Dave Watkins wrote:
As an update, the PSU was rated for 230W. The Kill-A-Watt said that the PSU was drawing around 130W or so. I had a spare Antec Sonata with a 380W PSU, so I swapped the stuff into the new case. Problem still occurs.Finding anything conclusive from the PSU is tough without spending some money on hardware to monitor the rails. Mostly because your PSU might be rated at a certain level, but due to heat and spikes during it's life it may get nowhere near that.
I found a script that just creates a 200MB file from /dev/zero, makes a few copies of it, and then uses 'od' to check if there are any non-zero values that have crept in. It's not the best check in the world (I'll probably switch to using md5 or something on the files), but it did reproduce the problem.
I have six *physical* drives: 2 IDE, 4 SATA. 3 of the SATA's are in a RAID5, so there are 4 *logical* drives.Thus far, running my torture script on either of the SATA-based logical drives has led to filesystem corruption and to the kernel forcing the mount to be read-only within about 5 minutes. I've been running the same script on one of the IDE drives for a half-hour now and no problems.
So, until any new data points come in, it's looking like the SATA card... - Joe
Description: S/MIME Cryptographic Signature