Small server running SLES9 and OpenExchange fell over yesterday. it was still reachable by ping, however no sevices were responding (SSH, SMTP, SMB etc). This is on a remote site so I don't have physical access to the machine, I had someone on site reboot this morning. The following thins are noted in /var/log/messages:

Dec 7 09:37:09 ZurichMail kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Dec 7 09:37:09 ZurichMail kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=95329888, high=5, low=11443808, sector=95329888
Dec 7 09:37:09 ZurichMail kernel: ide: failed opcode was: unknown
Dec 7 09:37:09 ZurichMail kernel: end_request: I/O error, dev hda, sector 95329888
Dec 7 09:37:09 ZurichMail kernel: ReiserFS: hda5: warning: vs-13070: reiserfs_read_locked_inode: i/o failure occurred trying to find stat data of [5162 7263 0x0 SD]

Dec 8 12:33:52 ZurichMail kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Dec 8 12:33:52 ZurichMail kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=103468943, high=6, low=2805647, sector=103468943
Dec 8 12:33:52 ZurichMail kernel: ide: failed opcode was: unknown
Dec 8 12:33:52 ZurichMail kernel: end_request: I/O error, dev hda, sector 103468943
Dec 8 12:33:52 ZurichMail kernel: ReiserFS: hda10: warning: vs-2100: add_save_link:search_by_key ([-1 8 0x1001 DIRECT]) returned -2

Dec 7 12:09:16 ZurichMail -- MARK --
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 804 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 815 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 816 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 817 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 818 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 819 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 820 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 821 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 822 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 823 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 824 (java).
Dec 7 12:08:55 ZurichMail kernel: Out of Memory: Killed process 825 (java).
Dec 7 12:09:16 ZurichMail kernel: Out of Memory: Killed process 826 (java).
Dec 7 12:09:16 ZurichMail kernel: Out of Memory: Killed process 835 (java).

These are just examples, of the errors, both sets occur quite a few times at various intervals.

I blame the Out of Memory errors on tomcat and a probably bodged installation of OpenExchange. I didn't install it and plan to go out and rebuilt the machine, however the hda errors (which occur a few times in the logs for the last few days) worry me as I don't want to travel, rebuild and then find the HD dies.

Any opinions?

In my experience, it's the drive. I'd wager that some of the apps your using are trying to use swap (for going to sleep, or because they just need it) and if it's not there, that's where your memory errors may come in. Did the machine physically fall over? May have some un-seated memory, or just dirty contacts on the memory if that's the case. I'd be more worried about that drive, though. I had one doing the same thing in a small office, and sure it still ran like that for almost a year before I replaced it, but "things" would happen. DHCPd decided to stop working one morning. After reboot it was okay, but a few weeks later, it was Apache, and on and on.

As previously posted, "hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }" is definately hard drive problems. Our SLOX server recently had this (about 2 months back), but as it was an IDE-RAID system, it didn't matter so much. Replace the drive, rebuild data, and it should work.