topic Re: vmunix scsi lbolt errors in Operating System - HP-UX

vmunix scsi lbolt errors

moonchild — Wed, 09 Jul 2008 17:39:30 GMT

rp8420 connected to ds2100 running HPUX 11.11

I see a lot of SCSI errors in syslog on devices:
dev: 1f032000 and dev: 1f036000.

ioscan -fnC disk hangs
diskinfo hangs

Any recommendations?

Re: vmunix scsi lbolt errors

TTr — Wed, 09 Jul 2008 17:48:44 GMT

Lbolt errors are an indication of either a hardware error or need of a SCSI patch. These two hex numbers correcspond to disk devices. They are /dev/dsk/c3t2d0 and /dev/dsk/c3t6d0. It is not likely that both went bad at the same time so maybe there is a problem with the ds2100 or the scsi cable or the scsi interface. Is there a lot of i/o on this scsi bus?

Search for lbort in this forum you'll find more info on how to tell which disks are etc.

Re: vmunix scsi lbolt errors

Rita C Workman — Wed, 09 Jul 2008 18:01:56 GMT

Call hardware repair......you disk(s) need replacing.

Rgrds,
Rita

Re: vmunix scsi lbolt errors

Michael Steele_2 — Wed, 09 Jul 2008 18:11:36 GMT

Hi Moonchild:

Can you post the output from 'logtool'.

/usr/sbin/cstm<<-EOF
runutil logtool
rs
EOF

Re: vmunix scsi lbolt errors

moonchild — Wed, 09 Jul 2008 18:53:18 GMT

disks come and go

Re: vmunix scsi lbolt errors

Michael Steele_2 — Wed, 09 Jul 2008 19:13:36 GMT

Hi Moonchild:

Disks come and go? I don't understand this comment. What is your SAN administrator doing to make disks come and go? Rezoning? What?

And paste in the cstm report anyway. I rely on logtool heavily for diagnosing disk errors. Really need this logtool report.

Re: vmunix scsi lbolt errors

TTr — Wed, 09 Jul 2008 20:49:26 GMT

> Michael Steel: Disks come and go...
SAN administrator

In the ioscan in his attachent. They are local disks on a ds2100.

Although not improbable, it is not very likely that both disks fail intermittently. Do check the disks but also check the enclosure, scsi cable, scsi interface and scsi related patches.

I had this happen to me many times. It is mostly caused by hardware but it can happen with missing patches when there is heavy i/o on the disks.

Re: vmunix scsi lbolt errors

Michael Steele_2 — Wed, 09 Jul 2008 21:13:19 GMT

Hmmm, very odd behavior that will be reported by logtool if you can ever get the logtool output.

Re: vmunix scsi lbolt errors

George_Dodds — Thu, 10 Jul 2008 01:14:50 GMT

Try swapping the scsi cable and also re-seat the disks in the ds2100.

If you still have the issue then the disks need replaced.

Re: vmunix scsi lbolt errors

Torsten. — Thu, 10 Jul 2008 04:43:31 GMT

The command

ioscan -fnH 4/0/2/1/0/4/1

does not an actual scan of the hardware, because the "-H" is below the LBA level.

This is in fact an "ioscan -k..." - a kernel based scan.

I guess a single bad disk may hang the bus, so the system cannot access the other disks.

Any more messages (syslog, ems mail to root)?

Re: vmunix scsi lbolt errors

moonchild — Thu, 10 Jul 2008 13:27:49 GMT

I am getting a lot of EMS events on all 3 drives as per attached example.

thanks in advance

Re: vmunix scsi lbolt errors

TTr — Thu, 10 Jul 2008 14:12:38 GMT

If your ds2100 and server are under HP maintenance, let HP take control of the situation, otherwise you need to start narrowing down the field. Start by replacing the cable which is the cheapest component. Then move on to check other things as I mentioned in my previous replies. There is nothing concrete here that will tell you exactlty where the problem is and what you need to do to fix it. It is very unlikely that all 3 disks went bad. I had asked if there is heavy i/o in this bus but you did not provide an answer. If you have a spare ds2100 try swapping it. If you have another interface in the server try moving the ds2100 to it (you 'll have to do vgexport/import). Check your patch level in the 8420 and look for SCSI related patches.