhost section. Unfortunately you will have to do this everytime you change your site configuration. This is very poor, there has to be a better way.

Installing JQuery

These are rather poor notes, you are probably better looking elsewhere, but the key thing is that the long and complex instructions for installing jquery that abound on the Internet should be avoided. The installation should be a mere couple of commands.

Either type:

$ gem install jquery-rails

Or as I preferred, edit your Gemfile (in the root of your rails project) and add:

gem 'jquery-rails'

My handwritten notes don’t say this, but I believe you would then need to run:

$ bundle install

And then (add --ui on the end of the following command to include the jquery UI):

$ rails generate jquery:install

These instructions have not been tested since re-writing, in particular there will be permissions issues.

Step 1: Install Wine

Please visit for instructions on how to install Wine:

• http://www.winehq.org/site/download

If you get a problem with applications appearing without any text, then you need to install some Windows fonts in Wine. One option is to simply download them from http://sourceforge.net/projects/corefonts/. You may also be able to install via a debian installer package:

$ sudo apt-get ttf-mscorefonts-installer

Install Wine packages:

$ sudo apt-get install libwine libwine-print libwine-sane libwine-alsa

Your requirements may vary, a non-inclusive list of available libraries is:

• libwine: the core wine library
• libwine-print: the printing library
• libwine-sane: the scanning library
• libwine-alsa: the alsa sound library
• libwine-oss: the oss sound library

You will notice winetools mentioned on several websites, my understanding is that this is no longer supported.

Step 2: Create a Wine directory for your application

By default Windows applications are installed in /home/user/.wine. If you want to install your application in a different directory then you must use the WINEPREFIX variable.

If you want your application to be installed in /usr/local/share/appname you must first create the directory:

$ cd /usr/local/share
$ mkdir appname
$ WINEPREFIX="/usr/local/share/appname" wineprefixcreate

In this directory you will now find a fake Windows drive and the Wine configuration files system.reg and user.reg.

Step 3: Configure your Wine directory

You can probably skip this step for typical installations.

$ WINEPREFIX="/usr/local/share/appname" winecfg

Step 4: Install your application

Installing from an .msi file

You can install .msi files with the msiexec.exe utility. This command is built into Wine, so you don’t have to install Windows Installer from Microsoft. If the .msi file is called msifile.msi you just have to type msiexec /i msifile.msi and the application will be installed.

Make sure you’re not using a native version of msi.dll, but the builtin one. To run the Windows install program, please type the following:

$ WINEPREFIX="/usr/local/share/appname" wine /path/to/setup.exe

Substitute “/path/to/setup.exe” with the installation path executable for your application

Step 5: Test your application

$ WINEPREFIX="/usr/local/share/appname" wine "drive_c/Program Files/app/appname.exe"

Step 6: Configure for multi-user operation

This section needs a re-write (or just refer to Example automation script below)

It is possible to configure the system so that a system Wine installation (and applications) can be shared by all the users, and still let the users all have their own personalized configuration.

# Creating individual user directory, e.g. /home/username/.appname
# Symlink directory structure to /usr/local/share/appname
# Remove symlink for user.reg (will be auto-created for each user on first run)

$ ln -sf /usr/local/share/appname/system.reg system.reg

To assist in this process, please refer to the following guide:

• http://www.winehq.org/site/docs/wineusr-guide/using-regedit

and perhaps even symlink these back to the administrator’s account, to make it easier to install apps system-wide later:

You might be tempted to do the same for user.reg as well, however that file contains user specific settings. Every user should have their own copy of that file along with the permissions to modify it.

You’ll want to pay attention to drive mappings. If you’re sharing the system.reg file you’ll want to make sure the registry settings are compatible with the drive mappings in ~/.wine/dosdevices of each individual user. As a general rule of thumb, the closer you keep your drive mappings to the default configuration provided by wineprefixcreate, the easier this will be to manage. You may or may not be able to share some or all of the actual “c:” drive you originally installed the application to. Some applications require the ability to write specific settings to the drive, especially those designed for Windows 95/98/ME.

Example automation script

This script was probably cobbled together from the superb ies4linux installation, or possibly from Franks Corner,

#!/usr/bin/perl
# Launch script for multiuser appname, master copy in /usr/local/share/appname.

use strict;

my $MASTERPREFIX = '/usr/local/share/appname';
my $WINEPREFIX   = $ENV['HOME'} . '/.appname';

use vars qw($MASTERPREFIX $WINEPREFIX);

# If we're not set up, create the user's magic symlink-copy of the master
# installation.

unless (-d "$WINEPREFIX") {
	system('mkdir', '-p', "$WINEPREFIX") == 0 ](| die "mkdir -p $WINEPREFIX: $?";
	system('lndir', "$MASTERPREFIX", "$WINEPREFIX") == 0 || die "lndir failed: is xutils installed? $?";

	# The profile directory will be recreated automatically by wine on the
	# first run, so it's OK to blow away eny existing copy hanging around in
	# unwritable symlink form from the install, and it's OK to just create a
	# blank one.
   
	$PROFILES = $WINEPREFIX . '/drive_c/windows/profiles';
	system('rm', '-fr', "$PROFILES") == 0 || die "rm -fr $PROFILES failed: $?";
	system('mkdir', '-p', "$PROFILES") == 0 || die "mkdir -p $PROFILES failed: $?";

	# WINE needs to be able to write to the following files, so they can't be
	# symlinks to root-owned ones.
   
	foreach my $file ('.no_prelaunch_window_flag', 'system.reg', 'userdef.reg', 'user.reg') {
		if (-l $WINEPREFIX . '/' . $file) {
			system('rm', '-f', $WINEPREFIX . '/' . $file) == 0 || die "rm -f $WINEPREFIX/$file failed: $?";
			system('cp', $MASTERPREFIX . '/' . $file, $WINEPREFIX . '/' . $file) == 0 || die "Cannot copy $MASTERPREFIX/$file to $WINEPREFIX/$file: $?";
		)
	}
	system('chmod', '-R', 'og-rwx', $WINEPREFIX) == 0 || die "Cannot chmod -R og-rwx $WINEPREFIX: $?";

}

exec('env', 'WINEPREFIX=' . $WINEPREFIX, 'wine', $WINEPREFIX . '/drive_c/Program Files/appname/appname.exe', @ARGV) == 0 || die "Failed to launch appname: $?\n";

exit;

References

• WineHQ
  • Wine Mailing List
• LTSP
  • LTSP Mailing List
• IES4Linux

• Franks Corner

  http://frankscorner.org/}

• Wine printing problems:
  • http://www.la-sorciere.de/Wine-HOWTO/wineprintconfig.html
  • http://groups.google.com/group/comp.emulators.ms-windows.wine/msg/b05da7237281dbab

This is just an aide-memoire, please don’t try and follow it.

Creating a template

# rm -f /etc/ssh/ssh_host_*
cat << EOF > /etc/rc2.d/S15ssh_gen_host_keys
#!/bin/bash
ssh-keygen -f /etc/ssh/ssh_host_rsa_key -t rsa -N ''
ssh-keygen -f /etc/ssh/ssh_host_dsa_key -t dsa -N ''
rm -f \$0
EOF
# chmod a+x /etc/rc2.d/S15ssh_gen_host_keys

Tip re. freenx-server

You will need to reconfigure the freenx-server:

$ sudo dpkg-reconfigure freenx-server

Updating template

$ sudo vzctl enter XXX
$ sudo apt-get clean
$ exit
$ sudo vzctl set XXX --ipdel all --save
$ sudo vim /var/lib/vz/private/XXX/etc/resolv.conf
$ sudo rm -f /var/lib/vz/private/XXX/etc/hostname
$ sudo vzctl stop XXX
$ cd /var/lib/vz/private/XXX
$ tar --numeric-owner -zcf /var/lib/vz/template/cache/debian-5.0-i386-minimal.tar.gz .

Create configuration template for 10 vms

$ sudo vzsplit -n 10 -f vps.tenth

Create new vm

sudo vzctl create XXX --ostemplate debian-5.0-i386-minimal --config vps.basic
sudo vzctl create XXX --ostemplate debian-5.0-i386-kde --config vps.tenth
sudo vzctl start XXX
sudo vzctl set XXX --ipadd 192.168.0.XXX --save
sudo vzctl set XXX --nameserver 192.168.0.254 --save
sudo vzctl set XXX --onboot yes --save
sudo vzctl set XXX --hostname server-XXXXXX.example.co.uk --save
sudo vzctl set XXX --diskspace $(( 1048576*2 )):$(( 1153434*2 )) --save
sudo vzctl set XXX --diskspace 10G:11G --save
vzcfgvalidate /etc/vz/conf/XXX.conf

Other tricks

sudo vzctl start XXX
sudo vzctl exec XXX passwd
sudo vzctl exec XXX ps aux
sudo vzctl enter XXX
sudo vzctl stop XXX

Removing a vm

$ sudo vzctl destroy XXX
$ sudo rm /etc/vz/conf/XXX.conf.destroyed

List running and non-running

sudo vzlist -a

Memory use

Check memory usage for guest 103

$ sudo vzcalc -v 103

Set minimum memory 256 x memory wanted

$ sudo vzctl set vpsid --vmguarpages $((256 * 256)) --save

Set maximum memory 256 x memory wanted

$ sudo vzctl set vpsid --privvmpages $((256 * 1024)) --save

References

• http://wiki.openvz.org/Debian_template_creation
• http://wiki.openvz.org/Resource_shortage
• http://www.linux.com/archive/feature/114214

First let’s take a look at where KDE squirrels away all these settings. KDE is complex, but it’s well-organised, and there is always a method to whatever KDE madness you find yourself in.

• /etc/kderc : This points to the file locations where configuration profiles are stored.
• /etc/kde3 : Global configuration files are here.
• /home/~username~/.kde : User’s individual settings go here.

You can see for yourself where your particular KDE installation is going to look for configuration files:

$ kde-config --path config
/home/carla/.kde/share/config/:/etc/kde3/:/usr/share/kubuntu-
default-settings/kde-profile/default/share/config/

KDE reads these in reverse order, so if it encounters any conflicts, the last value read is the one used. User’s individual configurations take precedence, so they are read last. There is a way to make an exception to the precedence rule: using $i, or the “immutable” key. You’ll see how this is used in the examples below. Anything marked immutable has precedence, no matter what order it’s in.

Installing Kiosktool

This is now unnecessary, as this has been fixed on Debian, you should now be able to simply do an aptitude install.

Kiosktool is simply an easy way of administering the above files. Kiosk Admin Tool is available in Debian Testing/Unstable and should be as simple to install as:

$ sudo apt-get install kiosktool

Unfortunately the current version of Kiosktool is broken for Debian, as Debian have changed the filename for the menu configuration from applications.menu to kde-applications.menu, and unfortunately Kiosktool is now unmaintained, so to install requires a few more steps than normal:

Install the build dependencies

$ sudo apt-get build-dep kiosktool

Before installing the dependencies, please copy them into the clipboard.

Download the source

Note that you do not need to be root to do this, and it helps later if you’re not! Change your repository if necessary, or remove it entirely to install from your default repository. Note however that at the time of writing Kiosktool is only available in testing, unstable and backports (the same versions in each).

$ apt-get -t etch-backports source kiosktool

Edit the source

This will download three files: a .orig.tar.gz, a .dsc and a .diff.gz. In the case of packages made specifically for Debian, the last of these is not downloaded and the first usually won’t have “orig” in the name. The .dsc file is used by dpkg-source for unpacking the source package into the directory packagename-version. Within each downloaded source package there is a debian/ directory that contains the files needed for creating the .deb package.

Navigate to this directory and open the following file in a text editor:

/path/to/kiosktool-1.0/kiosktool/menueditComponent.cpp

Replace all instances of applications.menu with kde-applications.menu.

Build the deb

From within the directory that was created for the package after downloading.

$ apt-get -b source kiosktool

To install the package built by the commands above one must use the package manager directly, like this:

# dpkg -i kiosktool-<version>.deb

Remove the build dependencies

It seems that some of the build dependencies may be incompatible with the LTSP enlightened sound daemon (esd), if you copied them (as recommended in Install the build dependencies section), then please purge them now. Otherwise you will need to review /var/log/dpkg.log for the relevant files.

$ sudo apt-get --purge remove package1 package2 etc

Using Kiosktool

After installation you should have a menu entry for it in System -> KIOSK Admin Tool, or use the kiosktool command to start it from the command prompt.

You may find that a copying dialog crops up at regular intervals and that often it will stall, please read:

• Fix kiosktool fish protocol stalls

References

This document was cobbled from the following sources:

• APT HOWTO Chapter 6 Working with source packages
• Lock Down Desktops with KDE Kiosk by Carla Schroder

Other references:

• http://techbase.kde.org/User:Danimo/KIOSK

The advice in the Edit the source section was copied from a mailing list post, which I now cannot find. The post also mentioned that patches were available for this issue, although I could not find any such patches, and anyway the method above is probably just as easy.

When making changes to profiles under kiosktool, it stalls when uploading changes to the local machine using the fish protocol, see screenshots.

Persistently clicking cancel followed by retry does deliver the goods, but sometimes it can take ten attempts or more, with the load average quickly escalating to ten times the server’s usual level.

I understand it requires the fish protocol, even though it is editing the local machine, because kiosk has to be run as a standard user, but requires root access to edit the profile.

The fish protocol usually works fine in konqueror.

Under our Kiosk configuration I have the base directory set as /etc/kde-profile and the option “On exit, upload profiles to a remote server” is unchecked.

Screenshots

Workaround

Keep cancelling the copy and then retry.

Never come out without successfully completing the copy, or you can end up with corruption in your personal desktop, or menus, or even end up unable to log in. You have been warned.

The Solution

The solution seems to be to edit the menu entry for Kiosk-Tool and change it to:

env KDE_FORK_SLAVES=true kiosktool

References

• http://lists.kde.org/?t=111633283900002&r=1&w=2 (Similar problem under SUSE)
• http://bugs.kde.org/show_bug.cgi?id=8137 (similar problem under FreeBSD)

Earlier this week I went up to the Westminster eForum to speak at the Keynote Seminar “Open source software: in business, in government”. The Westminster eForum provides an environment for policy makers in Parliament, Whitehall and government agencies to engage with key stakeholders. The event was sponsored by Sirius IT, the company which advised us on our move to Linux and who now provide ongoing support for our Linux servers.

I caught a train from Woking, arriving promptly at 61 Whitehall at 0845. I was greeted warmly by Bash, the person organising the seminar, who showed me around and took my slideshow presentation.

Rupert Goodwins, Editor at ZDNet UK, was Chairman of the meeting and introduced Karsten Gerloff, President of the Free Software Foundation Europe. I recognised some of the themes of his speech, including the definition of Free and Open Source and the “Four Freedoms”, and I was already aware of some of the developments in countries like Brazil. That said, I was not the target audience for his speech and I am sure that it must have been a powerful message for those MPs present.

There were four more speeches before mine, starting with Alan Lord of the Open Learning Centre, an Open Source Software Consulting and Services business based in Surrey. Alan discussed the challenges of Open Source, in particularly the challenge of procurement, given that standard procurement practices struggled with the concept of free software. Alan went on to reassure the audience regarding the risks and vulnerability of Open Source. This is something about which I feel passionately; to my mind, if you have the source code and access to that incredible resource “The Internet” then any risks and vulnerabilities are solvable. This is in stark contrast to proprietary software, where you are completely in the hands of the supplier.

Next up was Paul Holt, Corporate Sales Director at Canonical. I was impressed that he avoided the temptation to do a big sell on Canonical, and this was a feature of all the speeches, no-one seemed to be overtly pushing their wares. He did labour the point that Open Standards are what is important, much more so than Open Source, and in this he is undoubtedly right; the world would be a rosy place indeed if we all ran a combination of Open Source and proprietary systems, all of which worked together smoothly according to common open standards.

Andrew Katz from Moorcrofts went on to reassure the audience on the safety of running Open Source software as an end user. I must confess that this was not something about which I was worried, but in the face of the usual FUD from its detractors, it was probably worthwhile to tackle this issue. It was interesting to hear from Andrew that he often asks the Chief Financial Officer whether software contains any Free and Open Source software to be told that it doesn’t, only to find that quizzing the Chief Technology Office gives the opposite answer. Having written some simple scripts, I can’t imagine how you could do so without using the myriad of Open Source libraries; it is interesting to hear that, all too often, proprietary software companies find exactly the same!

At last it was my turn. Thanks to the magic of organisation, my slideshow was up and Rupert Goodwins had announced me. Yikes. I ran through my speech, much as I rehearsed it, even remembering to switch between slides. The speakers were only permitted to talk for a maximum of 5 minutes, with a yellow card being raised at 4 minutes and a red card at 5 minutes. Nothing like a little time pressure to calm my nerves! In the event I needn’t have worried about yellow and red cards, as I completely forgot to look out for them anyway; hopefully my speech ended broadly on time!

