HP Auto loader Problem

[Suraj_2]

Hi All

We are having a A class server running SAP. All of a sudden we started recieving the follwoing error in syslog and the HP DLT Auto Loader stopped working.

To confirm the Drive I have checked the Drive in our other RP5430 Server with Cable and there is no problem in the drive.

The ioscan o/p shows the drive as follows

autoch 0 0/0/2/0.0.0 schgr CLAIMED DEVICE HP ThinStor AutoLdr
target 5 0/0/2/0.7 tgt CLAIMED DEVICE

I tried to create the device file with -H and -e option with insf command, till device files are not creating. Even after connecting the another DLT Drive syslog starts continuously.

Is it a SCSI card problem. How can I confirm the SCSI initiator problem.There is no SCSI conflicts as there is only one device at that ID and even chekced the driver in kernel

syslog.log errors.

Jul 15 14:14:55 CRMDEV vmunix: SCSI: Unhandled interrupt -- lbolt: 7226160, dev: cb021002
Jul 15 14:14:55 CRMDEV vmunix: lbp->state: 6060
Jul 15 14:14:55 CRMDEV vmunix: lbp->offset: ffffffff
Jul 15 14:14:55 CRMDEV vmunix: lbp->uPhysScript: f87bb000
Jul 15 14:14:55 CRMDEV vmunix: From most recent interrupt:
Jul 15 14:14:55 CRMDEV vmunix: ISTAT: 0a, SIST0: c1, SIST1: 00, DSTAT: 80, DSPS: 00240c00
Jul 15 14:14:55 CRMDEV vmunix: lsp: 0000000042535300
Jul 15 14:14:55 CRMDEV vmunix: bp->b_dev: cb021002
Jul 15 14:14:55 CRMDEV vmunix: scb->io_id: 20001cd
Jul 15 14:14:55 CRMDEV vmunix: scb->cdb: 12 00 00 00 80 00
Jul 15 14:14:55 CRMDEV vmunix: lbolt_at_timeout: 0, lbolt_at_start: 0
Jul 15 14:14:55 CRMDEV vmunix: lsp->state: 5
Jul 15 14:14:55 CRMDEV vmunix: lbp->owner: 0000000000000000
Jul 15 14:14:55 CRMDEV vmunix: scratch_lsp: 0000000042535300
Jul 15 14:14:55 CRMDEV vmunix: Script dump [000000004403d000]:
Jul 15 14:14:55 CRMDEV vmunix: 09000080 00240c00 e25c0004 f87bb7f8
Jul 15 14:14:55 CRMDEV vmunix: 80080000 f87bb090 80080000 f87bb090
Jul 15 14:14:56 CRMDEV vmunix: SCSI: Resetting SCSI -- lbolt: 7226260, bus: 2
Jul 15 14:14:56 CRMDEV vmunix: SCSI: Reset detected -- lbolt: 7226260, bus: 2



with regards

Suraj

[Joseph Loo]

hi,

looking at bp->b_dev: cb021002 from syslog.log, it points to the SCSI controller with device file, /dev/rscsi/c2t1d0.

try powering off and on the autoloader and run ioscan again:

# ioscan -fn
to check status and also if device file appears. if device file do not appear, run insf again.

regards.

[Suraj_2]

Hi

Thanks for the reply.

Powering off/ on .. This I have already done.
Even ioscan and insf already done.. still no success.

Any help

Rgds
Suraj

[Joseph Loo]

hi,

correction on the device file, should be /dev/rmt/c2t1d0, check:

# ll /dev/rmt

could u also check to make sure the SCSI cable and terminator are connected properly.

regards.

[Alex Lavrov.]

What's the output of:

ioscan -kfnCtape

[Mahesh Kumar Malik]

Hi Suraj

Since another DLT drive also shows errors in syslog ,It seems that SCSI channel where drive is connected has gone to quasi stat. Advise to Power OFF the server to reset the affected SCSI channel and restart.

Regards
Mahesh

[Sudeesh]

Looks like you tried all the options possible in Online. I wouls suggest to power cycle the server.

Sudeesh

[Mel Burslan]

Right now, looks like your scsi bus either went dead or went asleep. You can try powering off your server and powering it back on (not just a simple reboot or going down to GSP prompt and issuing an RS command. We are talking about total power down). This will clear the problem if it is SCSI bus going sleeping due to an improper device addition or removal. But at the same time, I would call hardware support and have them get a new SCSI controller ready, or if this is the on-board controller, a new motherboard ready for immediate replacement. You do not want to bring the system back up and see that you have the same problem with no other alternative solutions at hand. Especially if your server is running SAP...

[Suraj_2]

Hi All

Suprisingly, after many power cycles ( thru GSP as well) and after connecting a SCSI DVD Drive and then the Auto loader, solved our problem.
Still am blank on how this happened..
Anyhow thanks for the support as usual..

Rgds
Suraj