finding out device

[vaman]

in hpux 11.0, I have a error in syslog,
Scsi: Request Timeout; Abort -- lbolt:833196885,dev: bc032000, io_id:3006f87
I need to find out which disk is giving problem, with information above error.

[RAC_1]

Find the disk as follows.

ll /dev/dsk/* | grep "032000" Seems that it is /dev/dsk/c3t2d0

The timeout may be related to heavily accessed disk. You can change the timeout with pvcgange -t 120 /dev/dsk/cxtxdx

Anil

[Geoff Wild]

See if this thread helps:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=219110

Lbolt errors tend to be hardware (lowlevel) SCSI errors, for instance the cable is a bit loose, is too long, isn't correctly
terminated or either of the SCSI adaptors (host adaptor in the HP, slave in the RAID
box itself) isn't working completely as it should.
Anyways, the HP (host) doesn't receive the "right answers" from the device (or too late). As the "returned" is very shortly afterwards, it just may be the response is just about on the time limit, so the cable may be a bit on the long side (the exact length supported depends on the kind of SCSI bus, single ended must be shorter then differential, and low-voltage buses have more chances of being sensitive to this then the older higher voltage ones.


Rgds...Geoff

[Sanjay_6]

Hi Vaman,

The disk in question is c3t2d0. Take a look at this thread from itrc on how to identify the offending device,

http://www2.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000063234348

the itrc doc id is A5872864.

hope this helps.

Regds

[Prashant Zanwar_4]

It might be just for some time. It can be due to the health data collection tool trying to access the data when device is busy a lot.

As explained previously, you can just do

ls -altrR /dev/rdsk/* | grep -i 032000

This will give you the devices.

And earlier also there are many discussion thread on this. Just look at those.

Thanks
Prashant

[Steven E. Protter]

Its possible to get an lbolt simply by swapping out a hot swap disk. You'll need to reboot to clear that.

Otherwise:

/dev/dsk/c3t2d0

Is either dead or on its way out. Heavy use can also trigger exceeding timeout.

Most of my experience has lbolt leading to a hardware call and eventual replacement of the disk.

SEP