Having been given the brief of the discussion, my first draft of the speech followed very closely to the other speakers, outlining the implementation of Open Source, the Challenges, and the Risks and Vulnerabilities. After reading my rather poorly written offering, I decided that I should leave the other speakers to follow that pattern and to treat myself instead as a case study, not attempting to address any particular issues, but merely to run through what we had done and why. My speech seemed fairly well received and I don’t think I made too many mistakes.

I had expected that there would be questions after each speech, but in the event it seemed that questions were to be left for the end of all of our speeches. I was somewhat astonished by the first question from Robert Onslow, in which he alluded to the fact that Open Source is not as flexible as proprietary software. I apologise to Robert, as I did laugh; which was most rude of me, but if Open Source is anything, then it is flexible and thus I was so taken aback by such a question. The fact is that Open Source is most often written as small components - “do one thing and do it well” - and combining these components together is what gives Open Source its incredible flexibility. And the fact, of course, that you have access to the source code and have the option of paying a developer to make changes to it.

I also laboured my hobby horse regarding the benefit of having access to the source code. I hear so many people, even Open Source users, saying that they don’t think that having the source code is really that big a benefit. I feel passionately that even non-developers can benefit from Open Source, partly through the wider community having access to the code, but also directly, by viewing the code to debug a problem, or even fix a simple typological bug. The best example that I have was when KDE changed one of their environment paths, which broke kiosk-tool, I was delighted to be able to track down that error and fit it. Similarly getting KAddressBook to work with LDAP resources needed a very simple code change.

There were a number of other questions, but none in which I played any key part.

After the session was a coffee break, and I got talking to Mark Taylor of Sirius and Rupert Goodwin and David Meyer from ZDNet UK. David Meyer showed me his netbook running Jolicloud, which seemed to be a very professional and user friendly implementation of Linux, in which Linux, Windows and Web software are offered seamlessly from the software centre. Very impressive, and I’ll certainly be looking at that soon.

I regret that I did not stay for the remainder of the day; whilst it sounded very interesting, the majority was related to the public sector, which is not directly relevant to me.

Thank you to Bash at Westminster eForums and Mark Taylor of Sirius. Oh, and Rupert Goodwins, for allowing me to share his taxi to the station when the heavens opened!

After a considerable time spent googling without any success, I decided to upgrade the Kernel from 2.6.26 to the backports 2.6.32 and normal operation was restored. I mention this only in the hope that the next person who googles qemu-kvm black screen might be saved a considerable time searching for the solution!

• Live TV - Pause Rewind etc
• Programme Guide
• Schedule Recordings
• MythVideo (for ripped DVDs)
• MythWeb (personal webpage for controlling Myth and viewing programme guide)

The best news is that Audio is now working over HDMI; although to be honest, I don’t believe I ever fixed this, so I suspect I just hadn’t tested it properly. All I did was select the Audio over HDMI setting in KDE Control Centre and it just worked.

So what is left to fix?

• System does not shutdown automatically - this is normal behaviour for MythTV, but I prefer to have it shutdown when not in use. This is a neat feature actually, as it will literally completely shutdown with a timer set in the bios to come on for the next schedule. Hopefully I can sort this out soon.
• MythTV Backend needs to be restarted every time the machine is powered up - this is a known issue and may even have been sorted; having been without Internet for the past three weeks, I have no idea where we are. Basically on first boot we just type $ sudo service mythtv-backend restart. Obviously if there is no fix, it would be simple enough to add to the start-up process. Subsequently fixed.
• My card has dual tuners, so can record one thing and watch another, or record two things. I have now read that some channels are on the same “mux” and that you can therefore record multiple programs on a single card. I need to research this further, but sounds exciting.
• For some reason it is not saving a bookmark when you stop watching a recorded program in the middle, so that you have to find where you were manually. This worked fine on our old machine, so just need to find out what is wrong. I can research this when I have time. Subsequently fixed.
• The case fans are a little noisier than they ought to be, so really need replacing for silent alternatives.

And, not MythTV related, but I am still struggling with KDE4, Akonadi and our Contacts in Kontact. I suspect this may be related to the use of MySQL for both MythTV and Akonadi. In the meantime my wife is using Gmail, but obviously would prefer to have Kontact back.

When I launched our first LTSP server, I was very keen to make the system as pleasurable for the users as possible; to this end, I left users able to customise their desktop pretty much how they liked. I decided that this free-for-all had gone too far, when I saw customers sitting in front of one user’s computer, sporting a desktop wallpaper of his baby son. At the time I just set a system default of a fairly neutral blue / dark blue radial gradient, and put an item on my to-do list to create a desktop wallpaper.

Those who read my earlier post regarding KDE Wallpaper may have seen examples, and basically I created a red background with company logo and a few lighting affects, making suitable wallpapers for most of the common screen sizes. It was when I went to implement these wallpapers that I realised that I had a problem - KDE only allows you to force one single wallpaper to all users, regardless of screen size. I could allow the users to set their own wallpaper, from a select list, but it would be impossible to restrict them only to that list. And what if they hop from one computer to another - the screen size will be wrong again. No, I needed a way of changing the wallpaper at login to the correct wallpaper for their screensize, and that meant a perl script running at login. The task could be broken into three parts:

1. Determine screen size
2. Choose best wallpaper
3. Set wallpaper

To determine the screen size, I used a small Perl Tk routine:

sub get_screen_size {
    use Tk;
    my $mw=tkinit;
    $mw->withdraw;
    return($mw->screenwidth, $mw->screenheight, $mw->screendepth);
}

To select the best wallpaper, I found this far more complicated that I would have liked - the issue is that NX users have screen sizes that are non-standard, so I couldn’t rely on exact matches. In any case, if I relied on perfect matches I would have been creating new wallpapers for each and every obscure monitor resolution. The best approach I found was to score each wallpaper based on aspect, width and height, calculating the percentage error for each compared to the users screen size, summing the errors and picking the wallpaper with the lowest error. Logically aspect would be far more important that the other errors, so I trebled the aspect error to make that the dominant criteria.

Lastly, I set the wallpaper using a simple dcop command:

`dcop kdesktop KBackgroundIface setWallpaper "$newwallpaper" 6`;

After testing it thoroughly, I inserted a shortcut to the script in all users’ Autostart directories last night.

I arrived this morning to the comment “it’s a bit bright any chance we could tone it down a bit”.

I was very polite.

$ sudo apt-get install libgd-tools

Create image in Gimp

Create a suitable image in Gimp, ensuring that it has a transparent rather than a white background.

From the menu - Image → Mode → Indexed and “Generate Optimum Palette” with “Maximum number of colours: 255”.

Save result as a PNG file, e.g. “router.png”

Convert to GD2

$ pngtogd2 router.png router.gd2 0 1

It seems that the HVR-2200 is something of a work-in-progress under Linux. Plenty of people seem to have it working, but it is not simple to configure. See the references at the end for further reading.

I now have this card working under MythTV, but I have not yet managed to get the programme guide updating via EIT (Freeview).

Those new to MythTV should visit: http://parker1.co.uk/mythtv_ubuntu.php

Installation

$ sudo apt-get install mercurial libncurses5-dev unzip

$ wget http://www.steventoth.net/linux/hvr22xx/22xxdrv_27086.zip
$ wget  http://www.steventoth.net/linux/hvr22xx/HVR-12x0-14x0-17x0_1_25_25271_WHQL.zip
$ wget http://www.steventoth.net/linux/hvr22xx/firmwares/4019072/NXP7164-2010-03-10.1.fw
$ wget http://www.steventoth.net/linux/hvr22xx/extract.sh
$ sh extract.sh
$ sudo cp *fw /lib/firmware
or
$ sudo cp *fw /lib/firmware/`uname -r`

$ hg clone http://kernellabs.com/hg/saa7164-stable/

$ cd saa7164-stable
$ make CONFIG_DVB_FIREDTV:=n
$ sudo make install
$ wget http://www.steventoth.net/linux/hvr22xx/firmwares/4038864/v4l-saa7164-1.0.3-3.fw 
$ sudo cp v4l-saa7164-1.0.3-3.fw /lib/firmware

$ sudo adduser <userid> video
$ sudo reboot

Create Channel.conf

$ sudo apt-get install dvb-apps
$ scan ./uk-CrystalPalace -o zap > ~/.mythtv/channels.conf

Scan

Now go to Mythbackend-Setup and perform a full scan.

Test

You should now have a working system, unfortunately I found that the programme guide, or EPG, was not updating.

Fixing EPG

(Thanks to “Neon Dusk” from the second Ubuntu forum link in the references below.)

Using channels.conf to get the channel information is OK, but sometimes it doesn’t fill in all the information required.

Make a note of your mysql password from /etc/mythtv/mysql.txt - there should be a field DBPassword

$ mysql -u mythtv -p
<enter mysql password>
use mythconverg;
select * from `dtv_multiplex`;

Copy the resulting output into a text editor, here was mine. Note the networkid is null.

6 rows in set (0.00 sec)

In mythtv-setup -> channel editor -> Edit transport Delete all existing transports (highlight transport then press d)

Go to channel scanner Select Scan Type: Full Scan (Tuned)

Use the values recorded from the sql statement and then do a scan. You’ll need to do this for the 6 muxes.

Hopefully this should find all the channels and if you repeat the sql commands you’ll see the networkid value filled in. This should then allow the EIT EPG to work

My resulting output was as attached:

9 rows in set (0.00 sec)

Test Again

Have a look at your programme guide, give it a few minutes - it isn’t instantaneous.

References

• http://parker1.co.uk/mythtv_ubuntu.php
• http://www.uluga.ubuntuforums.org/showthread.php?t=942403
  • http://www.uluga.ubuntuforums.org/showpost.php?p=9351505&postcount=271
• http://ubuntuforums.org/showpost.php?p=9292366&postcount=3
• http://www.zedlopez.com/strangeloopiness/2010/04/mythtv_on_ubuntu_lucid_gotchas.html
• http://www.kernellabs.com/blog/?p=721
• http://ubuntuforums.org/showthread.php?t=1167640
• http://www.hauppauge.co.uk/site/products/data_hvr2200mc.html
• http://www.mythtvtalk.com/forum/hardware/8096-there-status-hvr-2250-linux-driver-2.html#post46347
• http://www.linuxtv.org/wiki/index.php/Hauppauge_WinTV-HVR-2200
• http://www.mythtvtalk.com/forum/hardware/12229-hvr-2200-not-picking-up-all-channels.html
• http://ubuntuforums.org/archive/index.php/t-1410407.html

The tuner card is now working under MythTV to “Watch TV”; however the programme guide is still not updating correctly over EIT. In theory this should happen automatically after a full scan, but for some reason this seems not to be working. It may be that this is because I have simply not left the backend to idle for long enough for an EIT scan to be initiated.

More news when I have it.

I have updated my notes on the following page:

• Hauppauge HVR-2200

Fortunately it seems XPS is supported in Okular under KDE4 (which we don’t have yet), and can also be readily converted to PDF using the Ghostscript “gxps” tool. This has to be built from source, but proved relatively simple. As our users do not have access to the command line, I also had to also create a Konqueror service menu, to enable right-click Action “Convert XPS to PDF”.

I am pleased to say it all works beautifully, and I have typed up my instructions for Debian Lenny here:

• [Howto

  Convert XPS to PDF](/2013/03/12/convert-xps-to-pdf/)

Installing from Ubuntu onto Debian Lenny seems sub-optimal, but is quick and dirty.

Step 1 - Add sources

# deb http://ppa.launchpad.net/freenx-team/ppa/ubuntu intrepid main
# deb-src http://ppa.launchpad.net/freenx-team/ppa/ubuntu intrepid main

Step 2 - Install FreeNX in Debian

# aptitude update
# aptitude install freenx

Step 3 - Connecting

To connect, you need to download the NoMachine client. If you have issues with the arrow/cursor keys, please see:

• Fix nxclient arrow keys

References

• FreeNX
• FreeNX on Debian Wiki ← Read this
• FreeNX on DebianHelp
• FreeNX on LTSP
  • FreeNX
• NoMachine

I recently that the arrow keys on my laptop were not working correctly, usually triggering knapshot. The following webpage addresses this concern. You should go there now, only read on if that page is gone.

• http://wiki.laptop.org/go/Remote_Desktop

Keyboard mapping

The UK keyboard layout of my laptop is not correctly mapped by the nxclient (which sees it as pc102 instead of pc105). For instance, on a KDE session, the arrow keys are not working, and the Up key thinks it is a PrntScr key and starts ksnapshot! Not really handy… Here is a manual workaround.

When connected remotely to the server, open a terminal and register this keymap configuration with X.

Please note that these instructions are for the XO and may not be correct - I will correct the following as/when I have time or find problems.

