Geek with no Week!

Why was Wednesday so bad? Ian G - HOST - 1280th post - 12 Mar 2004 18:10
As promised in 'Not rocket science' here's a more technical outline of what failed:
"While someone was manually tarring up (compressing) the archived messages in the INN database the disk ran out of space. This was due to a bug in Solaris that doesn't report that correct disk utilisation. So when the tar file was just meant to squeak in under the max disk utilisation it didn't. When the INN server next wrote to its control file that file was corrupted. When the INN server tried to read it it couldn't and kept getting stuck in a loop.

Internet Services had to stop the server, manually delete some files, unmount the file system and fix the corrupted files. Then they had to restart everything."