05.09.08
Not Good
Checked the system logs yesterday and saw a bunch of write errors. Checking the RAID status showed this:
# cat /proc/mdstat
Personalities : [raid6] [raid5] [raid4]
md1 : active raid5 hdg[2] hde[3](F) hdc[0]
398296960 blocks level 5, 32k chunk, algorithm 0 [3/2] [U_U]
unused devices:
Normally, the last part of line 3 shows [UUU]. The underscore means it’s a bad drive. Unfortunately, I don’t know exactly when the drive fails, which just illustrates the need to have some kind of log monitoring. Ok, so now I have to figure out how to get EVMS to replace a drive. I want to do it as quickly as possible because another drive loss means a total loss of data. Which brings up another good point: RAID isn’t a substitute for backups (which I’ve been delinquent in doing). So I went to to my local Fry’s to get a replacement drive, but they don’t seem to make 200 GB drives any more. I ended up getting a 250 GB but the array will only use 200 GB of it. The other caveat is that the failed drive is IDE and the replacement is SATA. Will it work? In theory it will, I guess I’ll find out soon. Oh yeah, and I hope I pull the right drive out (they all look the same in the case). This is why making backups is a good idea. Meanwhile, changing out drives means shutting down the system. There goes my 81 day, 13 hour up time.