Tape Libraries and Drives
cancel
Showing results for 
Search instead for 
Did you mean: 

Caught signal for SCSI BUS RESET!!!

Ma zhenfang
Occasional Visitor

Caught signal for SCSI BUS RESET!!!

Hi everyone

In my project, I meet a problem which puzzlee me for a long time.
The enviroment are :

Hp SureStore E Tape Library 10/180 ( 4 Ultrium Drive ,140 slots), which
connect with brocade 2800 through fibre-SCSI bridge:

10/180 : bridge: switch: device

drive 1 --> ------------
--------------> Fibra/SCSI bridge (1/2) -->| | <---------- L2000 (Backup Host)
drive 2 --> | |
|brocade 2800| <---------- N4000 (Database Host)
drive 3-----------> | |
------> Fibra/SCSI bridge (1/2) -->| | <---------- xp512
rac --> drvie 4 --> | |
------------
Omniback 3.5 with split_mirror and sap option pack.

For about every 10 days, the backup will report error as following:

11/12/01 00:23:05 BMA.4007.0 ["/src/ma/dev/devseq.c /main/r31_split/21":1926] A
.03.50 b216
SeqWrite (/dev/rmt/0m): write()=-1: {1026}
SeqWrite (/dev/rmt/2m): write()=-1: {1026}

11/12/01 00:23:12 BMA.4007.0 ["/src/ma/dev/devseq.c /main/r31_split/21":2508] A
.03.50 b216
**** Caught signal for SCSI UNIT ATTENTION!!! ****

11/12/01 00:23:12 BMA.4007.0 ["/src/ma/dev/devseq.c /main/r31_split/21":1945] A
.03.50 b216
SeqWrite: write() I/O error. Tape drive status:
SK : 06h
ASC : 28h
ASCQ: 00h
FRU : 00h
B15 : 00h
B16 : 2Ch
B17 : 6Dh
LSB : 09h

SeqWrite: write() I/O error. Tape drive status:
SK : 06h
ASC : 28h
ASCQ: 00h
FRU : 00h
B15 : 00h
B16 : 2Ch
B17 : 6Dh
LSB : 09h


Another error log :



11/11/01 01:03:22 BMA.21491.0 ["/src/ma/dev/devseq.c /main/r31_split/21":2508]
A.03.50 b216
**** Caught signal for SCSI BUS RESET!!! ****

11/11/01 01:03:22 BMA.21491.0 ["/src/ma/dev/devseq.c /main/r31_split/21":1945]
A.03.50 b216
SeqWrite: write() I/O error. Tape drive status:
SK : 06h
ASC : 29h
ASCQ: 00h
FRU : 00h
B15 : 00h
B16 : 2Ch
B17 : 6Fh
LSB : 09h

The error log came from backup host (l2000) Omniback named debug.log and the backup task failed.

I am not sure the where error come from? from hardware or software (omniback II)? How to resolve it?

Thanks for any information about it.





zhenfang
1 REPLY
Michael Tully
Honored Contributor

Re: Caught signal for SCSI BUS RESET!!!

Hi,

If your using FC you will need A5158A
version B.11.00.06 at the minimum, but
8 is later. Minimum Patches PHKL_23939,
PHKL_24027, PHKL_25210, PHSS_23996 and
PHSS_23440

Disable EMS monitoring for the tapes as
per the attached document.

Make sure that you have the kernel
parameter 'st_ats_enabled' set to 0
(very important)on both servers.

Make sure that for each drive being used
within OmniBack, that each physical drive
has a lock file of the same name.
i.e. drive 1 being used on your L class
has a lock named 'lock1' drive 1 being
used on your N class also has 'lock1' as
it's lock file.

HTH
-Michael
Anyone for a Mutiny ?