I am having a problem mounting cloned volumes on a second "backup" server.
These are controller based clones created using Command Scripter and the
following steps:
>From production cluster (2 GS320's running 5.1):
cmdscript -f dsk69c < <list of commands>
The list of commands consists of:
reduce DISK30000 DISK40000
add mirrorset TEMPM1 DISK30000
add mirrorset TEMPM2 DISK40000
add stripe CLONE1 TEMPM1 TEMPM2
initialize CLONE1 nodestroy
add unit D91 CLONE1 disable_access_path=all
set D91 id=291
set D91 enable_access_path=BCK010-1,BCK010-2,BCK012-1,BCK012-2
Up to this point all is working well.
On the backup server (ES40 also running 5.1):
hwmgr -scan scsi
hwmgr -view dev |grep IDENTIFIER=291
ln -s /dev/disk/dsk111c /etc/fdmns/prd_u01_domain
mount -t advfs -r prd_u01_domain#u01 /u01
At this point the mount command will sometimes "freeze up". There are no
errors returned and nothing in any of the logs to indicate the problem. A
process shows up "advfs -F 0x40000" in uninterruptable sleep. Any advfs
command run on this volume freezes up as well. There is no apparent pattern to
which domains will have this problem. u01 will fail one time then work the
next and the same for the other 13 domains. We have tried this with the Oracle
database up and in hot backup mode and down.
All domains are single lun mirror sets or striped mirror sets. LSM is not in
use.
Any ideas?
TIA
Allen
Received on Fri Oct 26 2001 - 16:21:40 NZDT