$ xmodmap xmodmap.keys

Content of the xmodmap.keys file
keycode   8 =
keycode   9 = Escape
keycode  10 = 1 exclam exclamdown exclamdown exclamdown exclamdown
keycode  11 = 2 at notsign notsign notsign notsign
keycode  12 = 3 numbersign U0300 U0300 U0300 U0300
keycode  13 = 4 dollar U0301 U0301 U0301 U0301
keycode  14 = 5 percent U0306 U0306 U0306 U0306
keycode  15 = 6 asciicircum U030A U030A U030A U030A
keycode  16 = 7 ampersand U0302 U0302 U0302 U0302
keycode  17 = 8 asterisk U0324 U0324 U0324 U0324
keycode  18 = 9 parenleft U0307 U0307 U0307 U0307
keycode  19 = 0 parenright U0308 U0308 U0308 U0308
keycode  20 = minus underscore U0304 U0304 U0304 U0304
keycode  21 = equal plus U0303 U0303 U0303 U0303
keycode  22 = BackSpace Terminate_Server
keycode  23 = Tab ISO_Left_Tab
keycode  24 = q Q Greek_omega Greek_OMEGA Greek_omega Greek_OMEGA
keycode  25 = w W oslash Oslash oslash Oslash
keycode  26 = e E oe OE oe OE
keycode  27 = r R U0327 U0327 U0327 U0327
keycode  28 = t T U032E U032E U032E U032E
keycode  29 = y Y U0325 U0325 U0325 U0325
keycode  30 = u U U032D U032D U032D U032D
keycode  31 = i I U032C U032C U032C U032C
keycode  32 = o O U0323 U0323 U0323 U0323
keycode  33 = p P U0304 U0304 U0304 U0304
keycode  34 = bracketleft braceleft U0331 U0331 U0331 U0331
keycode  35 = bracketright braceright U0330 U0330 U0330 U0330
keycode  36 = Return
keycode  37 = Control_L
keycode  38 = a A ae AE ae AE
keycode  39 = s S ssharp ssharp ssharp ssharp
keycode  40 = d D eth ETH eth ETH
keycode  41 = f F thorn THORN thorn THORN
keycode  42 = g G
keycode  43 = h H sterling sterling sterling sterling
keycode  44 = j J EuroSign EuroSign EuroSign EuroSign
keycode  45 = k K
keycode  46 = l L
keycode  47 = semicolon colon masculine ordfeminine masculine ordfeminine
keycode  48 = apostrophe quotedbl currency currency currency currency
keycode  49 = grave asciitilde dead_grave dead_tilde dead_grave dead_tilde
keycode  50 = Shift_L
keycode  51 = backslash bar section section section section
keycode  52 = z Z
keycode  53 = x X
keycode  54 = c C ccedilla Ccedilla ccedilla Ccedilla
keycode  55 = v V
keycode  56 = b B
keycode  57 = n N ntilde Ntilde ntilde Ntilde
keycode  58 = m M mu mu mu mu
keycode  59 = comma less guillemotleft guillemotleft guillemotleft guillemotleft
keycode  60 = period greater guillemotright guillemotright guillemotright guillemotright
keycode  61 = slash question questiondown questiondown questiondown questiondown
keycode  62 = Shift_R
keycode  63 = KP_Multiply XF86_ClearGrab
keycode  64 = Alt_L Meta_L
keycode  65 = space
keycode  66 = Caps_Lock
keycode  67 = F1 XF86_Switch_VT_1
keycode  68 = F2 XF86_Switch_VT_2
keycode  69 = F3 XF86_Switch_VT_3
keycode  70 = F4 XF86_Switch_VT_4
keycode  71 = F5 XF86_Switch_VT_5
keycode  72 = F6 XF86_Switch_VT_6
keycode  73 = F7 XF86_Switch_VT_7
keycode  74 = F8 XF86_Switch_VT_8
keycode  75 = F9 XF86_Switch_VT_9
keycode  76 = F10 XF86_Switch_VT_10
keycode  77 = Num_Lock Pointer_EnableKeys
keycode  78 = Scroll_Lock
keycode  79 = KP_Home KP_7
keycode  80 = KP_Up KP_8
keycode  81 = KP_Prior KP_9
keycode  82 = KP_Subtract XF86_Prev_VMode
keycode  83 = KP_Left KP_4
keycode  84 = KP_Begin KP_5
keycode  85 = KP_Right KP_6
keycode  86 = KP_Add XF86_Next_VMode
keycode  87 = KP_End KP_1
keycode  88 = KP_Down KP_2
keycode  89 = KP_Next KP_3
keycode  90 = KP_Insert KP_0
keycode  91 = KP_Delete KP_Decimal
keycode  92 = ISO_Level3_Shift
keycode  93 =
keycode  94 = less greater bar brokenbar bar brokenbar
keycode  95 = F11 XF86_Switch_VT_11
keycode  96 = F12 XF86_Switch_VT_12
keycode  97 =
keycode  98 =
keycode  99 =
keycode 100 =
keycode 101 =
keycode 102 =
keycode 103 =
keycode 104 = KP_Enter
keycode 105 = Control_R
keycode 106 = KP_Divide XF86_Ungrab
keycode 107 = Print Sys_Req
keycode 108 = ISO_Level3_Shift
keycode 109 =
keycode 110 = Home
keycode 111 = Up
keycode 112 = Prior
keycode 113 = Left
keycode 114 = Right
keycode 115 = End
keycode 116 = Down
keycode 117 = Next
keycode 118 = Insert
keycode 119 = Delete
keycode 120 =
keycode 121 =
keycode 122 =
keycode 123 =
keycode 124 =
keycode 125 = KP_Equal
keycode 126 =
keycode 127 = Pause Break
keycode 128 =
keycode 129 =
keycode 130 =
keycode 131 =
keycode 132 =
keycode 133 = Super_L
keycode 134 = Super_R
keycode 135 = Menu
keycode 136 =
keycode 137 =
keycode 138 =
keycode 139 =
keycode 140 =
keycode 141 =
keycode 142 =
keycode 143 =
keycode 144 =
keycode 145 =
keycode 146 =
keycode 147 =
keycode 148 =
keycode 149 =
keycode 150 =
keycode 151 =
keycode 152 =
keycode 153 =
keycode 154 =
keycode 155 =
keycode 156 =
keycode 157 =
keycode 158 =
keycode 159 =
keycode 160 =
keycode 161 =
keycode 162 =
keycode 163 =
keycode 164 =
keycode 165 =
keycode 166 =
keycode 167 =
keycode 168 =
keycode 169 =
keycode 170 =
keycode 171 =
keycode 172 =
keycode 173 =
keycode 174 =
keycode 175 =
keycode 176 =
keycode 177 =
keycode 178 =
keycode 179 =
keycode 180 =
keycode 181 =
keycode 182 =
keycode 183 =
keycode 184 =
keycode 185 =
keycode 186 =
keycode 187 =
keycode 188 =
keycode 189 =
keycode 190 =
keycode 191 =
keycode 192 =
keycode 193 =
keycode 194 =
keycode 195 =
keycode 196 =
keycode 197 =
keycode 198 =
keycode 199 =
keycode 200 =
keycode 201 =
keycode 202 =
keycode 203 = Mode_switch
keycode 204 = NoSymbol Alt_L
keycode 205 = NoSymbol Meta_L
keycode 206 = NoSymbol Super_L
keycode 207 = NoSymbol Hyper_L
keycode 208 =
keycode 209 =
keycode 210 =
keycode 211 =
keycode 212 =
keycode 213 =
keycode 214 =
keycode 215 =
keycode 216 =
keycode 217 =
keycode 218 =
keycode 219 = multiply division ISO_Next_Group ISO_Prev_Group ISO_Next_Group ISO_Prev_Group
keycode 220 = XF86Search
keycode 221 = XF86Meeting
keycode 222 =
keycode 223 =
keycode 224 =
keycode 225 =
keycode 226 =
keycode 227 =
keycode 228 =
keycode 229 =
keycode 230 =
keycode 231 =
keycode 232 =
keycode 233 =
keycode 234 =
keycode 235 = XF86Display
keycode 236 =
keycode 237 =
keycode 238 =
keycode 239 =
keycode 240 =
keycode 241 =
keycode 242 =
keycode 243 =
keycode 244 =
keycode 245 =
keycode 246 =
keycode 247 =
keycode 248 =
keycode 249 =
keycode 250 =
keycode 251 =
keycode 252 =
keycode 253 =
keycode 254 =
keycode 255 =

Now, all arrows keys can be used, but the Up one. Go into KDE Control Center. In the Regional & Accessibility/Input Action pane, look for the Preset Action/PrintScreen. In the General tab, click on the Disable check box.

This is just a list of the hacks or fixes that I currently have in place on Iceweasel. These were not my invention, and surely I owe a debt of thanks to various sites; unfortunately I have no way of knowing which sites.

Some of these hacks are too facilitate system wide control, others help to maximise performance on a thin client system.

Iceweasel Global Preferences

/etc/iceweasel/pref/iceweasel.js

// This is the Debian specific preferences file for Iceweasel
// You can make any change in here, it is the purpose of this file.
// You can, with this file and all files present in the
// /etc/iceweasel/pref directory, override any preference that is
// present in /usr/lib/iceweasel/defaults/preferences directory.
// While your changes will be kept on upgrade if you modify files in
// /etc/iceweasel/pref, please note that they won't be kept if you
// do make your changes in /usr/lib/iceweasel/defaults/preferences.
//
// Note that lockPref is allowed in these preferences files if you
// don't want users to be able to override some preferences.

lockPref("extensions.update.enabled", true);

// Use LANG environment variable to choose locale
lockPref("intl.locale.matchOS", true);

// Disable default browser checking.
lockPref("browser.shell.checkDefaultBrowser", false);

// ========= LTSP SPECIFIC HACKS ========

// Disable "safe browsing" feature that hogs CPU, HDD, etc. etc. in LTSP
lockPref("browser.safebrowsing.enabled", false);
lockPref("browser.safebrowsing.malware.enabled", false);

// Disable annoying location-bar suggestion "feature" that is sludgy with thin-clients
lockPref("browser.urlbar.maxRichResults", 0);

// Disable offline/disk caching of web pages
lockPref("browser.cache.offline.enable", false);
lockPref("browser.cache.disk.enable", false);
lockPref("browser.cache.memory.enable", false);

// ========= PROXY HACKS =========

lockPref("network.proxy.type", 1);
lockPref("network.proxy.share_proxy_settings", true);
lockPref("network.proxy.http", "192.168.1.254");
lockPref("network.proxy.http_port", 800);
lockPref("network.proxy.no_proxies_on", "localhost, 127.0.0.1, 192.168.1.0/24");

// ========= MISC HACKS ==========

lockPref("browser.startup.homepage", "http://www.example.co.uk/");
lockPref("app.update.enabled", false);

Iceweasel Image Caching

I understand that, on websites with certain size of image, that the thin client xorg starts to take a ridiculous amount of memory, resulting xorg dieing and the thin client returning to the login page. The following may help:

$ sudo echo "MOZ_DISABLE_IMAGE_OPTIMIZE=1" >> /etc/profile

References

I must apologise for the lack of references, if you recognise some of your work in the above, please do accept my apologies.

This is currently a few scrappy notes just to jog my memory as to things to try. Messing about with monitor settings can blow your monitor, graphics card and probably both; although in practice sensible use should ensure that never happens to you.

Note that some of these commands may not exist on your system, if you need to install them, then remember to install into the chroot, e.g.:

# chroot /opt/ltsp/i386
# apt-get update
# apt-get install lshw
# exit

Erm, suppose I should mention this whole page was tested on Debian Lenny.

Cables

Do consider trying a new monitor cable - as a poor quality or damaged cable can cause problems that are difficult to trace. A few seconds to try a different cable might save you endless time.

A word about drivers

There are often multiple drivers for a particular graphics card, for example with nvidia cards you can sometimes use the opensource “nv” driver and sometimes the proprietary “nvidia” driver, and I believe there is a “nouveau” driver, which I have not yet experienced. Similarly in intel, stipulating that the driver should be “intel” may solve your problem. See lspci below for details.

X Authority

All of the following commands were run from an ssh session to the thin client itself, for this to work you need to install openssh-server in the chroot. Alternatively you can simply go to the client and type ctrl-alt-f1 to access to the client’s console and type the following commands from there. Obviously the benefit to the ssh approach is that you can do it remotely and cut and paste from your browser into the client.

We need to be able to access the xserver from the console, so we need to set the XAUTHORITY:

# export XAUTHORITY=`find /var/run -name Xauthority`
# echo $XAUTHORITY
/var/run/ldm-xauth-qOQkpTdNp/Xauthority
# export DISPLAY=:7
#

Useful Programs

lspci

In order to google your problems, you’ll need to know the make and model of your graphics card, to find this out, run the lspci command:

# lspci
# lspci -v | more

The first command is a summary, the second command gives you more information. The “VGA compatible controller” is probably for what you’re looking.

lshw

List Hardware will list all the hardware on your PC.

# lshw -C display

getltscfg

Report all the settings from lts.conf which apply to this terminal. This helps diagnose those occasions where the client is refusing to obey your lts.conf commands.

# getltscfg -a

xvidtune

xvidtune -show will report the current monitor settings as a modeline:

# xvidtune -show

You can also run xvidtune as a localapp when logged into the thin client:

"/usr/bin/ltsp-localapps xvidtune"

xdpyinfo

xdpyinfo reports a great many details from the xserver, including resolution, colour depth etc:

# xdpyinfo | more

gtf cvt

There are two tools for creating a modeline - gtf and cvt. My understanding is that cvt is the newer tool, in use from 2003, but tbh I really don’t know which you should use. Both tools work in the same way. Some examples:

# gtf 1920 1080 60
# cvt 1280 1024 75

xrandr

To save rebooting endlessly you can test monitor settings on the fly with xrandr:

# xrandr
# xrandr --newmode "1920x1080_60.00"  172.80  1920 2040 2248 2576  1080 1081 1084 1118  -HSync +Vsync
# xrandr --addmode VGA "1920x1080_60.00"
# xrandr --output VGA --mode "1920x1080_60.00"
# xrandr --auto

The first command provides you with useful information about your xserver, including the available modes, and most importantly the available “outputs” (these are your monitors/displays); in the above example my only output is a monitor called “VGA”. newmode creates the newmode (e.g. as output by gtf or cvt), addmode makes that mode applicable to the output, and the next command activates that mode. The final command “auto” will reset your display automatically, probably your previous settings.

References

For once almost all of this was my own work, and so all my fault ;). That said the following page was helpful to me in the section about xrandr:

• https://wiki.ubuntu.com/X/Config/Resolution

• Hauppauge HVR-2200

Introduction

This is a guide to upgrading from Debian Etch and LTSP 4.2 to Debian Etch and LTSP 5. This is largely based on the following page which should be your primary guidance on installation:

• http://wiki.debian.org/LTSP/Howto

Now that Debian Lenny has become the stable release, you should install that instead, see the above document for guidance, and/or my LTSP5 Install page.

Warning

This page seems a little outdated, most importantly it refers to the LTSP Backports repository, which is now obsolete. You should instead visit http://backports.org/ and add the Debian Lenny repository. All references to alioth.debian.org are redundant. You are probably best off going to this page:

• http://wiki.debian.org/LTSP/Howto

Back-up /opt/ltsp

This will copy your entire /opt/ltsp to /opt/ltsp42.

# cd /opt
# mkdir ltsp42
# cd ltsp
# tar cf - * | (cd /opt/ltsp42; tar xfp - )

Backup other configurations

# cd /opt/ltsp42
# mkdir conf
# cp /etc/ltsp.conf /opt/ltsp42/conf/
# cp /etc/ltsp-localdev.conf /opt/ltsp42/conf/
# cp /etc/dhcp3/dhcpd.conf /opt/ltsp42/conf/

This list is not exhaustive - if you can think of any other configurations outside of /opt/ltsp - back them up!

Backup tftpboot

# cd /opt/ltsp42
# mkdir tftpboot
# cd /tftpboot
# tar cf - * | (cd /opt/ltsp42/tftpboot; tar xfp - )

Remove lbe if installed

# cd /usr/local
# rm -r lbe

Add Backports Repository

This section is outdated - the latest packages are simply in Debian Backports for Lenny.

Given that you have chosen to upgrade to LTSP5, it is probably that you want to have the new technologies, which are currently only available in Backports. If you already have Backports listed in /etc/apt/sources.list then you may skip this stage.

Add the key(s) for the repository to your keyring(after verifying that you trust them)

# wget http://pkg-ltsp.alioth.debian.org/debian/pkg-ltsp-keyring
# apt-key add pkg-ltsp-keyring 

Add to /etc/apt/sources.list

deb http://pkg-ltsp.alioth.debian.org/debian etch-ltsp-backports main

Configure pinning in /etc/apt/preferences

If you do not have a preferences file, then an example one might be:

Package: ltsp-server
Pin: origin pkg-ltsp.alioth.debian.org
Pin-Priority: 998

Package: ltsp-server-standalone
Pin: origin pkg-ltsp.alioth.debian.org
Pin-Priority: 998

Package: ltspfs
Pin: origin pkg-ltsp.alioth.debian.org
Pin-Priority: 998

Package: *
Pin: release o=Debian,a=stable
Pin-Priority: 900

Package: *
Pin: origin www.backports.org
Pin-Priority: 700

Package: *
Pin: release o=Debian,a=testing
Pin-Priority: 500

Package: *
Pin: release o=Debian,a=unstable
Pin-Priority: 300

Package: *
Pin: release o=Debian
Pin-Priority: -1

Apt-get Update:

# apt-get update

How you are logged in?

You should be logged in either directly to the console, or via ssh (or perhaps FreeNX). If you are logged in from an LTSP Client, then there is a likelihood that your client will be terminated during installation.

The same clearly applies to your users. In any case some of these processes are CPU intensive (e.g. ltsp-build-client) and it’s a good idea to have everyone off the system (this may even be a necessity).

Remove Old LTSP Packages

# apt-get --purge remove ltsp-utils atftpd
# cd /opt
# rm -r ltsp

Install new packages

# apt-get -t 'etch-ltsp-backports' install ltsp-server-standalone ltsp-server
# dpkg -l ltsp-server | awk '/^ii/{print $2" "$3}'

Make sure the version ltsp-server is 5.0.40~ or greater.

Build the LTSP client environment

The ltsp-build-client script needs a lot of time if you have an older server.

#  ltsp-build-client --extra-mirror "http://pkg-ltsp.alioth.debian.org/debian etch-ltsp-backports main" --apt-key /etc/apt/trusted.gpg

Don’t be tempted to use a simple ltsp-build-client command, as you need to specify backports repository (as shown above).

The ltsp-build-client command will download a complete Debian filesystem into /opt/ltsp/i386 (or specify an alternate location with ‘–base’) and install the ltsp-client and ldm packages (the LTSP Display Manager).

The 2 ltsp configuration files inside the client are /etc/lts.conf and /etc/default/ltsp-client-setup. See the examples in /usr/share/doc/ltsp-client*. Also see the Edubuntu wiki.

Debian LTSP still uses NFS by default.

If you change the IP data after you have done the initial setup and run ltsp-update-sshkeys on the server.

The files the client will boot are installed on the server into /var/lib/tftpboot, unlike LTSP4.2 which installed in /tftpboot.

Configure DHCP Server

Edit /etc/dhcp3/dhcpd.conf.

As you already had DHCP working, under LTSP4.2, then it will just need some modification. In order to configure your dhcpd.conf, you should refer to examples in /usr/share/doc/ltsp-server/examples/dhcpd.conf or /etc/ltsp/dhcpd.conf and adjust for your network.

Specifically, you need to make the following changes:

Root Path

Take the IP address out of your root-path statement as LTSP5 doesn’t support that style of root-path

option root-path  "/opt/ltsp/i386";

Filename

Change the filename entries to:

filename   "/ltsp/i386/pxelinux.0";

Next Server

Etch is still affected by: http://bugs.debian.org/416868 so you need to add this line if the dhcp server and tftp server are the same:

next-server ip.address.of.server;

Restart dhcp3-server

# /etc/init.d/dhcp3-server restart

Watch for errors, if it will not restart then correct errors and try again.

Configure Exports

Edit /etc/exports to include the following line:

/opt/ltsp *(ro,no_root_squash,async,no_subtree_check)

Then restart the NFS Kernel Server:

# /etc/init.d/nfs-kernel-server restart

Start tftpd

By default, tftpd-hpa is started from inetd, so you may need to restart inetd after installing tftpd-hpa.

# /etc/init.d/openbsd-inetd restart

Alternately, edit /etc/default/tftpd-hpa to have tftpd-hpa start on it’s own.

RUN_DAEMON="yes"

And then restart with:

# /etc/init.d/tftpd-hpa restart

Test boot a client

Boot a PXE or Etherboot capable machine and enjoy!

