Dr. Christopher Kunz

I have recently installed SilverStripe to test if it fits into our shared hosting concept. Read after the break about my impressions.

Für alle, die mal wissen wollen, in welchen Dimensionen Google rechnet, kann sich in diesem Artikel über die “Google I/O”-Conference erleuchten lassen. Highlights sind:

Dean seemingly thinks clusters of 1,800 servers are pretty routine, if not exactly ho-hum.

Oder auch:

In each cluster’s first year, it’s typical that 1,000 individual machine failures will occur; thousands of hard drive failures will occur; one power distribution unit will fail, bringing down 500 to 1,000 machines for about 6 hours; 20 racks will fail, each time causing 40 to 80 machines to vanish from the network; 5 racks will “go wonky,” with half their network packets missing in action; and the cluster will have to be rewired once, affecting 5 percent of the machines at any given moment over a 2-day span, Dean said. And there’s about a 50 percent chance that the cluster will overheat, taking down most of the servers in less than 5 minutes and taking 1 to 2 days to recover.

Diese Ausfälle sind jedoch - zumindest auf Storageebene - kein Problem für Google, denn ihr eigenes verteiltes Dateisystem GFS hat für solche Fälle eine feine Fehlerbehandlung:

GFS stores each chunk of data, typically 64MB in size, on at least three machines called chunkservers; master servers are responsible for backing up data to a new area if a chunkserver failure occurs. “Machine failures are handled entirely by the GFS system, at least at the storage level,” Dean said.

Und wenn man genügend Rechner abnimmt (laut nicht von Google autorisierten Schätzungen mehrere hunderttausend Rechner), dann baut Intel auch spezielle Mainboards, so daß die Probleme, die mit konventionellen 19-Zoll-Racks einhergehen, durch speziell gebaute Racks umgangen werden können. Das ist allerdings keine Magie, denn andere machen das auch so (1&1, soweit ich mich erinnere, hatte vor Jahren komplett eigene Gehäuse, um die Rechnerdichte auf mehr als 1 Server pro HE zu bringen).
Ich denke, wir werden noch ein bißchen üben müssen, bis wir in diese Größenordnung vorstoßen.
(via fefe)

The Hardened-PHP Project is happy to announce another success story for Suhosin: Finally, we have the suhosin extension in Debian/Unstable. If you have the 4.0 version of Debian, you can now use the following packages to install PHP4 / PHP5 with Suhosin support:

irc:~# apt-cache search suhosin 
php4-suhosin - advanced protection module for php4
php5-suhosin - advanced protection module for php5

Great work, Debian guys!

I am going to speak on the PHP Conference UK which will be held at South Bank University, London, UK on February 10th, 2006. Yes guys, that’s in less than three weeks

php.net has the following to say about the event: “Not bad for 50 quid”, and I think that’s kinda true. Amongst others, book author Harry Fuecks and Derick “the ez photographer” Rethans will be speaking, so register now! The conference will be the first of its kind in the UK and anyone who’s interested in PHP should seriously consider joining us for the 10th. You should, however, expedite your decision - Early Bird Discounts are only on offer until February 3rd.

My talk will focus on the Hardening-Patch for PHP, outlining the usual stuff (obtaining it, installing it, configuring it) as well as some real-life demos of what it does and how the patch has influenced the mainline PHP distributions. I will also be available for signing CACert accounts and general networking.

And, to top it off, I’ll spend a few more days in London to visit all museums and places I didn’t get around to during my stay over New Year’s.