Message in syslog

[Georgi Georgiev]

I have AutoRaid HP C5447A and HP-UX 10.20
Message in syslog :



SCSI: Request Timeout -- lbolt: 789026821, dev: 1f000500
lbp->state: 4020
lbp->offset: ffffffff
lbp->uPhysScript: 780000
From most recent interrupt:
ISTAT: 22, SIST0: 04, SIST1: 00, DSTAT: 80, DSPS: 00000006
NCR chip register access history (most recent last): 168406586 accesses
21503, ISTAT<-20
4070, ISTAT<-20
7375, ISTAT<-20
2862, ISTAT<-20
17338, ISTAT<-20
31837637, ISTAT<-20
117, ISTAT: 22
7, SIST0: 04
5, SIST1: 00
3, DSTAT: 80
3, DSPS: 00000006
2, SCRATCHA: ff000866
6, DSP: 00780058
3, SCRATCHA1<-00
2, CTEST3<-04
1, STEST3<-82
lsp: 3fd6480
bp->b_dev: 1f000500
scb->io_id: d9f296
scb->cdb: 2a 00 01 d9 25 00 00 00 80 00
lbolt_at_timeout: 789023721, lbolt_at_start: 789023721
lsp->state: 10d
lbp->owner: 3fd6480
scratch_lsp: 0
Pre-DSP script dump [3ac4030]:
78346600 0000000a 78350800 00000000
0e000004 007803c0 80000000 00000000
Script dump [3ac4050]:
9f0b0000 00000006 0a000000 007803c8
721a0000 00000000 c0000004 0078035c

SCSI: Abort Tag -- lbolt: 789026821, dev: 1f000500, io_id: d9f296
LVM: VG 5 : PV 0 (device 0x1f000500) is POWERFAILED
LVM: Recovered Path (device 0x1f000500) to PV 0 in VG 5.
LVM: Restored PV 0 to VG 5.

[Chris Wilshaw]

These can be indications of a disk failure on the system.

use ioscan -fnCdisk to display the devices, and see if nay show up as NO H/W - this would indicate a total disk failure.

It is also possible that these are simply timeouts (the disk hasn't responded to a request in the default time). If this is the case, the timout period can be extended using

pvchange -t /dev/dsk/

The disk device that is potentially the problem can be identified from the line

SCSI: Request Timeout -- lbolt: 789026821, dev: 1f000500

In this case, 1f000500 probably translates to /dev/dsk/c0t5d0

if you do an

ll | grep 05

in /dev/dsk/, you should see something like

brw-r----- 1 bin sys 31 0x005000 Mar 5 14:57 c0t5d0

[Vincent Farrugia]

Hello,

Just to add.

Remember that c0t5d0 is NOT referring to a single disk, but to all disks put together, in case of the AutoRAID. c0t5d0 is a LUN distributed across all the disks evenly.

So it is a bit remote that an lbolt actually means a hard disk failure in this case.

I would add the timeout as shown and see what happens.

HTH,
Vince

[Stefan Farrelly]

Its either;

1. One of the disks in the autoraid physically spun down (which means its slowly dying) then spun up again (so now its ok) or
2. Youve overloaded your autoraid so much that its timing out on scsi connections (as per other replies).

Either way its not too much to worry about unless your autoraid has no protection against a disk failure. If it happens again then I would call HP to arrange an investigation or a replacement drive.

[Rainer von Bongartz]

This looks like a timeout of a SCSI request. This is not necessarily a hardware problem.

Do you have heavy load on the bus??

Check for the latest SCSI patches on your system.

Regards
Rainer

[Yogeeraj_1]

Hello,

I had a similar problem once!
see:
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x52e89276484dd611abdb0090277a778c,00.html

I did not encounter it again since then..

The 'lbolt' errors are generally considered to be a timeout issue from the server to the autoraid.

Hope this helps!
Best Regards
Yogeeraj

[Rita C Workman]

Well you can get these errors either because it was a true timeout situation... or...because your disk or cable or even controller is having trouble.

I ran into a problem with some bad controllers (HP known issue on fiber contoller..) and started seeing messages like this sporatically on some boxes...I replaced every last one and never saw the message again.

So you might want to add cable and controller on your things to check.

Just a thought,
Rita

[Keely Jackson]

Hi

Are your raids attached via Fibre Channel? If so check your patch level. We had similar Powerfailed errors and they were corrected by installing PHSS_23581.

Cheers
Keely