March 27, 2002
Well this is new...

Yeap, there were some..... "changes" over here at Basically we had (another) hardware problem. The hard drive that was replaced about 4 months ago decided to start making awfulbadugly noises, leaving the box in a bad place.

Short story is that the drives and some other hardware was replaced, but with some data loss and other messed-up-idness.

Read on for more...

First of all a huge thanks to a few people. Yohimbe ( sysadmin extraordinary) for mucking with hardware in his living room till 1am and Tom for coming in on a day off, and staying late another day to let me in to get the box out. Oh, and I won't forget Fred (the TDI guy) who donated some of his own time and $ getting the box up and going with it's new hardware.

The box is now back up and running with it's busted 30G drive replaced with 3 40G Maxtor drives using RAID 5 for speed and redundancy, and a nice Antec AX-1240 tower case (reviews) with a 400w power supply (one theory for the blowout was bad power, as the box was running on a 250w power supply.

The bad part about this crash was twofold. First of all, the backup script was backing up /var/spool/mail, a symlink to /var/mail. The script did exactly what it was supposed to do, and every night backed up a symlinks :( With the last hard drive problems I had mail on a separate partition, so it was all ok, but not this time. So anyone who left any mail in their main INBOX spool, sorry, but it's gone now.

The other thing was the databases, which relates to why you're reading this on a nearly default Movable Type page :) See, phpnuke was driven from a database. Every night the database backed up by script. Well, due to some changes in the way that somethingorother changed, the programs to do the database dump every night were failing. I noticed this, but never got a chance to go and fix it because "hey, it's a new hard drive, what could possibly happen?"

Well, guess I found out huH? Anyway, the mysql database's last good backup was around February somethingorother, and postgres's was around March 7. So a lot of stuff from ufies got lost, as well as other things (peer 2 peer being the main one. That one wasn't all that bad as it was postgres, but still...

Anyway, so that's the story. If something is messed up, or you don't have access to something you used to (shell users) like a database, or whatever, email me at If you lost data (by the way, I hope I mentioned that the /home partition is just fine, so any data in your $HOME is exactly as you left it), or mail, I'm sorry.

As you've probably noticed, this is a pretty default thing, so bear with me as I re-do some things, and learn the ins and outs of MT :) Now, back to getting it all back up and running!

Posted by Arcterex at March 27, 2002 11:37 PM