Today I saw errors like these on the console:
Nov 14 14:05:26 mustang vmunix: Defering I/O (errno 5) for block(0x2f8640,
0x2f8640) on device 8,6146
Nov 14 14:05:26 mustang vmunix: Defering I/O (errno 5) for block(0x328b80,
0x328b80) on device 8,6146
Nov 14 14:05:26 mustang vmunix: Defering I/O (errno 5) for block(0x346870,
0x346870) on device 8,6146
Nov 14 14:05:26 mustang vmunix: Defering I/O (errno 5) for block(0x347990,
0x347990) on device 8,6146
>From searching the archives I saw that this is most likely a failing disk.
So, I tried to umount it. It is the only disk in an external tower, just a
little more info. When I issued the umount command the terminal just hung.
I tried fuser on the mount point, hung. I tried to halt the server, hung.
The screen went black and kept issuing similar error messages like I saw on
the console. I eventually had to cycle the power to get it back. From the
firmware prompt I did a sho dev and the disk was there. Ok, so I started
up.(I commented the disk out of the fstab before I tried to shutdown, yay!)
I did a fsck on the disk, it fixed some errors, then I mounted it. It has
been running without incident for a couple of hours now.
Last week I was getting event errors and from uerf I saw that they were os
event type 199, CAM SCSI, same as this time. This was on the same disk. If
it is the case that the disk is toast, why after a reboot, fsck, mount it
will work for a week before giving me more grief? Disk failures can be
intermittent, consider yourself lucky that you had so much warning?
I know that it is fairly obvious that the disk is the problem, I just want
to learn as much about the problem while I still have a disk that is
failing.
Thanks!
Trevor Osatchuk
Process Solutions Canada Ltd.
Support and Integration Analyst
(780) 452-2227 Ext. 286
trevor.osatchuk_at_pscl.com
Any man whose errors take ten years to correct is quite a man.
- J. Robert Oppenheimer, speaking of Albert Einstein
Received on Thu Nov 15 2001 - 01:32:12 NZDT