fouldsy.com

Although I’ve ran CentOS on servers in the past (Open-Xchange, for example), I’ve never used Fedora much myself. Still pay attention to what they’re up to with Fedora though, as although Red Hat may be a tad slow + immobile at times due to their size, they do sometimes bring out cool tools. Their latest release of Fedora Core 6 drew me to Virtual Machine Manager, a GUI for Xen virtual machine management. I still have my Xen server sat by my feet (am kicking it now, for no particular reason!) and keep meaning to mess around with it again, and although I wouldn’t use virt-manager much with only a handful of VM’s, it looks very cool if you were running multiple VM’s, such as for web hosting. Although remote management, i.e. connecting to remote hypervisor isn’t quite up + running yet, it looks very promising.

Picked up a Wacom Graphire 3 tablet yesterday which is a pretty cool piece of kit. Had quite a few of these running when working at Greencroft in the art department, and was impressed by them. With a new hush hush project being worked on (apart from those that have had sneak peeks…), it’s ideal, as is so much easier doing graphics work with a pen + tablet rather than a mouse.

Running off the desktop rather than the MacBook, as even with 2Ghz RAM, it doesn’t like manipulating 600-700Mb graphics files, so back to trusty Kubuntu and the GIMP (no jokes please!). Works like a charm, and sod the 2 CD’s of tablet drivers and associated software to allow it to work properly. Most modern kernels will have wacom support compiled as a module, so it works as a regular hotplug USB device straight out the box. Did add the following into the xorg.conf though (thanks to http://www.shallowsky.com/linux/wacom.html) to ensure it used the full size of the tablet and the eraser:

# Wacom graphics tablet
Section "InputDevice"
    Identifier  "stylus"  
    Driver      "wacom"
    Option      "Type" "stylus"  
    Option      "Mode" "Absolute"
    Option      "USB"  "on"
    Option      "Threshold" "10"
    Option      "Device" "/dev/wacom"
EndSection
 
Section "InputDevice"
    Identifier  "eraser"
    Driver      "wacom"
    Option      "Type" "eraser"
    Option      "Mode" "Absolute"
    Option      "USB"  "on"
    Option      "Threshold" "10"
    Option      "Device" "/dev/wacom"
EndSection
 
Section "InputDevice"
    Identifier  "cursor"
    Driver      "wacom"
    Option      "Type" "cursor"
    Option      "Mode" "Relative"
    Option      "USB"  "on"
    Option      "Threshold" "10"
    Option      "Device" "/dev/wacom"
EndSection
# End Wacom section
 
Section "ServerLayout"
[ ... ]
        InputDevice    "stylus"   "SendCoreEvents"
        InputDevice    "cursor"   "SendCoreEvents"
        InputDevice    "eraser"   "SendCoreEvents"
[ ... ]

Definatley recommend a graphics tablet to others now I’ve spent a bit of time using one rather than simply seeing them in action elsewhere, and even when I’m skipping between apps or browsing the net, I’ll quite happily use the stylus. Helps combat the old RSI too I guess!

And no, you’ll have to wait to see what I’m working on

I was looking over some of the events taking place at LinuxWorld UK Expo at the end of October, and some of the workshops and seminars looked pretty cool. Don’t think I’ll head down as can’t exactly fob it off as work-related anymore, and don’t really have money floating round to jump on a plane or train down to London for a couple of days. Went down just for one day a couple of years back, and it was a long, but rewarding day - would recommend it to anyone already around the London area.

But, one of the sessions is being hosted by ‘Chris Lewis MCSE‘. Nothing against the guy, don’t know him from Adam, and never met him, but why on earth would you want to brag about being an MCSE when speaking at Linux convention? I’ll wager there are speakers that are RHCE or LPI certified that aren’t detailing it on the programme. I’m not judging him, but can give a ton of people I know in the Linux community that will instantly turn-on off to someone holding Microsoft certifications, even if they’re promoting migration from Windows to Linux-based networks in the enterprise.

Wish him the best of luck with his presentation though - he might need it!

It’s not often I trackback to other blogs as I’m not a fan of posts that simply link elsewhere and you don’t actually end up with any story or breaking news. But, I’ve followed Jeremy’s blog over at Linux Questions for a while now, and his latest article was just beautiful!

Like many, even those that aren’t necessarily big Linux advocates, I’m bored with statistics and reports trying to prove Windows servers are more cost-effective, econimcal, stable, secure and easier to manage than Linux servers. It’s quite simply untrue. The quote in question claimed “Windows 2003 Server, in fact, led the popular Red Hat Enterprise Linux with nearly 20 percent more annual uptime“, which is like pulling a rabbit out of my man-sized ass. You’re needing to run into some *serious* problems to have a margin of uptime that different.

One sentence I just loved from Jeremy’s article was:

“When you get a bunch of Linux boxes setup by MCSE’s it’s no wonder why things came out like they did”

which is quite true. Even though many reports aren’t backed by Microsoft, they do seem to be being conducted primarily by Windows-based/trained engineers (if that’s even the correct description of these people). I doubt many of these reports that look at things from a test-bed environment have to actually run them in the real world.