Changing the Client

See the following page for details: http://doc.ubuntu.com/edubuntu/edubuntu/handbook/C/customizing-thin-client.html

# sudo chroot /opt/ltsp/i386
# apt-get install rdesktop
# sudo ltsp-update-image

Setting root password for client

Screen scripts are how LTSP determines what type of login will run on what virtual screen. Most GNU/Linux machines have 12 virtual consoles, which you can access by pressing Control-Alt-F1, through Control-Alt-F12. There is a text based getty that is started on screen 1, but you normally can’t log into it, as there are no local users on the thin client.

However, for debugging purposes, you may want to set up root to log in on the thin client. You may need to do this if you’re debugging problems with local devices, for example. Fortunately, it’s easy to do: on the server, just chroot into the LTSP chroot, and set the password with passwd.

sudo chroot /opt/ltsp/i386
passwd

Further information

• LTSP5 Sound
• LTSP5 Local Devices

References

• http://wiki.debian.org/LTSP/Howto
  • http://doc.ubuntu.com/edubuntu/edubuntu/handbook/C/customizing-thin-client.html
• http://www.mail-archive.com/ltsp-discuss@lists.sourceforge.net/msg33415.html
• http://www.linuxagora.com/vbforum/showthread.php?t=749
• http://www.ltsp.org/twiki/bin/view/Ltsp/Debian#LTSP_5
• http://marc.info/?l=ltsp-discuss&w=2&r=1&s=ltsp5+debian+etch&q=b

Introduction

The selection of client hardware is an interesting topic, that causes much confusion for those unfamiliar with LTSP. When you start out with LTSP, the likelihood is that one of your motivations is to extend the life of your existing clients. But you’d be lucky indeed not to require a few new clients, and of course over time you will need to purchase replacements for hardware failures. So what are your options?

Options

• Existing PCs
• Thin clients designed for LTSP
• Thin clients designed for Citrix
• Barebones PCs
• Self-build PCs
• New PCs

Existing PCs

Well you obviously don’t want to throw out all your existing hardware, but do bear in mind that the electrical cost is likely to be around £40 a year, and so a replacement that brings this down to £15 a year, might be easier to justify than you might imagine. If you do want to keep your existing PCs, then do consider unplugging the hard drives. Hard drives are not required on LTSP clients and they consume unnecessary power. They also can often be noisy, unplugging an old drive can miraculously silence an annoyingly loud older PC.

Thin clients designed for LTSP

I have no experience of this route. One option is to purchase from Diskless Workstations, who support the LTSP project.

• http://www.disklessworkstations.com/

Thin clients designed for Citrix

I have had mixed success with thin clients that were designed for Citrix and/or Windows Terminal Services. The Wyse terminals tend to be very sluggish and quirky, and I find that I cannot recommend them for LTSP.

HP thin clients seem to work very much better, and are certainly an option. Such devices tend to be very low power. On the downside, they do tend to be surprisingly expensive, for what they provide. That said, if you have a good hardware budget, I believe you could do very much worse than buy some HP Linux Thin Clients but, as with any client hardware, do buy one and test before committing yourself to considerable expenditure.

Barebones PCs

Obviously the above is probably not true of dedicated LTSP thin client hardware, designed from the ground up to work with LTSP.

I have found that the best client hardware is to purchase low-cost barebones PCs, low-power and preferably fanless. I tend to only buy barebones PCs that include the CPU, in order to keep down the overall cost and keep commissioning time to a minimum. Previously this meant Via C7 CPUs, which haven’t the best reputation; although so far I have not had any failures. Nowadays I think you’d be searching for “intel atom barebones”.

Yes, you need to open the case, add some RAM, and maybe add an optical drive - perhaps 10 minutes work - the flipside is that you’re not paying for an operating system, and you’re not paying for a hard drive (which you don’t need).

Such barebones PCs currently cost about £125 (plus VAT), seem much less likely to have compatibility issues, perform much faster than dedicated “thin client” devices, and could easily be changed to fat clients, should your requirements subsequently change.

Self-build PCs

I have had great success building small Mini-ITX form factor PCs, on Via CPUs. The hardware was all purchased from LinITX.com, who I found very helpful. Build time was about an hour for the first, but this dropped to about 30 minutes for the remaining ones. The result are stunningly attractive tiny PCs which grace our showroom. Just do concentrate on low-power consumption and look where possible for “fanless”. LTSP uses very little of the local CPU, so heat is unlikely to be an issue.

New PCs

Full blown desktops are, I think, a poor option. These tend to be power hungry, and come with a Windows licence, hard drive and other bits that you will not need. There are a new generation of tiny PCs like the Acer Reevo and the Asus EeeBox, which may be a good option.

References

There are no references for this page, which was written without reference to other information.

As a died-in-the-wool KDE user, I’m just coming to terms with KDE4 and mostly loving it. A few niggles remain, but in the main it is very good indeed. But I suppose I would hesitate to recommend it to a new user. Having seen and heard much praise for the latest Ubuntu 10.04 Lucid Lynx, which uses the Gnome desktop, I thought I’d give it a go.

I started by downloading the ISO image from the Ubuntu website and followed my instructions to Make A Linux USB Key, basically by formatting in super-floppy format (which I find to be the most consistently successful method) and using Unetbootin to write the ISO to the USB Key. Using a USB key is more convenient, more portable and more environmentally friendly than burning a CD/DVD.

I took a spare laptop, an HP NX7300, and inserted the USB Key. I entered the BIOS settings and changed the boot order to boot first from USB. I then booted up the laptop.

HP are not know for their Linux compatibility, but Ubuntu booted up with an attractive boot splash, before logging straight into the very attractively themed Gnome desktop. The new colour scheme is purple, which is very attractive, although brown still exists in the OpenOffice splash screen to remind us of what used to be.

As expected, network manager failed to activate, and I knew from past experience that I would need to load the proprietary broadcom driver. Continuing to play it dumb, I waited to see if I would get any help from Ubuntu. After about half a minute, the restricted drivers’ manager appeared on the top panel, although there was no guidance as to why it had appeared, nor what to do with it. I clicked on it and then clicked on the option to install the Broadcom STR driver and clicked activate.

After a minute or so it had loaded the Broadcom driver from the USB stick and I noticed that the wireless light had illuminated on the laptop. Promising! Unfortunately it then suggested I reboot to load the new driver; which obviously I can’t do, as I am running off a USB stick and would effectively start from scratch again if I rebooted. I decided to just wait and, sure enough, after a while when I clicked on the wireless symbol on the top panel it offered my local wireless access point. I selected it, and entered the encryption key, and I was connected.

At this point I had a completely usable desktop environment, including a full office suite (OpenOffice.org), a working web browser (Firefox) and everything else you could possibly wish in a modern desktop environment, all working off a USB stick. I still find that just amazing.

I must admit that I am surprised by Gnome, it looks gorgeous, it feels very integrated, as if it had been designed as a whole, rather than being a collection of applications brought together. No it’s not my much loved Debian, and no it’s not my much loved KDE, but it does feel easy to use, very uniform and just… well “right”.

I was particularly impressed by Rhythmbox, which had the built in music store, podcast reader, and local collection. This does look like an application which would reassure an erstwhile iTunes user. It’s probably not as good as amaroK, but having pushed amaroK on my mother and regretted it mightily, I think Rhythmbox is a good choice, being infinitely more user friendly.

I even set-up my Gmail account in Evolution, simply by clicking on the mail icon in the top panel and following the wizard to set up my Gmail account. I played completely dumb, it provided all the necessary server information and network settings for me, and all I needed to know was my Gmail address and password. That is incredible, anyone could set up their email if it’s that easy.

Having tested it, just to decide whether to recommend it to friends and family, I find myself sorely tempted to install it on my own laptop.

After conducting further research, it seems that Nochex itself does not have the best reputation.

Further to this, I contacted them again, and once again they have failed to respond within 24 hours.

• Techcrunch: Google Kills Its Homepage Background Image Experiment Early

Now those that want pictures of kittens as their background can and the rest of us can keep our plain simple and fast search page. After spending the day using ixquick, I’ll be glad to return to google.com, which is still the best search engine that I have ever used.

When are we going to see a decent open source search engine, maybe distributed in some way amongst its user base?

And they have.

I can only imagine Google are wanting to reduce their monopoly and increase competition in the search engine market, and perhaps that is to be applauded.

Okay, so what can I use instead of Google? I think I’ll try Bing, supposedly their privacy policy is better as well.

Cendio SeamlessRDP

The Cendio SeamlessRDP component has two halves, a Windows executable as the “server component” for installation on the Windows server, and patches to the rdesktop application, to work with the server component. At the time of writing, these Cendio patches have now been merged into the rdesktop 1.5.0 and later core.

• You can download the source using Subversion.
• You can get a binary and unzip the files into c:\seamlessrdp (or similar).

Once you have a binary on your Windows server, then from you Linux client, type:

$ rdesktop -A -s "c:\seamlessrdp\seamlessrdpshell.exe notepad"

Under W2K8, both seamlessrdpshell.exe, and the Windows application you wish to run with it, must be published terminal services applications. And seamlessrdpshell.exe must have command line parameters enabled, by running “TS RemoteApp Manager’s” “Add RemoteAppProgram” dialog (remoteprograms.msc).

Fontis SeamlessRDP

Fontis have further developed Cendio SeamlessRDP, to enable connection sharing, allowing a single rdesktop connection to launch multiple applications. When run in seamless mode, rdesktop creates and listens on a control socket. A new option allows rdesktop to be run in slave mode, which notifies the master rdesktop instance of a new command to be run and then exits. The master instance sends a client-to-server message to the SeamlessRDP server component, which runs the new command.

I see this development as fairly fundamental to the use of SeamlessRDP.

Installation of Fontis patch

Download source for the latest version of rdesktop from CVS (press enter when prompted for a password):

$ cvs -d:pserver:anonymous@rdesktop.cvs.sourceforge.net:/cvsroot/rdesktop login
$ cvs -z3 -d:pserver:anonymous@rdesktop.cvs.sourceforge.net:/cvsroot/rdesktop co -P rdesktop

Alternatively, download the rdesktop CVS snapshot which has been tested with the patch. Download the rdesktop patch to the checked out directory:

$ cd rdesktop $ wget http://www.fontis.com.au/system/files/rdesktop.patch

Apply the patch and compile:

$ patch -p1 < rdesktop.patch
$ ./bootstrap
$ ./configure
$ make

Download updated server component and unpack. Alternatively, obtain the seamlessrdp CVS snapshot or download direct from CVS, apply the seamlessrdp patch and then compile:

$ cvs -z3 -d:pserver:anonymous@rdesktop.cvs.sourceforge.net:/cvsroot/rdesktop co -P seamlessrdp
$ patch -p1 < seamlessrdp.patch

Usage

The following examples assume the SeamlessRDP server component files (seamlessrdpshell.exe, seamlessrdp.dll and vchannel.dll) have been unpacked to C:\seamlessrdp. To start a SeamlessRDP session, invoke rdesktop in the same way as before applying the patch:

$ ./rdesktop -A -s "C:\seamlessrdp\seamlessrdpshell.exe notepad" server

To run a new command inside the existing SeamlessRDP session:

$ ./rdesktop -l "calc"

New command-line options:

• -M: Specify the path for the control socket that the rdesktop process listens on. By default, this is $HOME/.rdesktop/seamless.socket
• -l: Instead of starting a new rdesktop process, connect to an existing process’ control socket and tell it to run a command on the server. Can be combined with the -M option to use a non-standard socket.

Wow, this really is a second post in one day!

Following on from the DPI issue, I decided to throw science to the wind and just set DPI to 96x96 rather than 50x50, and the login screen looks better and the font size is improved (slightly larger).

Next, I decided that I ought to check that the TV Tuner was properly detected, but whilst it was detected correctly, it was unsure of the card number, and was unable to download the firmware. This is not entirely surprising as this TV card is fairly recent and (as usual) Hauppauge do not themselves support Linux.

This link provided me with the basis for fixing this issue, but in the event I only needed the following:

$ cd ~
$ wget http://www.steventoth.net/linux/hvr22xx/22xxdrv_27086.zip
$ wget http://www.steventoth.net/linux/hvr22xx/HVR-12x0-14x0-17x0_1_25_25271_WHQL.zip
$ wget http://www.steventoth.net/linux/hvr22xx/extract.sh
$ sh extract.sh
$ sudo su
# cp *.fw /lib/firmware/`uname -r`/
# echo "options saa7164 card=4" >> /etc/modprobe.d/options.conf
# exit

Then reboot. Hopefully tomorrow I will have some time to continue setting up the TV card under MythTV; but for now the aerial input is required by our old MythTV system in order to record Doctor Who :).

The DPI problem was fixed very easily by following the instructions on the link I provided in my previous post, I calculated that my Sony Bravia KDL 32S3000 had dimensions of 700mm x 390mm and a native resolution of 1366x768.

$ xdpyinfo | grep dimensions
$ xdpyinfo | grep "dots per inch"

Converting the millimetres to inches and dividing the resolution by the inches dimensions gave me DPI of 50x50, surprisingly low. I believe this is because the KDL 32S3000 is not a full HD TV but only 1080i. Setting this in /etc/X11/xorg.conf cured the problem with the KDM login manager.

Section "Monitor"
    Identifier "Monitor0"
    Option   "DPI" "50 x 50"
EndSection

Installed the Nvidia binary driver, which worked fine, except that, as I’ve found on a few installs recently the KDE screen DPI defaults to something absurd, resulting in unreadably small fonts. Easily fixed in KDE, via System Settings (as long as you know your way around!), but haven’t yet found out how to do it in KDM (the login manager). This is not a huge problem, as it is easy enough to type in your password, and anyway on a home theatre PC you probably want it automatically logging in. That said, I am determined to fix this problem when I have some time. Interestingly this was also the case on our existing mythbox, which has completely different hardware, the common thread being the nvidia driver. I suspect the issue may be related to the KDE virtual desktop, and the nvidia driver calculating this back to a reasonable DPI. This link may help, I will give this a go when I have some time.

Anyway, the good news is that, with the binary driver installed, the standard KDE compositing is working beautifully.

As to sound, this is still not working via HDMI, but is working fine through the sound outputs. This seems to be a widespread issue, which I think has been resolved, so it’s just a matter of waiting for the fix to filter downstream. In the meantime, I will continue to work as we do on our existing mythbox, with a 2.1 speaker system plugged into the HTPC.

It switched on okay, but the front LEDs were not working, neither the power nor the disk LED; so it seems that they are not wired up correctly, although I really can’t see how I could have got it wrong.

It booted into a Kubuntu 10.04 USB stick without any problems. The network card was detected correctly and I was able to browse the Internet without any configuration. The Nvidia card was detected correctly, using the open source driver, obviously I will install the Nvidia binary driver. The only strangeness was the lack of sound via the HDMI cable, but looking in the BIOS afterward I noticed an option “Audio via HDMI” which I have enabled, it was set to auto, but maybe that’ll make a difference.

It is difficult to say how quiet it is at present, as any sound it was making was being drowned out by our noisy old mythbox.

I can’t wait to have a decent chunk of time to start the installation.

The Intel Atom A330 ION motherboard had two prongs for the Power LED, but the Lian-Li PC-C37 case had a cable for a three prong connector. The cable on has two wires, so I got around this issue by levering up the plastic catch holding one of the cables and moved it into the middle connector, which then fitted the motherboard fine, just with the now vacant connector overlapping the connectors.

The second problem that I encountered was the question of which way up the PSU should go. Normally an ATX PSU sits at the top of a tower case with the fan downwards, pulling the hot air out of the case. In a desktop (horizontal) PC this is never going to be as effective. With my case, there is no space below the PSU, so that the fan would be drawing air from nowhere and exhausting it out the back of the PC. Not only will this not cool the case, it will also burn the PSU in a very short time. By mounting the PSU upside down, the cool air should be drawn in from outside, via the top vent, cooling the PSU and exhausting out the back. This does mean that the PSU is not helping to cool the case, but at least it is not adding heat into the case.

Unfortunately, mounting the PSU upside down has meant that the power cables are furthest from the motherboard, with the result that the main 24-pin lead does not reach. It seems that extension leads are available, and I’m to have a go at getting one from Maplin in Guildford today.

This lack of PSU fan cooling for the case does mean that we are very reliant on the two case fans. The case does allow for several more case fans if necessary, although the Intel Atom processor shouldn’t be generating too much heat.

Hopefully the following should be a slideshow of the build. I’m sorry to say that I didn’t take many interim photos, so this slideshow is rather a before/after:

These days building PCs is not a black art - the connectors are all labelled - and the motherboards generally have a helpful manual detailing exactly what goes where. Building a PC is primarily a challenge from the perspective of making sure that you buy compatible components and that you design in sufficient cooling.

With luck my next post will be the power-up and install of Kubuntu 10.04 with MythTV!

