Saw a great banner on "the greatest .nl forum of all ISP and webhosters" …

What’s that, a USB card reader with a dedicated server in it? If so, this could be a great solution for cooling and rackspace issues…

Now they really don’t have any reason not to support Linux 😉 *cut bs about FCC regulations and competitive advantages*

Linux Kernel Hackers write device drivers for free ! 

Via Slashdot, excellent reasoning 😉

Yes, indeed — Top Gear is back on BBC2. If you’re reading this now and didn’t know, you missed it. Maybe FinalGear could help you out… Welcome back Hamster, and Jeremy and James as well ofcourse! w00t! 🙂

testing.bitlbee.org is back since a few days… on a machine owned and hosted by yours truly. After the project suffered loss of their main testing public server due to hardware failure I stepped up to offer new hosting; as soon as the files were recovered from the other machine the server was put up again. Feel free to use it (or my other IM server, running stable bitlbee, at im.codemonkey.be) for all your Bitlbee needs. Happy IM’ing!

I want this ! Let me know if you see it somewhere, even if it’s more than $10 😉

Okay, that may be a slight overreaction, but there’s surely a surge of new versions fixing security holes these days ;(

*rant* 

Most of the time when I see someone asking how to kill a process on *nix I see someone responding "kill -9 <pid>".

Alas, this is NOT the way you kill a process … the process gets killed alright, and a little too hard. So for everyone not understanding, here’s a small explanation. Everyone who knows, just skip to the next post 😛

Normally when you (or something) kill a process, it gets a SIGTERM. When a program receives SIGTERM, it knows it should shut down and saves all its open files (or prompts the user), closes its file handles, etc etc. When you use -9, you send a SIGKILL which causes the OS to clean up the process without further ado (e.g. when it doesn’t respond to SIGTERM anymore), and shut it down. This also kills the process indeed, but doesn’t always have the intended effect when looking at data loss… Remember it 😉

*end of rant* 

Right when I don’t really have any time due to work (working until 9pm etc) some hardware started to fail…

– noc.phyxia.net died after I put back the second CPU. Jorne (hayano) brought home to me (thanks!), apparently the fan controllers on the mainboard are all failing, after a while the fas turn off and all fan readouts read 6335 rpm. I connected the CPU and Chassis fans to standard molexes and everything is working great now. It’ll be back up when I visit Versatel tonight. In the meanwhile its rails also arrived so it’ll be mounted properly too 😉

– datalock.powersource.cx, which is one of the most important servers in Brussels, not only failed a disk in its RAID5 raidset but there also seems something else to be wrong… It reboots every night at 0:30 with a CPU context corrupt panic. I’ll be replacing the CPU’s and CPU cooling, and possibly the RAM – hopefully it’ll then stay stable. I’ve got a 200GB Seagate disk but I’m afraid it’s not big enough to replace the 200GB Maxtor to fix the RAID.

– tequila.powersource.cx, which is in Amsterdam, broke a disk … this has been fixed yesterday, big thanks to my hoster Zylon for the remote hands. The box is now up’n’running again with a full RAID set.

 *argh* 🙂