But, the final sentence was wonderful, poking fun at the Yankee Group whom were behind the report:

“The completely hilarious thing about this is that the Yankee Group has absolutely no business even talking about uptime when they can’t keep their own web server up for more than 20 days. It incidentally runs Windows 2000.”

Check out Jeremy’s full post here - http://jeremy.linuxquestions.org/blog/_archives/2006/6/7/2013415.html

With Amarok throwing a wobbler and crashing once a day or so, and knowing that it’s only a couple of weeks until Dapper Drake makes it to final, is it worth waiting until 1st June or run an upgrade now and grab the shiny new Amarok 1.4 release?

I ran it when I first started playing with the 64-bit system and had Dapper, using Amarok 1.4-beta1 + beta2, and really liked some of the extra features such as *much* better media player support for my iPod Nano and Archos AV500, but since the rest of my system is so stable, I don’t want to risk upgrading just yet. Plus, I fired off some scripts a few weeks ago converting anything in .wma or .aac format to mp3 which means I’m not really needing Amarok 1.4 for updated taglib support for writing aac tags.

But, I’m a technology whore as someone recently describe me, and on cue, Amarok has just crashed

A little hiccup for a Monday. Not quite sure what went wrong, but in the middle of the morning our Open-Xchange server went belly up. No warning, no updates applied since the start of last week - just refused to handle user logins and dropped all existing sessions.

Easy enough to track down - the Java servlet couldn’t create a connection to the Postgresql database. After a bit of head scratching trying to figure out, simply re-installing ‘postgresql-jdbc‘ did the trick:

yum remove postgresql-jdbc
yum install postgresql-jdbc

followed by a restart of Open-Xchange:

/etc/init.d/openexchange restart

Bit of a puzzler as to why it suddenly stopped, as the correct PATH locations were still set and there hadn’t been any Java updates applied since the server was installed a couple of months back, but at least everything’s back to normal now.

Just got to figure out how to install Debian onto our new HP DL-320 webserver when they (quite helpfully) don’t ship with a CD drive… A nice ploy by HP to get you to then purchase a hot-swapable drive for £70-odd maybe?

Having spent a couple of weeks now playing with Xen, I have found it quite a struggle to find decent documentation explaining exactly how to install various Linux distros to act as virtual machines. After going through setting up Debian as my Xen server and a sample domU, any instructions on installing Gentoo, CentOS, Fedora, etc. all assumed you were already running those distros as the base distro. debootstrap could get me running Ubuntu, but that was about it and domi didn’t seem workable on Debian.

So, having trawled through some Xen newsgroup posts, I stumbled across jailtime.org, which provides pre-compiled, compressed images for CentOS 4.3, Gentoo 2005.1, Slackware 10.2, Debian 3.1, Fedora Core 4 + Fedora Core 5. As I’m not running my Xen server in a production environment, I don’t have any real qualms in putting my trust in someone not to have compiled something naughty into these images, though I don’t have any reason to believe there would be! Each distro is about 100Mb-120Mb in size, and each image takes about 5 minutes to extract and configure making it very quick + easy to have multiple distros running with the minimum of fuss.

Pick an image for the distro you require, and don’t forget to check the steps involved in using each image:

• CentOS 4.3
• Gentoo 2005.1
• Slackware 10.2
• Debian 3,1
• Fedora Core 4
• Fedora Core 5

A slight problem has been in that the sample configs are assuming /dev/sda to be the drive in use, and the images are pre-configured with fstab’s reflecting this, along with network configurations grabbing addresses via DHCP. After changing the Xen configs to use /dev/hda (or whatever hard drive layout you’re running) and then using ‘xm create‘ to load your new virtual machine, you’ll probably get an error thrown as the system can’t mount the drives, dropping you to a recovery console. Simply remount the drive read-write, edit fstab to reflect your hard layout, and reboot:

mount -o remount,rw /
nano /etc/ftsab
Ctrl-D

And bob’s your proverbial uncle Might look at expanding on some already-available scripts to handle the automatic creation of whichever domain is required, including the downloading of the image if not already available locally, though I’m already being distracted from at least two other little coding projects at the moment!

I downloaded Kubuntu 6.06 LTS over the weekend and it’s been sat in my backpack waiting to go onto the laptop. As I’m too used to being able to jump straight to a directory on various Linux boxes using fish:// or have WebDAV drives on-tap, I’d decided to stick Kubuntu onto my laptop for use at work. Having seen Dapper Drake running briefly on my desktop, it looked very slick.

But, it’s been a nightmare.

First off, I don’t agree with this idea of live CD’s and running an installer from inside it. The boot menu gives some nice options such as running a memory test and booting from the hard drive which previous versions didn’t, but why not then stick in an option to go straight to a normal install? Sure, you might want to give it a go to simply see how it runs, but performance will always be poor running from RAM + a CD, or for checking hardware compatibility, but I just want to be able to go straight to an install. Gentoo 2006.0 does the same thing, and failed to run through the installer a couple of weeks ago which annoyed me, though at least I could revert to a standard text-based install quite easily.

