Request Timeout; About Tab error in syslog

[TD Clark]

Hi everyone. I have a real strange problem that I've been looking at for a while and can't find anything on it - hopefully one of you can shed some light. We have a Sybase server running on an L-2000 with 11.0. Over the weekend we had a problem that occurred when Sybase was running i/o intense sql routines at the same time the system was running a i/o instense process. Below are the errors we received on the Sybase side when this occurred:
00:00000:00000:2002/04/06 18:23:52.24 kernel sddone: write error on virtual disk 38 block 931681:
00:00000:00000:2002/04/06 18:23:52.24 kernel sddone: I/O error
00:00000:00013:2002/04/06 18:23:52.24 server bufwritedes: write error detected - spid=13, ppage=5028705, bvirtpg=638465889, dbid=2
01:00000:00000:2002/04/06 18:23:53.25 kernel sddone: write error on virtual disk 13 block 230128:
01:00000:00000:2002/04/06 18:23:53.25 kernel sddone: I/O error
00:00000:00009:2002/04/06 18:23:53.25 server bufwritedes: write error detected - spid=9, ppage=11494128, bvirtpg=218333936, dbid=8

Below are the errors we have in the syslog during this time period:
Apr 6 18:23:52 dnmbt2 vmunix: SCSI: Request Timeout; Abort Tag -- lbolt: 215601719, dev: 1f041300, io_id: 4d95b1a
Apr 6 18:23:53 dnmbt2 vmunix: scb->cdb: 2a 08 02 da a0 70 00 00 10 00
Apr 6 18:23:53 dnmbt2 vmunix: scb->cdb: 2a 00 01 39 6d e0 00 00 04 00
Apr 6 18:23:53 dnmbt2 vmunix: SCSI: Request Timeout; Abort Tag -- lbolt: 215601819, dev: 1f041300, io_id: 4d95bb1
Apr 6 18:23:53 dnmbt2 vmunix: SCSI: Abort Tag -- lbolt: 215601819, dev: 1f041000, io_id: 4d95b10
Apr 6 18:23:59 dnmbt2 vmunix: SCSI: Request Timeout; Abort Tag -- lbolt: 215602419, dev: 1f041000, io_id: 4d95d66
Apr 6 18:24:00 dnmbt2 vmunix: scb->cdb: 2a 00 02 e1 a5 80 00 00 60 00
Apr 6 18:24:00 dnmbt2 vmunix: scb->cdb: 2a 00 02 e1 a3 00 00 00 60 00
Apr 6 18:24:00 dnmbt2 vmunix: SCSI: Request Timeout; Abort Tag -- lbolt: 215602519, dev: 1f041000, io_id: 4d95d44
Apr 6 18:24:00 dnmbt2 vmunix: SCSI: Abort Tag -- lbolt: 215602519, dev: 1f041000, io_id: 4d95d4b
Apr 6 18:24:03 dnmbt2 vmunix: LVM: PV 2 has been returned to vg[1].
Apr 6 18:24:07 dnmbt2 vmunix: LVM: PV 0 has been returned to vg[3].

I also checked the autoraid where vg1 and 3 reside and didn't have any hardware errors.

Any thoughts???? Any help would be appreciated....Thanks

[Sridhar Bhaskarla · Accepted solution]

Hi Todd,

It is not strange to get the Request timeout errors during IO intense operations on the system.

If you are sure that there were no problems with the disks during that time, you may consider changing the Timeout value by doing a pvchange -t value.

Note down the string after 1f in dev: field in the Request timeout error. You can get the corrsponding disks by doing a
ll /dev/dsk |grep #string.

Then do a pvchange -t on the disk.

#pvchange -t 120 /dev/dsk/c#t#d#

The default timeout is 30 secs.

-Sri

[TD Clark]

Sridhar,
Sounds like you have a great avenue for me to follow. I will make this change and see if the problem persists.

Thanks for your help!