Just a reminder of what I was wanting for our HTPC:

• Low power
• Silent
• Powerful enough to run KDE fully
• DVI and HDMI interfaces
• Decent on-board sound
• Plenty of USB slots
• Needs to visually go with our HiFi amp and Blu-Ray player
• Powerful enough graphics to support 1080p (or 720p if not possible)

The Case: Lian Li PC-C37

I started, perversely, with the case. Normally you would start with the cpu/mb, but I knew that, in order to get spouse approval, it had to look right. You can spend from £20 to £1000 on cases, and I dare say more. I wanted something that looked like a black HiFi separate, both in dimensions and styling. I would have thought that there would have been a considerable choice, but there really wasn’t, the only one I found was the Lian-Li PC-C37. It would seem to be a well made quality case, for about £100. I would have liked to have spent a little less, but really there was no choice whatsover.

Motherboard/CPU: Asrock A330ION

The Lian-Li PC-C37 takes both Mini-ITX and Micro-ATX motherboards. Well that’s just fine, given that my aim was a low power, silent and powerful PC, really the Intel Atom was the only choice and they invariably come in Mini-ITX motherboards. Well there is a vast array of Intel Atom motherboards on the market but, oddly enough, almost all of them had problems. Most of them only had 4 usb ports, which would be okay, just. Almost none of them had DVI, let alone HDMI. One motherboard stood out from the rest, the Asrock A330ION. Yes it’s the old Atom 330, but it allowed the faster DDR3 RAM, and the specification of the motherboard was fantastic, including:

• 2 x DDR3 DIMM slots (up to 4GB)
• Integrated NVIDIA® GeForce 9 Series
• DVI and HDMI
• 7.1 channel HD Audio
• Gigabit LAN
• 1 x PCI Express 2.0 x16 slot
• 4 x SATAII 3.0 Gb/s connectors + 2 e-SATA
• 4 rear USB plus support for upto 4 front USB

That feature set is almost unbelievable, (4+2 SATA!! 4+4 USB!! DDR3!!!) by the time I had read the specification, I barely cared about the price.

I also bought 2gb of compatible DDR3 RAM.

DVB-T Tuner

Well, as I said in my last post, we need to have DVB-T, and of course I have my Hauppauge Nova-T500 PCI card. Ah, but I have just bought a motherboard that only has PCI-Express; so a new card is required. I do hope my wife isn’t reading this. Anyway, I like Hauppauge products, and the HVR-2200 is the obvious PCI-Express low profile dual DVB-T tuner card. No decision to make really, other than should I pay twice the amount for the HVR-4400, which has satellite capability, err no, have to draw the line somewhere ;).

Hard Drives

Well, we’re making a lower power device, so we need low-power drives. Two were recommended to me, the Seagate Barracuda LP Series and the Western Digital Caviar Green Series. Starting to get rather concerned about the money I have been spending, I decided to opt for a relatively low capacity of 500gb, but to go for two drives, so that I could configure a RAID1 array. For those that don’t know, a RAID1 array means that you have the same data on both drives, so that if you lose one drive, you don’t lose your data. In short, your data is mirrored to the two drives. This doesn’t replace backups, but it certainly saves a headache if you lose one of your drives.

It was suggested to me that I should use a NAS for storage, but I don’t particularly want two machines on almost permanently. I already have an Icybox NAS, which I have stopped using after it ate two drives; so I don’t trust it and I don’t particularly want it. We’ll probably configure our HTPC to behave as a NAS as well as an HTPC.

• Western Digital 500GB Caviar Green 32MB SATA 69T6

Conclusion

All the hardware is now ordered, with the exception of the PSU, as I have one in my stores somewhere, hopefully it will be compatible! More tomorrow, when hopefully I will be starting the build.

Those that say that Linux is free if your time is free have never spent evening after evening setting up XPMCE.

In 2008 we upgraded to Mythbuntu 8.10. My wife was easy to persuade, simply because she could not see how Linux could possibly be worse than Microsoft Windows XP Media Center Edition. And so it proved, it wasn’t perfect by any means, but it was just good enough that we never regretted the change.

Conventional wisdom is that, if you have a working MythTV system, then you should leave it well alone; but I decided that I wanted to move to KDE from the standard XFCE. It was only after the installation was well under-way that I realised that this meant KDE 4.1, which was an unusable version. Kicking myself for such stupidity, I quickly upgraded to 9.04 and 9.10. Both upgrades went almost flawlessly, and the system has been pretty good ever since.

Unfortunately we’ve come to hate the Evesham emedia station that runs our MythTV. I stupidly bought four PCs from Evesham and have always regretted it. Evesham have a good reputation, but my experience is that the hardware is very poor indeed. Specifically it has always been very noisy and suffered from overheating. In addition we have issues with USB support on this PC in Ubuntu. Something to do with OHCI, UCHI etc, I forget what exactly. It’s also ugly as sin. In short it was not a good buy and we have always regretted it.

If you’re interested you can read a review on this god-awful PC. Having read this very positive review, I can only assume that Evesham reduced the component quality significantly by the time I bought this machine.

So we’ve decided to replace our media centre. I have considered XBMC and Boxee, but with such a poor Internet connection, we really do still need to get our TV via our aerial. And that means MythTV, as far as I know.

I considered buying an Acer Revo, but a huge prejudice against Acer put me off this decision (oh god, not Acer, not ever). I did not consider for a second buying a dedicated XP MCE machine and putting MythTV onto it - we’ve experienced what can go wrong first hand - and I wanted 100% control over the components.

So what are we looking for in our new media centre (or Home Theatre PC as the jargon now has it!)?

• Low power
• Silent
• Powerful enough to run KDE fully
• DVI and HDMI interfaces
• Decent on-board sound
• Plenty of USB slots
• Needs to visually go with our HiFi amp and Blu-Ray player
• Powerful enough graphics to support 1080p (or 720p if not possible)

In my next post I will run through the hardware I’ve chosen.

The Shortcut

Try erasing the contents of your USB key, then skipping down to Install Unetbootin below. You could be lucky, and save yourself some fiddling about!

Step 1: Make sure that your USB key is unmounted

Open a terminal and type:

$ mount

This will list your current mounts, including your USB key (if mounted). Take a note of the mount point (e.g. /media/disk) and type:

$ sudo umount /media/disk

Once you are confident that the drive is not mounted, you may proceed.

Step 2: Determine the device

Find you USB key by typing:

$ sudo fdisk -l

Disk /dev/sdc: 64 MB, 64487424 bytes
4 heads, 32 sectors/track, 984 cylinders
Units = cylinders of 128 * 512 = 65536 bytes
Disk identifier: 0x5187ef1e

Device Boot      Start         End      Blocks   Id  System
/dev/sdc1   *           1         983       62896    6  FAT16

Step 3: Delete the old partitions

This will lose any data on this drive!!!

Ignore the number after the letter, so in the above example the device is /dev/sdc, not /dev/sdc1; the 1 is partition 1, but fdisk works on the whole drive, not just a single partition.

$ sudo fdisk /dev/sdX

Command (m for help):

• Type “p” and enter to view the existing partitions (generally only one)
• Type “d” and enter to delete the partition (if there are multiple partitions, you may need to enter 1 for the first)
• Type “p” to view the remaining partitions (there should not be any!)

At this point, nothing has really happened, you can “q” to quit and nothing will have been touched.

Step 4: Create the new partition

• Type “n” to make a new partition
• Type “p” to make this partition primary and press enter
• Type “1” (one) to make this the first partition and then press enter
• Press [enter] to accept the default first cylinder
• Press [enter] again to accept the default last cylinder
• Type “p” to view what you have done

This should display the current configuration of your drive:

Disk /dev/sdc: 64 MB, 64487424 bytes
4 heads, 32 sectors/track, 984 cylinders
Units = cylinders of 128 * 512 = 65536 bytes
Disk identifier: 0x5187ef1e

Device Boot      Start         End      Blocks   Id  System
/dev/sdc1               1         984       62960   83  Linux

Step 5: Change partition to W95 Fat32

• Type “t” to change type
• Type “b” for Win95 Fat32
• Type “p” to view what you have done

Step 6: Make partition bootable

• Type “a” to add bootable flag
• Type “1” (one) to select first partition
• Type “p” to view what you have done

Step 7: Write partition changes

Up to this point, nothing has actually been changed; if you are happy with the changes you have made, then:

• Type “w” to write the new partition information to the USB key

Step 8: Format the partition

The final step is to create the fat filesystem:

Note that this time we DO quote the partition number

$ sudo mkfs.vfat -F 32 /dev/sdx1

This process ends by sync’ing the disk.

Step 9: Mount the USB key

For most systems, this can be easily achieved by simply unplugging and replugging in the key, and taking option to Mount when prompted. Equally you can mount manually by typing:

Change “mountpoint” below for an appropriate name

$ sudo mkdir /mnt/mountpoint
$ sudo mount /dev/sdx1 /mnt/mountpoint

Step 9: Install Unetbootin

$ sudo aptitude install unetbootin

Step 10: Use Unetbootin

Run Unetbootin from your menu, either select a distribution and version, or take the option to install an already downloaded disk image.

Next select the usb drive.

Given the tiny size of my USB key, I am using slitaz, which is a minimal distribution.

Have a look at the contents of the USB drive, in at least one test, unetbootin had only copied a single file across. Just run unetbooting again if that happens.

Step 11: Reboot and Test

That’s it, you should now be able to reboot onto the drive.

Step 12: Troubleshooting

My first attempt was using FAT32, which did not work. Next I tried FAT16, which also did not work, then I tried EXT2, by change partition type to “83 Linux” and then typing the following command (replacing the “x” for your correct drive letter):

$ sudo mkfs.ext2 /dev/sdx1

This time the USB key was detected as USB-HDD and booted perfectly. It does seem that everyone’s has different experiences, and there is an element of trial and error to get things working.

If you have a usb key over 512mb and are using FAT32, then you might try reducing the size of the partition to 512k, and create a second partition with the remaining space.

See Alternative Methods for other options.

Step 13: Superfloppy

I understand that the problem with USB booting, is that many BIOS’s understand the superfloppy format - i.e. a formatted but unpartitioned drive. This is easy enough to create under Linux, simply delete the partitions in fdisk and format with:

# mkdosfs -I /dev/sdx

or possibly on older hardware:

# mkdosfs -v -F 16 -I /dev/sdx

The problem is that superfloppies will not be auto-mounted under Linux, with the result that Unetbootin cannot write to them. But you can mount superfloppies manually:

$ cd /media
$ sudo mkdir sdx
$ sudo mount /dev/sdx sdx

Then use unetbooting to write to /dev/sdx, and not forgetting to umount afterwards:

$ sudo sync
$ sudo umount sdx

Alternative Methods

Some distributions offer their own USB key creation software, including Ubuntu 8.10+, PuppyLinux 4.10+, and Slitaz, to name but a few.

Also, it is possible to obtain disk images that can be written to a usb key directly, thus circumventing these problems. For example:

• Debian Installer
• Debian Live

References

• Install a Live Ubuntu image on a USB stick
• PenDriveLinux
• Installing Debian Sarge from a USB memory stick
• Debian Wiki USB Boot
• Debian EeePC Install

Job control turned off nbd0 not found

These instructions do not apply for Ubuntu.

If you receive an error message stating initramfs “job control turned off”, then look further up for a line stating that /path/to/nbd0 not found. If you’re on Debian then this is an indication that the client is trying to boot via NBD instead of NFS. Check the following:

Firstly, see if there is a directory /opt/ltsp/i386/images, this is the default directory for the ltsp-update-image generated unionfs files. If it exists you should probably delete it, reboot the client and all may be well.

If not, then on the server, use the chroot command to maintain the LTSP chroot:

$ sudo chroot /opt/ltsp/i386

Now edit /etc/default/ltsp-client-setup and change the value of the root_write_method variable to use bind mounts instead of unionfs, it should look like this afterwards:

root_write_method="bind_mounts"

Next, create the file /etc/initramfs-tools/conf.d/ltsp and add the following line (set the value of the BOOT variable to nfs):

BOOT=nfs

Regenerate the initramfs:

update-initramfs -u

Hit CTRL-D to exit the chroot now. Make sure LTSP uses the new initramfs to boot:

$ sudo ltsp-update-kernels

Clients cannot login via LDM

If your clients get to the LDM login page, but none of them can login, and particularly if you are aware of recent updates affecting ssh, then:

$ sudo ltsp-update-sshkeys
$ sudo ltsp-update-image  (skip this for debian*)

*It is important that you do not ltsp-update-image for debian (unless you have moved to nbd in which case you presumably know what you are doing), if you do, then you are likely to find that the client will attempt to boot from the generated image.

See https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/144296 for more details (which may not be relevant for Debian).

No client sound

See LTSP5 Client Sound.

Display Problems

See LTSP5 Display Troubleshooting.

Client has loud hissing noise

The microphone has likely been activated by default. With normal PCs the user fixes the problem once and that fix it remembered, with LTSP the problem repeats on every boot. The solution is to mute the microphone on every boot. To achieve this we need to login to the client shell and run “alsamixer” to mute the microphone. Silence, ah bliss. Take a note of the name of the microphone audio channel. The amixer program is also very helpful - see LTSP5 Client Sound for details of how to use this.

Next set the volume on this channel in lts.conf:

Volume parameters take the form of:

"channel_name"_VOLUME, where “channel_name” is the name of the audio channel (you can find audio channel names by running alsamixer or “amixer -c0 scontrols” at a client shell) with unfriendly variable characters replaced or removed. In other words, forward slashes, hyphens, and spaces are replaced by underscores, and anything in parentheses is dropped completely.

So, if you have a channel called “Front Mic”, then FRONT_MIC_VOLUME should already work with no need for special scripts.

Thanks to Gadi for explaining this on the LTSP mailing list.

LDM only prompts for password on login failure

Login window asks for Username and then Password. But, if you enter the username incorrectly then it says “Password incorrect. Try again.”. Of course, no password will work because the username is wrong.

Edit /opt/ltsp/i386/etc/ssh/ssh_config and add this line to the bottom:

NumberOfPasswordPrompts 1

Then for Ubuntu or NBD setups (not Debian) run:

$ sudo ltsp-update-image

Then reboot your thin client.

Credit to Gavin McCullagh from the Edubuntu mailing list for this workaround (read more).

LDM password expiry loop

There is an odd loop that arises with the LDM password expiry process:

• User enters username and hits enter
• User enters password and hits enter
• LDM tells user password has expired enter new password
• User enters new password and hits enter
• LDM tells user to verify password
• User enters new password again and hits enter

It seems that after this process it always returns “No response from server”, and usually logging in with your the password will work just fine. Unfortunately the error leads people to think that the password change has failed and they seem unable to resolve the problem.

But on other occasions, typing the new password does not work, but it is possible that the operator set the password incorrectly.

If there is a bug, it may be related to the following thread:

• ldm problems with debian etch

LDM login screen shows white background

In Debian Etch, the LDM login screen shows a white background with an ugly LTSP logo in the middle. Actually the background is a light grey, but people describe it as white. The LTSP logo was not designed to work with a light grey background, and that’s the reason it looks so ugly. This problem has been fixed in Debian Lenny, and you should now upgrade from Etch.

No known workaround at this time, but see How to Create a custom LDM theme for details of how to change the logo, which really is the only issue with the grey background.

LTSPFS mounts appear strange when running df

util-linux 2.13 fixed this bug; however it is still present in etch. workaround is to make /etc/mtab a symlink to /proc/mounts:

# mv /etc/mtab /etc/mtab.old
# ln -s /proc/mounts /etc/mtab

LTSP client hostname displays as (none)

Workaround set up dns on the thin-client network, or:

# mv /etc/mtab /etc/mtab.old
# ln -s /proc/mounts /etc/mtab

LDM support for mostly insecure logins (LDM_DIRECTX)

Requires getting /root to be writeable:

# echo 'copy_dirs="$copy_dirs /root"' >>/opt/ltsp/i386/etc/default/ltsp-client-setup

KDE local device icons on desktop

In KDE there may not be any local device icons appearing on the desktop. If you find this to be the case, but the device does exist under /media, then you may need to script /usr/sbin/ltspfsmounter code to create symlinks on ~/Desktop.

X Configuration

If X doesn’t automatically configure properly (it uses X.org’s built-in configuration, which doesn’t work as well on the version on X.org in etch), you may want to switch back to the slower but more reliable X configuration using xdebconfigurator:

# chroot /opt/ltsp/i386 apt-get install xdebconfigurator
# cp /opt/ltsp/i386/usr/share/doc/ltsp-client-core/examples/configure-x-xdebconfigurator.gz /opt/ltsp/i386/etc/
# gunzip /opt/ltsp/i386/etc/configure-x-xdebconfigurator.gz
# chmod +x /opt/ltsp/i386/etc/configure-x-xdebconfigurator
# echo CONFIGURE_X_COMMAND=/etc/configure-x-xdebconfigurator >> /opt/ltsp/i386/etc/lts.conf
# echo 'copy_dirs="$copy_dirs /var/cache/debconf"' >> /opt/ltsp/i386/etc/default/ltsp-client-setup
# echo 'temp_copy_dirs=" "' >> /opt/ltsp/i386/etc/default/ltsp-client-setup

Introduction

Time to install a new server - yay! The hardware is an HP Proliant DL360 G5, with a single quad-core processor and 4GB ram. This page is about creating a complete desktop server and, as such, much of this is actually nothing to do with LTSP, but is all about setting up a functional and attractive desktop server.

