Everything is a Freaking DNS problem

Fosdem is over .. and it was ... overcrowded :(

Honestly trying to squeeze into an overcrowded bar, then on saturday overcrowded rooms, or even not being able to enter that room (Mozilla and Embedded) , Fosdem is starting to become the victim of it's own success.

Some people are suggesting Fosdem to move to the Arenberg campus in "Brussels-East" dunnow if Leuven can actually host enough beds for Fosdem :)
However the Beer event problem would be solved but Philip will have to make arrangements with 'The greatest bar of Western Europe"

The talk about Xen on ARM was interresting however the grande finale missed, the MiniOS just didn't boot :( Kettle was interresting and I should start spending time with it :) But then again .. so are a zillion other things Too bad the SWOT analysis between Postgress and MySQL got cancelled.
but it left me some breathing space :)

The evening ended with a mixed crowd of local Linux geeks and Drupal folks in the restaurant on walking distance.

On sunday morning I realized it must be the Fosdem weekend when you are on the E19 direction Brussels around 0900 and there is no traffic :)

I was right on time for the Drupal 6 and 7 talks from Gabor and Dries , which off course meant I was going to be too late for Ian's talk. Luckily I catched the important parts. The virtualbox talk disappointed me .. this was a marketing talk for endusers, not a talk suitable for Fosdem :(

Pascal learned that integration Amazon basically is a fine dns problem :) Then after some chatter with the MySQL crowd I headed into the Conary talk.. I was expecting a bit more information around their rbuilder system
I wondered into too much talks on sunday afternoon . the MySQL Proxy talk , the end of atogs talk which he didn't want to repeat :) Karan's talk .. etc.

I had to miss Simon's talk about Posgtress HA on sunday for the obvious reason , but luckily I could catch him on saturday to get a short sneak preview..

But more about those obvious reasons in a separate post :)

My Pics are over here. You'll be recognising Jan Kneschke, LVB , Thomas Bonte , Geert Vanderkelen, Dries, Gabor,Matt Casters and others.

It seems like for the past 4 years February is the month that O'Reilly really loves me and decides to publish one of my articles.

This years version was cowritten together with my collegue Johan Huysmans and tackles the creation of Highly Available Gateways

Altough the every HA situation is different and this is a pretty easy setup it's a good start for other setups.

Enjoy the read

PS. Yes I know , in 2006 I also had a January article :)

Last century I was sometimes noticed to be using VMWare to run a weird platform on my Linux Desktop , or to run some test installations. With the introduction of Qemu and later Xen there was no need for me to use the proprietary closed alternative.

So I got the question earlier this week to build a virtual instance of our bootstrapping environment so that field engineers could take that virtual machine on their laptop and do installations from there.

One of the collegues told me that booting a fresh VMWare instance of the network would be no problem so I took up on the challenge.

First of all a registration procedure so the nice folks over at VMWare can spam me with their marketing stuff during the next couple of months. I needed to register as I seemed to need a Serial Number.. whow .. that was ages ago since I last needed such a beast.
A full 101Mb download later I had an RPM ready to be installed.

"Cool an RPM" I tought, that means that I will be able to clean up all the mess they leave behind with a simple rpm -e.
Wrong guess. it seems like VMWare starts copying around files in different places on your filesystem and actually even wants to compile stuff against your running kernel. All fine and well, but the result is a bunch of unidentified files that are cluttering my filesystem.

The next step required me to start the gui to create a virtual machine, suddenly the load of my machine skyrocketed. Load 11 whew.. yep the VMware process was the guilty one.. luckily it went back to normal after a while .. but I`m still not sure what happened there.

Click, click, click and before I knew it VMware was allocating 4Gb on myfilesystem for an empty VM. Noo.. don't allocate it yet.. retry.. ok .. take it when you need it .. that's better. And why do you think I only want 256Mb for this machine ? Hmm.. where's that config file.. aah.. clickerdy click again .

Start VM, F12 to boot of the network, wait, reboot done.

And thus we joined the era of transferring an unmanagable image that everyone will copy around wile slightly modifying things and never placing them in version control . hence ending up one day with something nobody knows how we got there..

What did we learn, that the way I bootstrap a Virtual machine or a physical machine really doesn't matter, and is still just a matter of mapping a MAC address to a HOSTNAME.

When working on the system, to check if the install was done correctly it felt like I was on a remotely bad connected machine, not really slow, but really slow but still (could have been because I logged on via the gui). Certainly not like the paravirtualized Xen machines I`m used to work with, more like the VirtualBox experience.

The big problem with giving someone an image to play with however is that you loose all control over what is being deployed , configured or changed and it just becomes a change management nightmare. At least reinstalling won't take much time.

But I`ll go back to KVM and Xen for my daily work.. I've seen enough GUI's for this month :)

For those of you reading my blog the old fashioned way with a browser you might have seen the ProfOss Virtualisation conference Badge popping up on the right.

For the others:

So come and hear me speak at ProfOSS next January !

Tarry has an overview of the commentaries about Oracles latest Virtualisation announcement.

Two things hit my brainwaves, First ,seemingly Larry is claiming that his Xen package is better than
the others since he supports Live Migration and all the others don't. I don't know where he gets the idea.. I have to admit I don't remember which year it was .. but it was somewhere in december that I first starting with Live Migratiion of Xen machines and it was also on a CentOS platform. No fancy gui, no hardcover manuals that had it all documented. But fast and seamlessly working live migration, ready for everybody to use.

