03-06-2013 03:50 AM
Hello,
When we run luxadm inq for /dev/rmt/*cbn, we can see several errors like this" Error: SCSI failure. - /dev/rmt/41cbn"
Please can you explain why this is happening and how to solve it.
Operating system is Solaris 10.
Thank you.
Solved! Go to Solution.
03-17-2013 06:22 AM
We have 24 tape drives and there are 34 cbn (they are not sequential).
You should only have 0 - 23.
Suggestion:
Clean up device names:
devfsadm -C -c tape
Next, ensure you have persistent binding in place to ensure device names stay the same across reboots.
03-06-2013 03:58 AM
Maybe best place to ask this question is on a Solaris forum?
NBU needs OS for I/O, meaning that any device problems needs troubleshooting at O/S level.
/var/adm/messages may be a good starting point to find additional info/errors.
03-06-2013 06:43 AM
Hello Marianne,
I will ask in solaris forums. The odd thing is that backups work for those drives and robtest is able to move tapes to those drives.
Thank you.
03-06-2013 08:31 AM
03-06-2013 03:16 PM
On my server (SunOS nbumaster 5.10 Generic_147441-19 i86pc i386 i86pc) I get:
Error: Invalid pathname (/dev/rmt/0ubn)
error for all but the "n" dev - I have two tape drive so only /dev/rmt/0n and /dev/rmt/1n work.
How many tape drives do you have on the server?
have a good day,
GlenG
03-07-2013 06:42 AM
Hi GlenG,
Our server is sparc SUNW,Sun-Fire-V490.
We have 24 tape drives and there are 34 cbn (they are not sequential).
The cbn's which gives that SCSI error change. Some time it may be 41cbn, other 69cbn ...
Thank you.
03-07-2013 07:07 AM
I agree with Marianne and suggest you seek help from the Solaris community.
If I had similar problem I would open an SR with Oracle.
Also, I would suggest checking the firmware levels of all the hardware involved. I had an odd problem with NBU loosing contact with the robot/lib (SL48) that in the end was firwmare in our Brocade 5000 switch.
have a good day,
GlenG
03-17-2013 06:22 AM
We have 24 tape drives and there are 34 cbn (they are not sequential).
You should only have 0 - 23.
Suggestion:
Clean up device names:
devfsadm -C -c tape
Next, ensure you have persistent binding in place to ensure device names stay the same across reboots.