Server Specification

The following page, which I wrote for LTSP4.2, should largely hold true:

• Server Sizing

Server Set-up

Physically install the server and plug the network into a Gigabit port on your network switch and install your disk drives and configure your RAID volume.

Debian Base Install

Download the netinst ISO from Debian. Burn the ISO to disk. Insert the disk into the optical drive of your server and boot. Follow the installer prompts, until you get to the section on partitioning.

Partitioning

There is no way that I can tell you what to do here - there are just too many variables - but here is what I did:

In addition to the following, you will need to create a swap partition which should be roughly equal to your RAM. Opinions do differ on this point, and if you have plenty of RAM, then you should be able to have a much smaller swap partition, perhaps only 1 GB. For low-memory systems, you should probably create a swap partition that is much larger than your RAM, but expect the performance to suffer as a result! Server based systems should be given adequate RAM, and the swap should really only be used occasionally.

Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/cciss/c0d0p5       885436    150672    689784  18% /
/dev/cciss/c0d0p1        85528     28829     52283  36% /boot
/dev/cciss/c0d0p2        85549      5664     75468   7% /rescue
/dev/cciss/c0d0p6     48062440    184344  45436620   1% /home
/dev/cciss/c0d0p7     48062440    184324  45436640   1% /opt
/dev/cciss/c0d0p8     28834716    176200  27193792   1% /srv
/dev/cciss/c0d0p11     2529920     69384   2332020   3% /tmp
/dev/cciss/c0d0p9      6728280    293660   6092840   5% /usr
/dev/cciss/c0d0p10     2885780    256356   2482836  10% /var

With an LTSP system it is sensible to have a separate /opt partition; it doesn’t need to be as large as this, but I choose to use /opt for my virtual machines, 10GB should be adequate. On servers it is sensible to have separate /usr, /var and /tmp file systems. The /srv is simply going to be for shared documents - so your requirements may be different - perhaps you have a NAS for that purpose. A separate /home is recommended for all systems. The /tmp partition here is HUGE, this is because I will be installing a virtualised Windows server using KVM in snapshot mode, which uses /tmp for working files; you should be able to get away with 0.5GB. Lastly the /rescue partition is optional, it should be a mirrored copy of /boot and then should be removed from /etc/fstab, as a rescue partition.

Installing Debian

My personal preference is to install a base system only, unchecking even the “Standard system” when prompted. This leaves you with a perfectly working system on which to build your LTSP server. Follow the prompts to remove the install CD and reboot.

Install OpenSSH Server and Screen

These packages make it easy to remotely administer your server, openssh creates the connection, screen enables you to use multiple terminals and to disconnect and reconnect to each (Simple Tutorial). Screen is optional but recommended, this tutorial will make no further mention of it, you should just use it as appropriate, or if you don’t wish to use it, simply remove it from the following command.

Log into the server as root and install openssh:

# apt-get install openssh-server screen

You may need to create some static routes in /etc/network/interfaces, depending from where you will be ssh’ing.

Return to the comfort of your workstation and ssh into the server in the usual way (this isn’t a tutorial on ssh!).

Install K Desktop Environment (KDE)

We use KDE, so these instructions are specific to KDE; however you can simply change the install command below according to the desktop environment that you prefer.

N.B. Note that I am installing the minimum software, if you prefer to have KDE in total, then just install kde-desktop instead of kde-core kdm.

N.B. Change kde-i18n-engbi for the correct package for your language.

# apt-get install kde-core kdm kde-i18n-engb

N.B. Even installing a base KDE will result in a large number of packages being downloaded, go get yourself a cup of coffee.

Reboot

Now reboot the server, and ensure that you get a graphical login on the console.

Install LTSP

N.B. Please read the Debian LTSP Howto before continuing.

N.B. Please note that there are now lenny-backports available.

# apt-get install ltsp-server-standalone

You can safely ignore the error message about DHCP server not being “authoritative”, and don’t worry that your DHCP server doesn’t start - you need to configure it before it will start cleanly.

Build LTSP Client

N.B. Please note that there are now lenny-backports available.

$ ltsp-build-client

The ltsp-build-client command will download a complete Debian filesystem into /opt/ltsp/i386 (or specify an alternate location with ‘–base’) and install the ltsp-client and ldm packages (the LTSP Display Manager).

N.B. If you change the IP data after you have done the initial setup and run ltsp-update-sshkeys on the server.

N.B. The files the client will boot are installed on the server into /var/lib/tftpboot, unlike LTSP4.2 which installed in /tftpboot.

Set-up NFS Share

$ nano /etc/exports
/opt/ltsp       *(ro,no_root_squash,async,no_subtree_check)
$ invoke-rc.d nfs-kernel-server reload

The 2 ltsp configuration files inside the client are /etc/lts.conf and /etc/default/ltsp-client-setup. See the examples in /usr/share/doc/ltsp-client*. Also see the Edubuntu wiki.

Configure DHCP Server

# cp /etc/dhcp3/dhcpd.conf /etc/dhcp3/dhcpd.conf.lenny
# cp /etc/ltsp/dhcpd.conf /etc/dhcp3/
# nano /etc/dhcp3/dhcpd.conf

Edit /etc/dhcp3/dhcpd.conf, updating the IP addresses according for your network.

Test Client Boot

Now is the time to test that your client is booting okay.

Audio Set-up

Next we need to set-up the audio, and to that end please follow the instructions on our LTSP5 Sound page:

• LTSP5 Sound

Local Devices Set-up

Next we need to set-up local devices, and to that end please follow the instructions on our LTSP5 Local Devices page:

• LTSP5 Local Devices

Optional - Installing FreeNX

If you wish to be able to remotely support your server with a GUI, then you may like to install freeNX:

• Install FreeNX

Then visit the NoMachine website and download the client for your PC.

Optional - KDE Style Changes

N.B. These instructions will increase server load, but do make KDE look a great deal more attractive - your choice.

I apologise in advance for this section; I know that “Bling” has no place on a server, but equally I know that getting Linux adopted in your workplace is tough, and a little bling goes a long way towards user acceptance. Unfortunately these changes WILL increase the load on the server, if you are not happy with this, then simply skip this section.

Crystal Windows Decoration

# apt-get install kwin-style-crystal

Go to the KDE Control Panel, Appearance & Themes and change the Windows Decorations to Crystal. Then, under the General tab, make the top corners rounded, and increase the titlebar size to 21. Under the Buttons tab, change the buttons to Kubuntu-Edgy. Under the Background tab, I suggest you switch off transparency, which will but an unnecessary load on the server. Under the Overlay tab, change both active and inactive to Simple Lighting.

NuoveXT

# apt-get install kde-icons-nuovext

Go to the KDE Control Panel, Appearance & Themes and change the Icons to nuoveXT. Under the Advanced tab you may like to increase the Desktop icon size to 48, which suits these icons.

Desktop Background

I believe one of the most attractive backgrounds is a simple two colour elliptic gradient. To achieve this, simply go to KDE Control Panel, Appearance & Themes, Background. Change the Background picture to “No Picture” and set colours to Elliptic Gradient. You can choose any two colours, but these two look good:

#2E3B4E
#7698CA

Other colours may not go well with the rest of my suggested changes.

Panel

The default panel doesn’t look that great, you can download any number of Kicker Wallpapers from KDE Look, but one that I think looks very good is this one:

• Apple-like Kicker Wallpaper

It looks plain on that page, but much better in KDE.

Add Debian Multimedia Repository

The Debian Multimedia is a useful resource for multimedia codecs, converters and players. In particularly it includes Adobe Flash Player. This is not an official repository, although it is a very common one. I have decided to take the risk of installing this software on my system, you may not agree!

Edit /etc/apt/sources.list and add the following repository:

deb http://www.debian-multimedia.org lenny main

Download and install the following:

• debian-multimedia.org keyring

# dpkg -i debian-multimedia-keyring_2008.10.16_all.deb
# aptitude update

Optional - Iceweasel

In my opinion, Konqueror, whilst an amazing piece of software, is not going to be suitable as your main web-browser. Mozilla Firefox is the answer, and this is available in the Lenny repositories, named “Iceweasel”.

# apt-get install iceweasel icedtea-gcjwebplugin flashplayer-mozilla mozilla-acroread mozilla-mplayer

Unfortunately, being a GTK application (not native to KDE) it looks ghastly, so…

Optional - GTK Applications in KDE

GTK applications in KDE always look horrible. In Kubuntu this is taken care of for you, but in Debian it is only a quick install away:

# apt-get install gtk-qt-engine

There are other ways of achieving this, e.g. qtcurve, but for me this seems perfectly satisfactory. Having installed, you visit your control panel and you will find a new option under “Appearance & Themes” called “GTK Styles and Fonts” change this to use your KDE style for your GTK applications, and restart KDE.

Optional - Other Applications

Applications you might consider installing for your users are:

• KUser : apt-get install kuser
• The Gimp : apt-get install gimp gimp-help-en gimp-data-extras
• Inkscape : apt-get install inkscape
• Scribus : apt-get install scribus-ng scribus-ng-doc scribus-template (Development branch, but more up-to-date than “scribus”)
• OpenOffice : apt-get install openoffice.org-kde openoffice.org-help-en-gb openoffice.org-l10n-en-gb openclipart-openoffice.org
• Freemind : apt-get install freemind
• KCalc : apt-get install kcalc
• Microsoft Fonts : apt-get install ttf-mscorefonts-installer
• KeePassX : apt-get install keepassx
• Unison : apt-get install unison unison-gtk
• Wine : apt-get install wine wine-utils
• Kontact : apt-get install kontact
• Ark : apt-get install ark
• Dia : apt-get install dia

Optional - Install KVM Virtualisation

If you intend adding a virtualised Windows server on top of the LTSP server (not recommended - but we do it and it works well), then you need KVM (or an alternative) installed:

N.B. I suggest you right-click on the following link and Open in New Tab, so that you don’t lose your place in this document.

• How to install KVM

Optional - UMask

I dislike that every created file is given read access to all users, to change this behaviour we need to change umask. In KDE the easiest way to achieve this system wide is by editing /etc/profile and changing the umask line from the default line of umask 022 to your preferred setting. We use umask 007, which gives full access to the user and group, but no access to other.

Default permissions umask 022: -rw-r--r--
New permissions     umask 007: -rw-rw----

Your view may well vary from mine.

Set Client Root Password

You will almost certainly need access to the client, in order to troubleshoot:

# chroot /opt/ltsp/i386
# passwd root
Enter new UNIX password:
Retype new UNIX password:
passwd: password updated successfully
# exit
exit
# ltsp-update-image

References

• Debian LTSP
  • Debian Lenny Howto
  • Debian Lenny Backports Howto
• HP Proliant Debian Lenny Installation Instructions by HP Alternative link
• GNU Screen Manual
  • Simple Tutorial
• Linux Partitioning mini-FAQ
• Installing KDE
• Installing Adobe Flash in Debian
• Ubuntu Thin Client Howto

Paths

The ldm-theme is loaded from a symlink at:

/opt/ltsp/i386/usr/share/ldm/themes/default

Which points to further symlink at:

/opt/ltsp/i386/etc/alternatives/ldm-theme

Which points back to :

/opt/ltsp/i386/usr/share/ldm/themes/ltsp

Copy existing theme

Copy this ltsp directory, e.g.:

# cd /opt/ltsp/i386/usr/share/ldm/themes
# cp -r ltsp custom
# cd custom
# ls -al

Set symlink to new theme

Then you need to change the symlink to point to your new theme:

# cd /opt/ltsp/i386/etc/alternatives
# rm ldm-theme
# ln -s /usr/share/ldm/themes/custom ldm-theme

Customise your custom LDM theme

Now you can customise your ldm theme to your heart’s content. Please note that editing the bg.png currently has no effect (hence the existence of this page!), but you can at least tidy up the login page by replacing the logo.png with something tidier than the current ltsp.org logo, which looks dreadful on the grey background (because of a transparency issue).

Change background colour

Firstly change the background colour to your desired background colour (personally I quite like the light grey of “efebe7”), so left it like that.

You can change the colour of the background by opening the greeter-gtkrc in your theme and looking for the following line:

    bg[NORMAL]        = "#efebe7"

Change the colour for whatever colour you wish. Under KDE you can use the KColorChooser for this purpose, or indeed the colour picker in Gimp.

Create new logo

Next open your organisation’s logo in the Gimp (or whatever logo you wish to appear on LDM login). Add a shadow (Script4u → Shadow → Drop Shadow), I would suggest making the X/Y offsets 0/0. You can play with the radius and opacity to create the effect you desire.

Next create a new layer of the same colour as the background (e.g. efebe7 unless you changed it) and put that background at the bottom of the layers. The gimp makes this easy by enabling you to key in the efebe7 in the colour picker.

Save the resulting image as logo.png into your custom ldm theme.

Command

ltsp-update-kernels

Introduction

The ltsp-update-kernels script copies it into the chroot environment and automatically updates the symlinks for the boot loader in the process.

ltsp-update-kernels updates the current client kernel by using the current kernel from the chroot and copying it to the tftpboot directory (/var/lib/tftpboot/ltsp/i386/). It also copies the initramfs across and sets up the PXE linux boot files.

This command must be run each time the kernel changes in the chroot. There should be no harm in running this command each time you update the chroot, just in case.

Command

ltsp-update-image

Comment

For systems that use NBD (not Debian), this command will create a client image from the contents of /opt/ltsp/i386, and store it in /opt/ltsp/i386/images.

Any change you make in the chroot will only take effect for the clients after you run ltsp-update-image. As such, this is the command that you will run most often.

As Debian uses NFS, this command is not necessary, unless you have chosen to change to NBD, in which case you presumably know that you are “different”.

If you are on Debian and you have a directory /opt/ltsp/i386/images, then this is likely to mean that you have run this command by mistake. This may trigger the clients to attempt to use NBD, so you should probably remove this directory, or explicitly tell your clients to use NFS - see LTSP5 Troubleshooting for details.

Command

ltsp-build-client

Introduction

The ltsp-build-client command will download a complete Debian filesystem into /opt/ltsp/i386 (or specify an alternate location with --base) and install the ltsp-client and ldm (LTSP Display Manager) packages. As such, you will probably only run this command the first time you install LTSP5.

You should move /opt/ltsp to /opt/ltsp_old before you start.

Man Page

You should see the man page appropriate for your system, but the Debian Lenny man page current states:

ltsp-build-client(8)                                                                                                                                                                                               ltsp-build-client(8)

NAME
    ltsp-build-client - Builds an LTSP client system for use with the LTSP server

SYNOPSIS
    ltsp-build-client  <options>

DESCRIPTION
    ltsp-build-client  builds  an  ltsp client chroot using standard packages for use with the ltsp server.  This chroot installation  can be modified in a number of ways as outlined in the OPTIONS section.  If present, defaults
    will be read from the configuration file /etc/ltsp/ltsp-build-client.conf, and will be overridden by using the command line options. ltsp-build-client is usually run as the root  user.   Visit  http://www.ltsp.org  for  more
    information.

OPTIONS
    Not all options are available on all installations. For a complete list of available options, use the --help and --extra-help options.

    --arch arch
                Builds a chroot with the specified architecture. Must be a supported architecture of the distribution, and compatible with the host architecture.  Defaults vary by distribution.

    --base dir
                Sets the default path for the ltsp chroots to be installed into. Defaults to /opt/ltsp/.

    --chroot file
                Sets the default name for the ltsp chroot. Defaults to the architecture selected with --arch or the running architecture.

    --conf file
                Load configuration defaults from specified file. See CONFIG FILE below for more information.

    --dist distribution
                Distribution : Sets the distribution to be installed in the client chroot.  Defaults to the running distribution.

    --extra-help
                Displays all available commandline options.

    --help    Displays the ltsp-build-client help message.

    --mirror mirror_url
                Select the mirror to download packages from. file:/// and http:// URLs are supported.

CONFIG FILE
    ltsp-build-client will read a configuration file to override built-in defaults. the default location for this configuration file is /etc/ltsp/ltsp-build-client.conf, and can be specified with the --conf commandline option.

    values are NAME="VALUE" pairs, with # representing commented lines. most commandline options have an equivalent. an example Debian based system with an alternate base location, architecture, and distribution:

        # set alternate default location for ltsp chroot (--base)
        BASE="/srv/ltsp"
        # build an amd64 chroot by default (--arch)
        ARCH="amd64"
        # build a Debian sid environment (--dist)
        DIST="sid"

SEE ALSO
    ltsp-update-sshkeys(8), ltsp-update-kernels(8).

