My overnight backups are failing with SCSI errors whilst verifying the tape (FBACKUP).

[Managed Services]

The server is currently running HP-UX B.11.11.

I have been having some problems for some time when my script that runs fbackup to backup and verify data was failing on the verify 'bit'.
Since the installation of two HP patches(PHKL_32089 and PHKL_34187) about 4 weeks ago, the backup script completed succesfully until the last few days.

The errors reported during the process are attached.

Do I need further OS patches or is this a hardware issue
Note HP replaced our tape device only a month or two ago.

[Peter Godron]

Hi,
lbolts are normally hardware related problems.
Check your scsi cables and terminators.
The failing device appears to be connected to Controller 4 as target 1.
Check with ioscan -fn

Please also read:
http://forums1.itrc.hp.com/service/forums/helptips.do?#33 on how to reward any useful answers given to your questions.

[Steven E. Protter]

Shalom,

lbolts come from two sources.

1) Hardware. You will eventually need to replace something, usually the disk.

2) Switching out hot swap disks. This creates a temporary lbolt that will go away next bolt.

I believe item 1 to be the case.

I recommend checking the disks with cstm or mstm or xstm.

Plus what Peter said.

SEP

[Andrew Merritt_2]

'lbolt' is the equivalent of a timestamp, it is not in itself an indication of an error. What you are seeing are SCSI errors, a bus hang followed by a timeout.


Given that we're talking about a backup to tape, I suggest you can disregard SEP's suggestions which relate to disks. It is possible that there might be a hardware problem with the tape device, so that should be checked, but as that's just been replaced it's unlikely.

Can you check the version of the OnlineDiag package you have installed? Run 'cstm' and see what version that shows. The 'dm_stape' monitor can cause problems in older versions as it used to poll the device, which can interrupt a backup. The current version does not do any polling, so should be OK, though it is possible to disable that too, since there have been suggestions that it can still cause problems.

Andrew

[Managed Services]

Thanks for the replies.
All looks ok with the ioscan (see attached).

The version number when I run 'cstm' is
Version A.45.00
Product Number B4708AA

[Managed Services]

Hi, Has anyone any further thoughts on this problem as it is causing some concern here.
Thanks for any thoughts

[Andrew Merritt_2]

A.45.00 should have a version of dm_stape that does not poll the tape device, but I would recommend that you upgrade to a currently supported version of OnlineDiags (see http://www.docs.hp.com/en/diag/stm/stm_upd.htm#table ) though I don't believe doing so will affect the problem you are currently seeing.

You can try disabling the dm_stape monitor, to see if that is causing the problem. See the first "ANSWER:" in James George's response on this thread - http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=999085

If the backup still fails, then obviously it's something else causing the failure, which you'll have to track down, and you can re-enable the monitor. If not, leave dm_stape disabled. That's going to be the default situation in future releases of the OnlineDiags, anyway.

Andrew