>> I think you've got really serious filesystem-level problems. Have you >> tried running any hardware diagnostics? Are you sure you're using a >> stable kernel version?
> I run fsck on the filesystem (gfs) -- no problems found. The disks are from > a san and the diagnostic programs say there's nothing wrong. I also have > other db clusters running on different filesystems (also gfs) and I have > never had any problems with them.
Some RAM checks wouldn't be out of place either.
Memtest86+ has now been running for 20+ hours and no errors has been found. I was also unable to reproduce this problem, but it only happened after a few days of constant activity anyway so I guess it's not so easy to replicate. Any other pointers where to look at? Your help is well appreciated.