AVAILABILITY
    ltsp-build-client is part of ltsp package and the latest versions are available in source form from https://launchpad.net/products/ltsp.

MAN PAGE AUTHOR
    Peter Savage
    updated 2006, 2009 by Vagrant Cascadian

Command

ltsp-update-sshkeys

Introduction

Secure Shell Handler or SSH is a network protocol that allows data to be exchanged using a secure channel between two computers. LTSP uses it to enable secure communication between clients and the Server. SSH uses public-key cryptography to authenticate the client, this command rebuilds the keys.

You would run this command only at first install, or upon subsequent change of server IP addressing or DHCP or possibly hosts. Basically if you get the “workstation not authorised” message, then you should run this command.

All the following commands are entered on the server. Normally this would be having logged into your normal LTSP desktop and opened a terminal session, but it could equally be from an ssh session to the server. The key thing is that you do not run these commands from a client shell, as the client system is read-only.

These instructions are for Debian, Ubuntu users will also need to update the client image afterwards.

Step by Step Process

$ sudo chroot /opt/ltsp/i386
# mount -t proc /proc /proc
# mount -t sysfs sys /sys
# apt-get update
# apt-get upgrade
# umount sys
# umount /proc
# exit
$ sudo ltsp-update-kernels

The mount proc and mount sysfs commands are not always required, and can be omitted in many instances, but if in doubt I don’t believe they cause any harm.

The ltsp-update-kernels command is only required if the client kernel has been upgraded, but again I don’t believe it causes any harm to run it unnecessarily.

I have run through these notes again and corrected the odd mistake and they do seem to work. At least until something changes!!

I have not yet got the init.d script working, but am able to successfully start adempiere and reach the login page.

I need to review the postgresql security settings as “trust” for all is probably unsafe.

Lastly, I am an adempiere newbie, so please take everything I say with a pinch of salt.

Good luck!

Install a Debian system

Please visit Debian for guidance on how to do this. Personally I like to install an absolutely bare system as my starting point (I even uncheck the option for “standard system”).

It is possible, I believe, to do a headless (non-GUI) installation, but these instructions do require a desktop environment, as some of the setup screens require a GUI.

For example to install a minimal KDE on a Debian base system:

# apt-get install kdm kde-core

Install tofrodos

Required for the database migration.

# apt-get install tofrodos

Install Postgresql > 8.2

Fortunately Debian includes version 8.3, so a simple apt-get should do the trick:

# apt-get install postgresql postgresql-doc oidentd

And a ton of dependencies will also be installed.

Once installed, you should have a functional postgresql database installed and running. Debian places the database in /var/lib/postgresql/version the default database is named main and is located in /var/lib/postgresql/version/main. We assume in this guide that Adempiere’s database will be created in the default cluster.

Install Java

sun-java6 is the version to install, unless you are using the ajax interface in which case you can even use openjdk. To do add contrib non-free to /etc/apt/sources.list repositories.

# apt-get install sun-java6-jdk sun-java6-fonts sun-java6-jre sun-java6-plugin

Configure postgresql for java

Now we must add the JAVA_HOME variable to the /etc/postgresql/8.3/main/environment so postgresql can find the java virtual machine:

JAVA_HOME = '/usr/lib/jvm/java-6-sun'

Authorise users for postgres

Now we must enable users to access the database as the postgres user, which is the system user of the database and will be used by the RUN_ImportAdempiere.sh script to import the database data.

Add the following lines to /etc/postgresql/8.3/main/pg_hba.conf:

You may wish to change the IP address to the external IP

# Database administrative login by UNIX sockets
#local   all         postgres                          ident sameuser

# TYPE  DATABASE    USER        CIDR-ADDRESS          METHOD

# "local" is for Unix domain socket connections only
local   all         all                                 trust

# IPv4 local connections for user postgres:
host    all         postgres       127.0.0.1/32         trust
# IPv4 local connections for user adempiere:
host    all         all            127.0.0.1/32         trust
# IPv6 local connections:
host    all         all            ::1/128              trust

trust or md5 - trust is probably wrong

Configure postgresql listen address

Running the database configured in localhost isn’t probably what you want, as all the clients will need access to this database and the address you tell the script the database runs from will be remembered in the future (but it can be altered). So you should probably put you server external IP address (the address of the eth0 interface i.e.) instead of the 127.0.0.1 address of the previous line.

If this is a test server and you will only be accessing it locally, then leaving it as 127.0.0.1 is fine.

Restart Postgresql

We must now restart the postgresql server to active the changes in the configuration files:

# /etc/init.d/postgresql-8.3 restart

Install Subversion

# apt-get install subversion

Download Adempiere from SVN Head

This could take a very long time, depending on your Internet bandwidth

$ svn co https://adempiere.svn.sourceforge.net/svnroot/adempiere/trunk */svnworkdir/adempiere

Build Adempiere

Open a terminal window.

$ export JAVA_HOME='/usr/lib/jvm/java-6-sun'
$ cd */svnworkdir/adempiere/utils_dev
$ ./RUN_build.sh

Install Adempiere

# unzip */svnworkdir/adempiere/install/build/Adempiere_354a.zip -d /opt

Now the extracted folder /opt/Adempiere exists in /opt:

$ cd /opt
$ sudo mv Adempiere adempiere-3.54a
$ sudo ln -sf adempiere-3.54a adempiere

Now it looks like this:

$ ls -l /opt
$ total 4
$ lrwxrwxrwx  1 root root   15 Oct 18 19:22 adempiere -> adempiere-3.54a
$ drwxr-xr-x 10 root root 4096 Oct 18 19:11 adempiere-3.54a

Create the database

Switch to the postgres user:

debian:/opt$ sudo su - postgres
postgres@debian:*$ export ADEMPIERE_HOME="/opt/adempiere"
postgres@debian:*$ /usr/bin/createdb adempiere
postgres@debian:*$ /usr/bin/createuser adempiere
Shall the new role be a superuser? (y/n) y

Set database password

postgres@debian:*$ echo "ALTER USER adempiere WITH PASSWORD 'your-chosen-password';" | psql
ALTER ROLE

Check database

postgres@debian:*$ echo "select * from pg_user; select * from pg_roles;" | psql
  usename  | usesysid | usecreatedb | usesuper | usecatupd |  passwd  | valuntil | useconfig
-----------+----------+-------------+----------+-----------+----------+----------+-----------
 postgres  |       10 | t           | t        | t         | ******** |          |
 adempiere |    16385 | t           | t        | t         | ******** |          |
(2 rows)

  rolname  | rolsuper | rolinherit | rolcreaterole | rolcreatedb | rolcatupdate | rolcanlogin | rolconnlimit | rolpassword | rolvaliduntil | rolconfig |  oid
-----------+----------+------------+---------------+-------------+--------------+-------------+--------------+-------------+---------------+-----------+-------
 postgres  | t        | t          | t             | t           | t            | t           |           -1 | ********    |               |           |    10
 adempiere | t        | t          | t             | t           | t            | t           |           -1 | ********    |               |           | 16385
(2 rows)
postgres@debian:*$ exit

Load database

$ export ADEMPIERE_HOME="/opt/adempiere"
$ psql -h localhost -d adempiere -U adempiere < $ADEMPIERE_HOME/data/Adempiere_pg.dmp

Set-up Adempiere

$ export ADEMPIERE_HOME="/opt/adempiere"
$ cd $ADEMPIERE_HOME
$ sudo chmod +x *.sh
$ export JAVA_HOME='/usr/lib/jvm/java-6-sun'
$ sudo ./RUN_setup.sh

Please note that this section must be able to write to the display, so either the desktop session needs to be root, or use sudo. If you get an error about connecting to DISPLAY then try using sudo.

The following screen should now be displayed. Change the database name to adempiere, database type to postgresql, set your mailserver settings. You may need to set the database server to localhost, if the hostname does not work.

Migrate to Head

$ cd */svnworkdir/adempiere/migration
$ chmod +x *.sh
$ ./migrate_postgresql.sh 354a-trunk commit | psql -U adempiere -d adempiere > 354a-trunk.lst
NOTICE:  ALTER TABLE / ADD PRIMARY KEY will create implicit index "ad_document_action_access_key" for table "ad_document_action_access"
NOTICE:  ALTER TABLE / ADD PRIMARY KEY will create implicit index "r_contactinterest_key" for table "r_contactinterest"
NOTICE:  CREATE TABLE / PRIMARY KEY will create implicit index "c_ordersource_pkey" for table "c_ordersource"
NOTICE:  CREATE TABLE / UNIQUE will create implicit index "c_ordersource_ad_client_id_key" for table "c_ordersource"
WARNING:  there is no transaction in progress

Start Server

$ cd $ADEMPIERE_HOME/utils
$ sudo nohup ./RUN_Server2.sh &
$ sudo tail -f nohup.out

The output of nohub.out will indicate that the server started in only a few seconds, this is only the JBoss server, the site will not be accessible for several minutes.

Client Login

On the server, open a browser and go to: http://hostname.yourdomain/admin/ and follow the Adempiere ZK Webui

http://localhost/admin/ will not work.

For guidance visit http://www.adempiere.com/index.php/InstallClient.

Optional: Install init.d script

Not essential for a test server (and I have not yet got this working, I suspect because I am not using user = adempiere.

# ./RUN_Server2Stop.sh
# sudo cp unix/adempiere_Debian.sh /etc/init.d/adempiere
# /etc/init.d/adempiere start 

Read the Functional User Manual

• http://www.adempiere.com/index.php/Functional_User_Manual

References

• http://www.adempiere.com/index.php/Debian_and_PostgreSQL_Install
• http://www.adempiere.com/index.php/Adempiere/Libero_Linux_install_from_subversion_head
• http://www.adempiere.com/index.php/Manual_Installation_From_Trunk
• http://www.adempiere.com/index.php/ADempiere_Install_Linux&PostgreSQL
• http://www.adempiere.com/index.php/ADempiere_Install#JDBC_Connection_Error
  • http://www.adempiere.com/index.php/CreateDatabase
  • http://www.adempiere.com/index.php/InstallComplete
• http://www.adempiere.com/index.php/ADempiere_with_Postgresql_on_Ubuntu_9.04_Jaunty

RAM

LTSP website recommends 256mb for the core system plus 50mb per user:

Total RAM = 256mb + (25 users x 50mb)
          = 1506mb

However Edubuntu recommends a different formula:

Total RAM = 150mb x 25 users
          = 3750mb

We recommend you choose the greater of the two, and double-it if you can (see update below).

Post install update

This memory has proved a little low, mainly because we are also virtualising Windows on the same server. If you don’t require Windows, or you are providing that service from a different server, then the higher of these two figures should be sufficient; I would certainly recommend the higher figure, because of the high memory demands of Firefox and Adobe Flash.

CPU

LTSP website states that each user is likely to require around 3% of a CPU:

Total CPU = 3% x 25
          = 0.75
          = 75% of a CPU

So a single processor should be sufficient; however this requirement doubles or even trebles if we opt to virtualise a Windows server on the installation. We would say that a minimum of a quad-core would be sensible if virtualising Windows on the LTSP server. If this is a problem for you, simply host the Windows server elsewhere; which is probably better anyway.

Post install update

We have now upgraded both servers to dual quad-core, which is proving ample capacity. I believe a single quad-core would have been ample had we not been virtualising windows on the same server.

Network

LTSP recommend a Gigabit connection for the server and fast Ethernet for the terminals. Many fast ethernet switches include a gigabit connection, alternatively gigabit switches are becoming quite affordable.

There can be issues running LTSP via multiple switches - so if you’re buying a new switch - consider the greatest number of ports that you are likely to require.

Post install update

This has all provided to be correct - Gigabit on the server is nearly essential if you want decent performance and similarly, PCs with only old 10-base-t internet are noticeably slower to boot.

Solving VirtualBox Mismatch

If you install virtualbox-ose under Debian, along with virtualbox-ose-modules-2.6, there can be a mismatch between versions, which will prevent VirtualBox from running. At the time of writing virtualbox-ose is at version 1.66 and virtualbox-ose-modules-2.6 is at 1.62.

Unfortunately you need to build the kernel module yourself.

Step 1: Determine modules package name

$ dpkg -l "virtualbox-ose-modules*" | grep ii
ii  virtualbox-ose-modules-2.6.30-2-amd64 3.0.12-dfsg-1+2.6.30-8squeeze1 VirtualBox modules for Linux (kernel 2.6.30-

Take a note of this package name for the next step.

Step 2: Uninstall the Debian package

$ sudo apt-get remove virtualbox-ose-modules-2.6.30-2-amd64

Step 3: Install VirtualBox Source

$ sudo apt-get install virtualbox-ose-source
$ sudo module-assistant auto-install virtualbox-ose

This will produce a custom virtualbox-ose-modules package in /usr/src, which will then be used to install.

Look out for error messages, if it states that a deb file already exists, remove it and try again.

Step 4: Load Kernel Driver

This step was carried out automatically in the version that I have.

$ sudo modprobe -r vboxdrv
$ sudo modprobe vboxdrv

Yes once again I come to complete my tax return, last time it was a far from successful experience.

Given that that page is now a year old and my next year’s return is required, I thought it was an opportune time to update it. Unfortunately nothing much has changed, the list of personal tax applications has shrunk, as TaxCalc’s market domination continues. The only Linux-compatible competitor seems to still be ftax and it requires Acrobat Reader, which isn’t available for 64-bit Linux.

The problem is that you don’t really want to mess about with your personal tax, especially when that includes on-line submission. You need to have faith in the company that offers the product, and you need them to have sufficient numbers of customers that, if there is a problem, it gets fixed quickly.

In short, the only meaninful competition for TaxCalc would seem to be HMRC’s own web-based offering. If it weren’t for my positive experiences in using TaxCalc, the convenience of being able to import the previous year’s return, and the fact that I have a Windows virtual-machine on my laptop, I would probably try out the HMRC system. As it is, sadly, the Windows-only TaxCalc wins again.

Introduction

I created this page when completing my tax return for 2007/8, and am now updating it for 2008/09 tax year (January 2010). I currently use TaxCalc running on Linux under my Windows virtual machine (in VirtualBox OSE), I would like to stop relying on Windows; so what are the options for submitting a UK Self-assessment tax return under Linux?

• Linux alternatives
• HMRC On-line Self-Assessment
• Install TaxCalc under Wine
• Run TaxCalc in VirtualBox

Linux Alternatives

The following is HMRC’s list of approved software:

• http://www.hmrc.gov.uk/efiling/SA_efiling/soft_dev.htm

The list updated as at January 2010 is as follows:

• Acorah Software Products Ltd (TaxCalc) - Windows only
• Andica Limited - Windows only
• BTC Software Limited - Professional only
• CCH Software - Professional only
• Drummohr Technology Ltd - Professional only
• Forbes Computer Systems Ltd - Windows only
• IRIS Software Ltd - Professional only
• Isokon Limited - Professional only
• Keytime Objective Ltd - Not available for download
• Legatio Ltd (Ftax) - Windows, Mac and Linux
• PTP Software Ltd - Professional only
• Quality Management Software Ltd - Windows only
• RFA Limited - Professional only
• Sage (UK) Ltd - Professional only
• Thomson Reuters - Not seemingly available for download

The comments are my own personal impression at the time of visiting the websites, these opinions may be wrong and you should make your own enquiries before making a decision.

Some of these, like Sage, are clearly not consumer level products, and my impression is that the above list is somewhat misleading, as in practice TaxCalc seems to have the personal market pretty much sown up. Am I wrong? Let me know in the Comments section at the bottom.

Of those that are consumer level products only ftax seems to support Linux. Ftax requires Adobe Acrobat Reader, which is not currently available for 64-bit Linux, also the demonstration shows that it is a smart walk-through of the actuals forms; which does fall short of the wizard mode under TaxCalc. So to run ftax I am only swapping one set of compatibility issues for another. If you are running 32-bit Linux, then I would definitely investigate ftax further, perhaps by searching reviews of ftax.

If you are going to try ftax, then you will notice the lack of any help for Linux on their site. In fact there is help, but it is difficult to find, given their lack of a search facility. Google to the rescue:

• google “site:ftax.co.uk linux”

HMRC On-line Self-Assessment

Another option is to use HMRC’s own software, which I believe is web-based. The only issue with this is that it appears that you need to enrol in the on-line service, even if you have previously used TaxCalc (or other Internet-filing service); see the following page for details:

• http://www.hmrc.gov.uk/sa/using-online.htm

Nevertheless, this is probably a good option, judging from the on-line demo:

• http://www.hmrc.gov.uk/demo/individual/self-assessment/

However, the system requirements for this service state:

“If you use HMRC’s tax software and; or the Data Provisioning Service, you will also need a minimum browser level, Internet Explorer 5.0 (Windows users) Safari 2.0.4 (Mac users) or Opera 7.0 (Windows or Mac) Your Internet browser must have JavaScript and cookies enabled and be capable of supporting 128-bit SSL. Other operating systems and browsers, such as Mozilla Firefox, may allow you to access the site and use the pages but might not display the site as designed or allow you to access all the functionality.”