Anyways, aside from the annoyance of it loading up a live version of Kubuntu, the Espresso installer just didn’t work. Tabbing between menu items jumped all over, the actual order of the installer options didn’t seem very logical, and when the page for disk partitioning came up, the installer closed down. As it wouldn’t start again, I restarted the system. The installer wouldn’t run the second time round, so I just went to reboot back in to Gentoo.

“Grub error 22“. Bugger.

Quite impressively, the live CD managed to erase the hard drive. Not just corrupt the bootloader or partition table, but actually wipe the drive. Even Knoppix couldn’t repair it! Although I easily dropped a Windows image across the network back onto it (minus a dozen or so Windows updates since I don’t slipstream them with running a separate system for handling updates with the RM network), it wasn’t too bad as I don’t store any work on it, but it took out a Gentoo install which took a while to do. For those that have used Gentoo, you’ll know what this means!

To be fair, the Kubuntu website is now (under)stating “there is a bug in the live CD installer known to cause data loss, please wait for beta 2 before using it“, but that wasn’t there when I downloaded it, and it must be one hell of a bug! Again, to be fair, it is a beta, but the Kubuntu team are normally pretty solid and this is not at all what I was expecting. Annoyingly, whilst running the live version of Kubuntu, it did seem to have ACPI working properly to detect power usage, battery life and temperatures so it’s not all bad news.

I think I’ll use a 5.10 CD and simply change the sources.list to update to Dapper rather than try running another flight or TLS CD, and can only hope people haven’t lost any data trying to run this release in order to help out with testing + development of what will hopefully become a great release!

After some advice from Frank (Osterfeld?) with regards to Akregator, I updated my system to KDE 3.5.2 yesterday to iron some problems I was having, atlhough I’ve been meaning to update KDE anyway. With Gregarius seemingly working quite happily, I was keen to now get a desktop reader working properly.

Upgrading to Akregator 1.2.2 seems to have made things easier to manage in terms of multiple selection + deletion, and I haven’t come across any problems with incorrect/duplicate/old items being delivered. I actually quite like Akregator sitting in the system tray quietly showing when new items are available.

Now it’s onto writing a plugin for Gregarius to handle the deletion of items the way I want it to. Already been spotted lurking in their IRC channel the last couple of days so figure I may as well do something about it Other than that, everything seems to be working perfectly. Will write up some quick instructions how to setup synchornisation between Gregarius and a desktop reader once I have some rough code for the plugin working.

A few people have asked what happened to my stack of Ultra 5’s and the Gentoo network servers. To be honest, I just didn’t have the time or drive! Although it was good fun at first, they were just so slow compiling source it got annoying. They’re still sat there and I occasionally play with them, but not much!

Instead, I now have my new test server up and running based off my old Athlon XP 2800+. It’s powered by Debian at the moment rather than Gentoo as I found a wonderful guide on setting up Xen and creating your virtual machines. For those not familiar with Xen, it’s an open source server virtualisation system along the same lines as VMWare, though without all the pretty buttons to set everything up! It allows you run multiple operating systems at the same time from the one system, meaning you cut down on hardware + maintenance costs, and for situations like I’m running, means you can run multiple servers from the one machine

Now that I’ve got it up and running it’s actually very, very impressive. The documentation is clear and they have very good IRC channel (##xen on Freenode), but I still spent a good while trying to get the networking going which had everyone stumped. Am good at coming up with these things! I had dom0 setup, which is your host server, and had created a couple of domU’s, your virtual machines. But, I had no networking from the virtual machines. The idea is your host system handles the routing using virtual interfaces and a network bridge, but it wasn’t working. Whenever you tried to ping anything, it would return “Destination Host Unreachable“, indicating something at least was working.

After checking through various settings, re-building virtual machines, running tcpdump to monitor traffic, etc. I finally stumbled across the problem whilst trying to manually create the bridge using the ‘/etc/xen/network-bridge‘ script. The cause? Xen looks for, and only for, a physical connection on eth0. It sets up peth0 which seems to handle the virtual routing alongside xenbr0, your network bridge. I was using eth2, not eth0, so although everything appeared fine, peth0 had no link with eth2 and so all domU virtual machines would have no network connections. Switch yer cable into eth0, bring the interface up and everything works.

Very annoying, and possibly something catching quite a few people out since there’s a number of identical problems strewn across forums, message boards and blogs. Gonna check if a bug has been filed for this, as Xen should check for which eth is active and link peth0 with it. Now that it’s resolved, I’m finding it very impressive - you simply ssh into your virtual machines as normal and since you have a standard Debian install running, can do whatever you want. Will start building on the network services over the next few days and linking the virtual machines together simulating a virtual server rack to see how Xen handles it, but it’s certainly a very powerful tool