Second one is he claming that since Xen was re-engineered by Oracle to be faster than the competition.
The way you read it there is that Oracle took Xen, modified it then started redistributing it.
Is that really what happened ? Are they redistributing the source, or are they violating the GPL ? Coz if they are
redistributing the source everybody just got a faster Xen and if they aren't ..
I don't know but there sure is room for rumour here. Or is this just a bunch of marketing people and IT journalist that are mispresenting the facts. Fact is that one needs to spend lots of time verifying the facts of stories one read on the internet today.

I`m also seeing people crying that Oracle is finally stepping to the open source side.. .. I`m wondering on which planet they have been living.. Oracle has been supporting different Open Source products for ages already and they even are the owners of core components of typical daily used packages, so where do those authors get the idea that Oracle is finally stepping over ?

Tonight I don't have time to go and hunt back the original sources and see who actually said what . Honestly I don't care that much about Unbreakable, Last time I tried caring was when one of my european customers was interested in it, and no one at Oracle.be could help me . It's Like Dell shipping Ubuntu.. not on this part of the world.. and sadly it's also like the OLPC .. Give one Get one only on the other side of the Ocean.

It's all good marketing .. until you actually want to get it .. :(

Just give me the source code .. then I don't have to sit trough presentations from marketing people that can only read from spec sheets for some product , but fail to read the fine print and leave me disappointed again ...

Henning Sprang seems to have a different view on the concept of a Vendor LockIn than I do

On his blog The daily laziness: OpenQRM vs. vendor lock in??? , he describes how openQRM Locks into using .. openQRM.

Well, not really .. at least not in my opinion, although openQRM still has a long way to go and the proposals for a meaner and leaner Henning gives are certainly valid,

openQRM however does not force you to do anything you don't like. The source is available and free, you can modify its behaviour , you yourselve can spend time on in, learn and modify the platform no one .
(given the complexity I agree you won't dive into it in just 5 minutes but you nothing (b)locks you)

Certainly in the virtualization field openQRM gives you the freedom to migrate your machines from one virtualization technology to another, Today VMWare, tomorrow KVM, the week after Xen
or all 3 or them from one management console (can come handy after a merger) all with the same interface. None of the commercial products out there will even think about giving you a GUI to manage their competition.

On top of that when you go out and buy VMWare , you can only manage VMWare instances and you get a framework that will also force you to work in a way you can't change. You have only one company to talk to, VMWare, (maybe via its integrators but they can't fix issues , they don't have the source) with openQRM you can hire other people besides Qlusters , you can change to another organisation to support you.

Try that with VMWare , I fear you are Locked In

I wonder if I shouldn't totally switch careers and become an industry analyzer. Traveling to all the nice places , getting invited to all interesting parties just by predicting
what will happen in the future. And all that offcourse for a big fat paycheck

It's not like this is my first prediction coming true

I don't know on which planet I have been for the past couple of years , days or hours but since when do
VMware’s Vmotion, XenSource’s Xenmotion or Virtual Iron’s Virtual Iron support Replication ?

Live Migration yes, but Replication , No.

I discussed this kind of technologies with Mark and Vincent , Moshe and others already a zillion times.. Continuously mirroring or realtime replication of a virtual machine is really difficult to do. And I haven't heard from a working scalable solution yet .. (Shared Memory issues such as we had with openMosix still are amongst the issue to be tackled)

Live Replication would mean that you mirror the full state of your virtual machine realtime to another running virtual machine. Every piece of disk/memory and screen you are using has to be replicated to the other side of the wire realtime. Yes you can take snapshots of filesystems and checkpoints of virtual machines. But continuous checkpointing over the network , I'd love to see that.. (outside of a lab)

So with a promise like that .. our good friends the CIO will be dreaming and the vendors will be blamed for not delivering what was promised to them.

But on the subject of using just Live Migration features as an alternative for a real High Availability solution , I know different vendors are singing this song, but it's a bad one.

Using Live migration in your infrastructure will give you the opportunity to move your applications away from a bad behaving machine when you notice it starts behaving badly, hence giving you a better overall uptime. If however you don't notice the machine is failing, or if it just suddenly stops working, or if your application crashes you are out of luck.
Live migration won't work anymore since you are to late, you can't migrate a machine that's dead. The only thing you can do is quickly redeploy your virtual machine on another node, which for me doesn't really qualify as a Clustered or HA solution.

Real HA looks at all the aspects of an application, the state of the application, the state of the server it is running on and the state of the network it is connected to. It has an alternative ready if any of these aspects fail. Session data is replicated, data storage is done redundantly and your network has multiple paths. If your monitoring decides something went wrong another alternative should take over with no visible interruption for the end user. You don't have to wait till your application is restarted on the other side of the network, you don't have to wait till your virtual machine is rebooted, your filesystems are rechecked and your database has recovered no it happens right away .

But Virtual Machine Replication as an alternative for HA ? I'd call that wishfull thinking and vapourware today

I have to admit .. I prefer being called a good soul over other things :)

Before I forget again

Initrd for domU

mkinitrd --preload xennet --preload xenblk /boot/initrd-2.6-xen.img 2.6.18-8.1.8
.el5xen

Now at least I can google for it again :)