Re: TSM on Solaris with a 3583 Tape Libray

First of all, Thanks Richard for your response. We have looked into your
suggestions and the SCSI cable connections are not loose, the SCSI chain
isn't too long and the here are the errors we get:

Dec  3 21:10:17 twcsds001 scsi: [ID 107833 kern.warning] WARNING:
/sbus@b,0/QLGC,isp@0,10000 (isp2):
Dec  3 21:10:17 twcsds001       Parity Error
Dec  3 21:10:17 twcsds001 scsi: [ID 107833 kern.warning] WARNING:
            /sbus@b,0/QLGC,isp@0,10000 (isp2):
Dec  3 21:10:17 twcsds001       Parity Error
Dec  3 21:10:17 twcsds001 scsi: [ID 107833 kern.warning] WARNING:
            /sbus@b,0/QLGC,isp@0,10000 (isp2):

We have replaced all the Ultra SCSI III (sbus) cards. These are the latest
            cards Sun has out. We have upgraded IBMtape, the library code
            and the tape drive microcode. The IBM CE's can exercise the
            drives without any problems but we can't open the drives with
            tapeutil. We are at loss what to try next. Any suggestions are
            greatly appreciated.

            Linda Seeba                               Phone: (314) 234-5190
            I/T Specialist - Sr                      Fax: (314) 232-4581
            IBM Global Services               E-Mail:seeba AT us.ibm DOT com



Richard Sims <rbs AT BU DOT EDU>@VM.MARIST.EDU> on 12/03/2003 09:33:19 AM

Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>

Sent by:    "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>


To:    ADSM-L AT VM.MARIST DOT EDU
cc:
Subject:    Re: TSM on Solaris with a 3583 Tape Libray



>We have TSM 5.1 server running on Solaris 8 with a 3583 IBM LTO tape
>library that is scsi attached.  Every month or so we get tape drive errors
>that hang the library so bad that we have to reboot to recover. TSM and
>tapeutil can not see the drives until after the reboot. The IBM CE's
>verified that the hardware if not the problem. They are saying the
problem
>is somewhere between the library and the O/S. ...

Linda - Ah, SCSI...

The classic cause of weird problems like this is a loose SCSI cable
connection,
a too-long SCSI chain, or faulty SCSI termination.  I'd think that the CEs
would have looked or mentioned such, but perhaps they deem it non-IBM
customer
equipment (Sun!) they don't want to get involved in.  That's something to
check.  Replace any dubious looking stuff - particulary if people are
walking on the cables or they suffer extreme bends.

In such an event, you may be able to find error messages in the Solaris
syslog or the like; and you may be able to run diags or at least drive
access tests outside of TSM.  It could also be a bad SCSI card inside the
Sun (or even the 358x: maybe that aspect didn't get examined).

   Richard Sims, BU