I'm running Ubuntu 12.04. Initially, I had 3 different HDs retained from various upgrades. Suddenly, the oldest and slowest started generating I/O errors in syslog and, considering it to be dead, I removed it and the errors went away.
A little while later, I upgraded the main HD and kept all the old ones under various mountpoints. Within about 2 weeks, I was getting I/O errors and took the drive back to the shop for a replacement (Caviar green from memory). They confirmed it was dead and I got a new one that worked fine.
A couple of months ago, I started to get I/O errors on one of the older slave drives. A bit of experimenting showed I had a very dirty fan (3rd fan in the case) and, wondering if this was causing a power drain somehow I replaced it. Problem solved. Until a week or so later when I started getting the I/O errors again. Unplugging the 3rd fan solved the problem.
About 2 weeks after that, the I/O errors started again. This time it was resolved (intended as a temp fix) by unplugging the 3rd HD.
Today, I got a second monitor. Plugged it in, and got it configured. Changed some WM settings and rebooted and got errors on startup - one of your disks has errors, unable to mount HD with UUID XXXXXXX, unable to mount drive for fstab entry /home/media/Photos (that sort of thing). And the syslog is full of these:
Jul 20 18:44:49 local kernel: [ 486.236060] ata4: SATA link up 1.5 Gbps (SStatus 113 SControl 310) Jul 20 18:44:49 local udevd: '/sbin/blkid -o udev -p /dev/sdb1'  terminated by signal 9 (Killed) Jul 20 18:44:49 local udevd: timeout 'udisks-part-id /dev/sdb1' Jul 20 18:44:49 local kernel: [ 486.244446] ata4.00: configured for UDMA/100 Jul 20 18:44:49 local kernel: [ 486.244473] sd 3:0:0:0: [sdb] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE Jul 20 18:44:49 local kernel: [ 486.244483] sd 3:0:0:0: [sdb] Sense Key : Aborted Command [current] [descriptor] Jul 20 18:44:49 local kernel: [ 486.244493] Descriptor sense data with sense descriptors (in hex): Jul 20 18:44:49 local kernel: [ 486.244498] 72 0b 00 00 00 00 00 0c 00 0a 80 00 00 00 00 00 Jul 20 18:44:49 local kernel: [ 486.244517] 00 00 08 18 Jul 20 18:44:49 local kernel: [ 486.244525] sd 3:0:0:0: [sdb] Add. Sense: No additional sense information Jul 20 18:44:49 local kernel: [ 486.244534] sd 3:0:0:0: [sdb] CDB: Read(10): 28 00 00 00 08 18 00 00 08 00 Jul 20 18:44:49 local kernel: [ 486.244552] end_request: I/O error, dev sdb, sector 2072 Jul 20 18:44:49 local kernel: [ 486.244562] Buffer I/O error on device sdb1, logical block 3 Jul 20 18:44:49 local kernel: [ 486.244596] ata4: EH complete
For now, I've unplugged this HD too. Each time I've continued to run with these errors for a while, the HD has ended up dead, and this one has a lot of backed-up, but otherwise irreplaceable, photos and videos on it.
During the same period I've had a PSU burn out and been replaced with a higher wattage model, I've spent time looking into the syslog error messages and found others who have similar issues, but usually for a particular make of drive and my problems have occurred with all sorts. The machine is not particularly high spec (Gigabyte GA-MA74GNT mobo, Phenom II X4 955, 8GB RAM, NVIDIA GeForce 210 pushing a 1920x1080 and a 1240x1024, DVD writer, 1 HD and 2 fans) although the PSU is not a particularly high quality one (DTech 450W) but online power calculators suggest it should have enough wattage. The PC is used for dev work which can be a bit peaky on CPU load at times, but averages about 0.8 to 1.0 (on four cores). The problems seem to occur whether the PC is powered up for a few days, or whether it's powered down when not in use.
So, please, does anyone have any obvious ideas of what to look at to fix this. I'm suspecting the PSU, but I'd like to be more sure about this before I start replacing stuff at random.
NB: Drives are recognised by the BIOS, just not during Ubuntu boot.