Okay, so it doesn’t mention Linux, but I wouldn’t be unduly concerned about that. No the issue seems to be the browser, and, reading between the lines, this would suggest to me that the site is using extensive Javascript, which is largely blocked (quite rightly) by Firefox. The chances are that there would be issues; although it wouldn’t be massive problem for me to fallback on IE under a Virtual Machine, your mileage may vary.

Install TaxCalc under Wine

TaxCalc themselves state:

“We do have some customers that run Windows emulation software to enable the use of Taxcalc on other Operating Systems (e.g. MAC OS). Although we have had some positive feedback on this we do not officially support the use of TaxCalc on Non-Windows Operating Systems at this time.”

So the next option is to consider installing TaxCalc under Wine, my concern with this is that I would expect that a simple application like TaxCalc would indeed install fine under Wine. My concern would be that I would complete my tax return and go to upload to the tax office, and that point hit a snag. Also, I am running on Debian Lenny 64-bit, surely that must effect Wine as well - can 64-bit Wine run 32-bit Windows applications? The existence of this page would suggest that there may be issues:

• http://wiki.winehq.org/WineOn64bit

Having downloaded TaxCalc, I opened the installation program in Wine. The installation was successful, and the first run prompted for me to either enter the registration key, or connect to the Internet to find the registration key. I decided to do the latter, in order to test connectivity, and it worked fine. Then it crashed and I was unable to get back into TaxCalc again.

Update for 2010

A quick search on the WineHQ AppDB showed that someone has successfully installed TaxCalc under Wine; however he has not tested the online submission, which is crucial for me. Given that he has used winetricks to install IE6 under Wine, then there is a good chance that it will work. If you try it, do update the WineHQ AppDB and consider posting a comment in the new comments section below!

• TaxCalc 2009 under Wine

Run TaxCalc in VirtualBox

It is outside of the scope of this document to describe this in detail, suffice to say it works flawlessly. The downside is that you are effectively running under Windows, which can hardly be considered a success.

Conclusion

So what have I decided? Well this year I will again continue with TaxCalc, installed on Windows XP running through VirtualBox. Yes I could give ftax a go, but I am fairly unimpressed with their website, and lack of any obvious linux help. Also Googling for information on ftax and linux leads you back to this page - and, if this page really is the best help available for ftax and linux, then you’re in real trouble.

So the best hope for Linux users is HMRC’s own on-line service. It is web-based and free, and whilst it does not promise Firefox support, I suspect that may not be an issue. If you would like to be able to use TaxCalc natively under Linux, then why not let them know: by emailing them at sales@taxcalc.com.

# unopkg add --shared sun-pdfimport.oxt

For further information please visit:

• http://api.openoffice.org/docs/DevelopersGuide/Extensions/Extensions.xhtml

• Installed winetricks
• winetricks: fakeie, vb6run, mdac25, mdac28, jet40, vcrun6, wsh56, allfonts, riched30, mdac27, native_mdac, native_oldaut32, gecko
• Installed Adobe Reader 8.1.2 from Intranet Downloads
• Accepted Adobe licence via “wine regedit”, HKEY_CURRENT_USER/Software/Adobe/Acrobat Reader/8.0/AdobeViewer, add a DWORD “EULA” = “1”
• Installed SMR2000
• Updated SMR2000 by running update self-extracting zip files

Scanning with LTSP 5 in Debian Etch or Lenny

At the time of writing these instructions worked for Etch, but have not been completed for Lenny.

Introduction

This is a description for a scanner physically connected to a terminal (not to the server).

For LTSP scanning to work, you just need to remember that the roles are reversed - that is the client is the scanning server and the server is the scanning client. With the exception of the fact that you have to chroot to get at the client, the set-up is an ordinary Linux network scanning set-up.

Configure client

Install client packages

You have to be root in the chroot LTSP client environment for the next steps:

$ sudo chroot /opt/ltsp/i386
# mount -t proc /proc /proc
# mount -t sysfs sys /sys
# apt-get install libexif12 libgphoto2-2 libgphoto2-port0 libieee1284-3 libltdl3 libsane libsane-extras sane-utils
# umount sys
# umount /proc

User ‘saned’ needs an additional group:

Still in the chroot:

# usermod -a -G scanner saned

Modify /etc/inetd.conf (Etch)

Still in the chroot, open /etc/inetd.conf (e.g. with vi) and add at the end of the file (after ‘#:OTHER: Other services’) this line:

sane-port stream tcp nowait saned.saned /usr/sbin/saned saned

Modify /etc/default/saned (Lenny)

The problem with this is that it will run for all users, which seems a bit pointless, should be possible to use an ldm rc.d script, but this did not work correctly for me.

Still in the chroot, open /etc/default/saned and change RUN=no to RUN=yes.

On an upgraded server, this was not enough, I also needed to symlink:

# cd /etc/rc2.d
# ln -s ../init.d/saned S99saned

The initial openbsd-inetd links are wrong. (Etch)

Still in the chroot, Delete openbsd-inetd links with:

# rm /etc/rc*/*openbsd*

Create new one’s:

# ln -s ../init.d/openbsd-inetd /etc/rc2.d/S20openbsd-inetd
# ln -s ../init.d/openbsd-inetd /etc/rc3.d/S20openbsd-inetd
# ln -s ../init.d/openbsd-inetd /etc/rc4.d/S20openbsd-inetd
# ln -s ../init.d/openbsd-inetd /etc/rc5.d/S20openbsd-inetd
# ln -s ../init.d/openbsd-inetd /etc/rc0.d/K20openbsd-inetd
# ln -s ../init.d/openbsd-inetd /etc/rc1.d/K20openbsd-inetd
# ln -s ../init.d/openbsd-inetd /etc/rc6.d/K20openbsd-inetd

Check if everything is fine:

# ls -l /etc/rc*/*openbsd*

Leave chroot

# exit or [Ctrl]+[d]

Configure server

Install server packages

Thanks to the Debian package system it’s easy to get almost all software (as root):

# apt-get install libsane libsane-extras sane-utils

Set user grants

You have to set the saned group for every scanner user:

# usermod -a -G saned ~username~

saned.conf Configuration (client)

Modify the saned.conf file. Put the IP address of your server, e.g.:

/opt/ltsp/i386/etc/sane.d/saned.conf

#
# saned.conf
#
# The contents of the saned.conf  file  is  a  list  of  host  names,  IP
# addresses or IP subnets (CIDR notation) that are permitted to use local
# SANE devices. IPv6 addresses must be enclosed in brackets,  and  should
# always  be specified in their compressed form.
#
# The hostname matching is not case-sensitive.
#
~serverip~

net.conf Configuration (client)

Modify the net.conf file. Put the IP addresses & names of your server, scanner terminal; client and localhost here, e.g.:

/opt/ltsp/i386/etc/sane.d/net.conf

# This is the net config file.  Each line names a host to attach to.
# If you list "localhost" then your backends can be accessed either
# directly or through the net backend.  Going through the net backend
# may be necessary to access devices that need special privileges.

localhost
~server-ip~
~server-hostname~

Modify the dll.conf (client)

Enable only two lines, one with ‘net’ and the other one with the name of your (!) scanner backend (for me: snapscan -> works with Epson scanner).

/opt/ltsp/i386/etc/sane.d/dll.conf

# enable the next line if you want to allow access through the network:
net
snapscan
#abaton
#agfafocus
#apple
#avision
#artec
#artec_eplus48u
#as6e
...

Here you find an overview of supported backends:

• http://www.sane-project.org/sane-mfgs.html

Modify dll.conf (server)

Modify dll.conf. Enable only the line with ‘net’ (disable all others):

/etc/sane.d/dll.conf

# enable the next line if you want to allow access through the network:
net
#abaton
#agfafocus
#apple
#avision
#artec
#artec_eplus48u
#as6e
...

Modify net.conf (server)

Modify net.conf. Put the IP address of your scanner terminal client (!) here, e.g.:

/etc/sane.d/net.conf

# This is the net config file.  Each line names a host to attach to.
# If you list "localhost" then your backends can be accessed either
# directly or through the net backend.  Going through the net backend
# may be necessary to access devices that need special privileges.
~client-ip~

Alternatively (and better in my opinion) is to script the scanner program, so that it starts with:

SANE_NET_HOSTS=client-ip
export SANE_NET_HOSTS
scanimage > test.pnm

See Alternative Configuration.

Restart your client terminal

Restart hplip (Etch only)

$ sudo /etc/init.d/hplip restart
Stopping HP Linux Printing and Imaging System: hpiod hpssd.
Starting HP Linux Printing and Imaging System: hpiod hpssd.

Test

If it doesn’t work - take a look at your syslog.

Alternative configuration

If your client doesn’t have a static ip and hostname (as with the default ltsp set-up), then you can either configure /etc/dhcp3/dhcpd.conf to provide you with a static ip and hostname, or use this alternative scanner configuration.

Step 1: Remove all client references above

Work back through the above configurations, commenting out all references to your client-ip and client-hostname.

Step 2: Write a script

Here is a simple test script to determine your client, and list your connected scanner:

SANE_NET_HOSTS=`echo $SSH_CLIENT | cut -d " "  -f 1`
export SANE_NET_HOSTS
scanimage -L

You would need to modify this script to do something more useful with scanimage.

References

• http://wiki.ltsp.org/twiki/bin/view/Ltsp/Debian#Scanning_with_LTSP_5_in_Debian_E
• http://wiki.ltsp.org/twiki/bin/view/Ltsp/Scanners
• http://www.enterprisenetworkingplanet.com/nethub/article.php/%203637076
• http://www.jumako.de/cms/index.php?option=com_content&task=view&id=26&Itemid=31

$ sudo -i
# apt-get install nagios3
# cd /etc/nagios3
# htpasswd -c htpasswd.users nagiosadmin

Then point your browser at http://localhost/nagios3/.

This document is for the purpose of building in IPCop Firewall onto a MSI Axis 700 barebones PC. The firewall will also be used as a NTP time-server, Content-filtering proxy server and OpenVPN server.

Requirements

• MSI C7 VIA CN700 5.1 Audio Lan DDRII quicklinx:4FY9WS
• D-Link DFE 580TX - network adapter - 4 ports quicklinx:1YH9WS
• LG Electronics DVD ROM 16X BLACK DRIVE ONLY quicklinx:4N6SWS
• Seagate Barracuda 7200.10 80GB S300 8MB quicklinx:4MRKWS
• Crucial 512MB 240Pin DIMM DDR2 PC2-4200 Non-ECC quicklinx: 36TCWS

N.B. I believe 512mb is insufficient and I would recommend purchasing Kingston Memory - KVR667D2NG/1G

Key Steps

• Assemble above
• Disable “EHCP on chip” under “Integrated peripherals” in BIOS (prevents NIC being loaded correctly)
• Disable “ACPI” under Power Management (from memory)
• Download IPCop at IPCop.org and burn to CD
• Boot CD
• Follow prompts to install (you may find this Howtoforge useful)
• Connect to web admin page
• Install Copfilter (for Virus protection and more)
• Install Advanced Proxy (for user authentication and control)
• Install URL Filter (for content filtering)
• Install Zerina OpenVPN
• Configure Zerina OpenVPN

Install Copfilter

Download Copfilter.

# scp -P 222 copfilter-version.tgz root@ipcop_green_address:/root
# ssh -p 222 -l root ipcop_green_address
# cd /root
# tar xzvf copfilter-version.tgz
# cd copfilter-version
# ./install

This should complete with a message such as this:

Copfilter ~version~ installation completed successfully!

Install Advanced Proxy

Download Advanced Proxy.

# scp -P 222 ipcop-advproxy-version.tar.gz root@ipcop_green_address:/root
# ssh -p 222 -l root ipcop_green_address
# tar -xzf ipcop-advproxy-version.tar.gz
# ipcop-advproxy/install

Install Enhanced Proxy Log Viewer

Download Advanced Proxy Log Viewer.

This package enables the capability to show usernames within the proxy log viewer.

Copy the file proxylog.dat to the directory /home/httpd/cgi-bin/logs.cgi and set the file permissions to 755 (rwxr-xr-x).

Install URL Filter Add-on

Download URL Filter.

# scp -P 222 ipcop-urlfilter-version.tar.gz root@ipcop_green_address:/root
# ssh -p 222 -l root ipcop_green_address
# tar -xzf ipcop-urlfilter-version.tar.gz
# ipcop-urlfilter/install

Install Zerina OpenVPN

Zerina OpenVPN.

If you are updating, first stop the OpenVPN Server through the GUI.

These instruction include a dangerous bodge to arbitrarily alter the IPCop version requirement, only proceed if you are happy to accept the inherent risk in so doing.

# scp -P 222 ZERINA-version-Installer.tar.gz root@ipcop_green_address:/root
# ssh -p 222 -l root ipcop_green_address
# mkdir /root/zerina
# mv ZERINA-version-Installer.tar.gz /root/zerina/
# cd /root/zerina
# tar -xzvf ZERINA-version-Installer.tar.gz
# sed -i s/1\.4\.18/1.4.21/ install
# ./install

Configure Zerina OpenVPN

• Zerina OpenVPN

Configure OpenVPN for Network-Manager

Download and extract the ZIP file from IPCOP OpenVPN, and extract the contents into (for example ~/.openvpn).

$ sudo apt-get install network-manager-openvpn
$ cd ~/.openvpn
$ sudo su
# openssl pkcs12 -nocerts -in default.p12 -out key.pem
# openssl pkcs12 -nokeys -clcerts -in default.p12 -out cert.pem
# openssl pkcs12 -nokeys -cacerts -in default.p12 -out ca.pem
# exit

These converted files can then be used in Network Manager.

If you’re struggling to connect to OpenVPN with Network Manager, consider that you need to ensure that the settings are mirrored on both IPCop and on the client’s Network Manager, for example I struggled to connect until I realised that LZO Compression was switched on on IPCop but off in Network Manager.

Static Routes

There is no way to set this in the GUI, so connect via SSH (see above):

In /etc/rc.d/rc.netaddress.up add:

if [ "$1" != "NOTGREEN" ]; then
if [ "$GREEN_DEV" != "" ]; then
    ifconfig $GREEN_DEV $GREEN_ADDRESS netmask $GREEN_NETMASK broadcast $GREEN_BROADCAST up

    ####CUSTOM EDIT FOR STATIC ROUTE####
    route add -net 10.100.1.0 netmask 255.255.255.0 gw 10.2.0.1
    ####CUSTOM EDIT FOR STATIC ROUTE####

Identifying Network Interfaces in Linux

If your network card supports it, you can blink the LED on the specified network interface in Linux like this:

# ethtool -p eth# 5

This blinks the LED on the interface for five seconds-without interrupting network traffic.

• Cool Command

Flushing the DNS Cache

There may be times that you wish to flush the DNS cache. IPCop uses dnsmasq, log onto ipcop using root and run the following:

# ps -ef | grep dnsmasq
# kill -s SIGHUP <PID> (replacing <PID> with the PID shown by the previous command)

References

• IPCop.org
  • User forum
• Howtoforge
  • Continued…
• Addons
• DansGuardian
• IPTables on Wikipedia
• IPCop Review
• Snort
• Installing OpenVPN
• Resolving add-on conflicts

$ dd if=/dev/sda bs=1M | ssh root@blah "cat > /root/disk.img"

Haven’t tried it yet, but it sounds incredible. The bs=1M is essential or the copy will take forever.

For example,I’ve just found that Kmail will not auto-complete email addresses from an LDAP KDE resource. A simple bug; although strictly speaking it’s not a bug, as it was a deliberate feature. Fortunately there is a fix, with a modification made to one of the components. Unfortunately this fix has not found its way into Debian Lenny or Squeeze, and Sid is now KDE4.

I could install from the KDE SVN repository, but that version has not been tested with Debian, and I don’t really want to have such a significant part of a production server to have come from outside of Debian. Perhaps I am being overly cautious.

Anyway, I changed to /usr/src and ran apt-get source kdepim to download the source from Debian. I used “find” to locate the offending component, and used diff to determine what differences there were between the modified KDE version and the standard Debian version. I was pleased to note that only one line had changed. I copied in the new component and ran apt-get build-dep kdepim (with help from Dominic in #surrey). Then a quick ./configure, a very slow “make”, and a quick sudo checkinstall make install and the fix was made.

There are many many more benefits to open source than the ability to modify the source, but I just wanted to point out that, even for non-developers, there is a very real and very direct benefit to having the source.

May the source be with you.

Update - the actual patch may be found here.

Our Debian Lenny LTSP5 Install has hit a road block: currently the site are using a Citrix server at another location over a 256k connection to pick up their emails. I really didn’t feel that I could ask them all to collect their emails over a 256k connection, so what are the alternatives?