Everything is a Freaking DNS problem

Now this really is an annoying DNS problem :)

To start with the Postgresql talk , I ran in just too late as the slides about slony were just finished :(

Skipped out on Nigel's Talk, after not having time to play with the OpenPower at the office I`m assuming that Macbar knows enough about it already ,
Then the awkard break before your own talk it's the waiting and the people dropping in one by one , It's much easier to start directly after someone else :) Ray introduced my talk as "The talk you have all been waiting for" , hmm.. he shouldn't have done that. It made me fall back into my old bad habit of speeding trough my slides in just about 35 minutes rather than the scheduled 45 I had.

I'm starting to hear myselve thinking during a talk that I gave before "Hmm.. I've already said this" then realising 2 slides later that I actually didn't say that yet during that talk but during a previous one. That really might have confused people. I really have to work on my presentation skills.

But then again it's the content that counts and discussing other ideas with people afterwards, actually much of the discussion already happened before :)

Back up the hill, to Collingwood College, for lunch then back down the hill for the Kerberos talk. Lunch was good at this conference, the only donwside being that you had 1 hour time and you also needed about a 5-10 minutes up the hill and back down which didn't really give you the opportunity to eat and rest.

I sneeked out during the break to get a drink, and the machine ate my money :((

Alex has been quite active on de MySQL Cluster list so I`m kind of interested to see what he has to say about it. Hmmm. what's a NetApp Filter ? :)
He is giving a general introductory talk on LAMP HA, nothing really new there for me .. turns out he is about to get his book on MySQL Cluster published really soon now.
I spoke to him afterwards on what to expect from the book, mainly regarding MySQL 5.1 etc..
The Register's talk on their Apache infrastructure is matchining really with his paper, good talk, but as he concludes they are extremely lucky that their site is mostly static, this ain't working with fully personalised an dynamic sites.

Pieter also just posted his report .. which is quite normal as my laptop is bridging his internet access ;)

The next couple of blog entries have been written offline and pasted here when I finally had time and network access, so some stuff might be out of sequence :)

Pieter and I flew from Zaventem to Newcastle on tuesday, then the Metro to from Newcastle Airport to Newcastle Train Station , then the train from Newcastle Station to Durham station, then the Bus to Collingwood college where we are staying.

So here we are now UKUUG Spring Conference Day 1 :)
Our first talk was the LDAP tutorial, the lack of IP feed is really annoying, seems the Durham Uni is using some kind of wifi auth that is only working for some weird platforms ..
The LDAP tutorial was fairly good, altough the number of things I learned from it was minimal, but that mostly to blame myselve as I've been working with LDAP for a couple of years now

Something went wrong with one of the keynote speakers laptop so the only keynote we got was from James Youngman from Google crawling and indexing
Some notes, according to Google "HA comes in for free, with scaling" , now how comes that when I say that HA shouldn't be more expensive when taught about upfront people don't believe me but when Google says it ..
There is a paper about GoogleFS which is in userland not in kernelspace on labs.google.com
Use cheaper hardware, use multiple instances, replicate , hmm.. that sounds familiar to the the kind of architecture I'm working with daily :)
Automate yourselve out of a job each 18 months, yep that's true that way I can move on to doing new and other interresting stuff and spreading knowledge ..

Then it was Pieter's turn to talk about Juliux, we really need to get that domain registered and small site to get started with. People are really interrested by the concept and some even want to contribute it seems, great stuff

Then the CFengine talk , I was really looking forward to this talk ... oh my .. this Kpresent presentation is the most annoying I have ever seen.. what's with the baby and dolphin pictures in this presentation ? Apart from that the presentation was
good for an introductory talk I was hoping to learn a lot here.. well it didn't happen.. I learned some stuff however I`ll be looking at replacing filereaper cfengine soon on some platforms,

My biggest question was how do you do the Initial key distribution ? Seems like you can do something with cfrun initially, I have to look into that

Ahhr.. the network really, really sucks here.. even the wired stuff in the terminal room upstairs doesn't work.. , after the "registration" I should be able to access the network but it keeps telling me I`m not registered, however when I try again it tells me I`m already registered, let's not even mention the fact that wifi was only available for some proprietary platforms, and it didn't even work for these people.

The Exim talk , hmm I should look into ho how this stuff can be done on Postfix also :)

Finally I figured out I had ip connectivity at my room in Collingwood College but as we had to hurry to the Conference Dinner then to "Vix" these postings will still be late ;)

If you want to read a short and clear intro to HA , and terminology you should have a look at Werner Fisher's first chapter of his thesis "Implementation of a Disaster Resilient Linux Cluster ..." , but just the first chapter :)

SourceForge.net: David Santo Orcero Returns to User Tools Development

Yeah. David (our original userland maintainer) promised to be working on the 2.6 userland tools also.. this is good .. !

Warning, commercial content below :)

Newtec is looking for young potentials with an interest in Open Source and Software development.
For that matter they are organising a Workshop to meet those people where I will be talking about the Xen usage at Newtec and Wouter Zoon's of Andromda fame will be talking about Model Driven Architecture at Newtec.
Meet us there

I just caught myselve typing "dd if=/dev/zero of=/dev/sda" on one of my test machines... luckily I did't hit the return key .. I use the above mentionned command to make sure a device boots from network after I reboot it hence bootstrap again, looks like I have been testing that kind of stuff too much recently ;)
A reboot did just fine to get this machine in to 2.6.15-openmosixbeta ;)

While testing FC5 my /home was flooding and I at first didn't really look at it. Then later I experienced often high loads on my machine by some mono process , upon killing it nothing really naughty happened. Didn't pay further attention.

Then a couple of days ago.. I mentionned it to Philip and figured out it was beagle , I therefore removed beagle from my default session startup, problem solved .. I tought ..
Until just now my disk was full and I really knew I hadn't been downloading large stuff recently. So I started figuring out where my space was gone
How come I wasn't surprised that my $HOME/ contained a .beagle that was abusing 4.6Gb of disk space ?

I tought I disabled it and guess what.. there it was .. still running

sdog@mine2 ~$ ps -ef | grep mono | wc -l
88

Euhm ? So let's disable this thingie.. ah it has a control panel. I want no indexes whatsoever and no logging. Ok.. Submit.
40 minutes later .. yep .. again 350 Mb of Indexes in that directory I solved the problem "permanently" now

root@mine2 ~# rpm -e beagle

I really hope this isn't the prequel of lots of mono problems on our favourite platform
Off course the Gnome people will blame the Fedora people but still there is no need for 88 running instances of one binary

Last weekend we had a small Belgian demoscene reunion, meeting up with people I hadn't met for over 10 years but back in those days we travelled Europe to go to events such as Assembly, The Party etc. As expected most of us ended up with some kind of computer related job, some of them even ending up in the Game development industry just as they always wanted. It was weird to be discussing differnt shading techniques again , discussing old demo's again and trying to figure out at which party we couldn't sleep due to the techno party that was being held in the sleeping room.

The fact that I havent played a computer game for ages really showed , when they were discussing the Unreal 3 engine I really tought I had missed out on a new demo Future Crew had released after Second Reality.. :)

I`m tired of waiting so my quick fix was

127.0.0.1               localhost www.google-analytics.com

Now lots of pages load quicker again ..