Veritas-bu

[Veritas-bu] status 96 errors

2004-04-20 17:12:20
Subject: [Veritas-bu] status 96 errors
From: Jerald.Iverson AT aiminvestments DOT com (Iverson, Jerald)
Date: Tue, 20 Apr 2004 16:12:20 -0500
from the veritas web site:
/*
Situation:
In a multi-drive environment, if multiple jobs are submitted, and there are
not enough available tapes to fulfill mount requests from all available
drives in the Storage Unit, only the jobs that successfully obtain a drive
with a tape will continue. The other remaining jobs will subsequently fail
with a status 96, "Unable to allocate new media for backup, Storage Unit has
none available."

For example:
A customer has a 4 drive robot. One night, 27 incremental jobs kicked off
from one class. There were no available tapes in the pool that the schedule
grabs from ("INCREMENTAL"), however there were 3 available tapes in the
scratch pool. NetBackup grabbed all 3 tapes, and placed them into the
"INCREMENTAL" pool, and mounted them in the first 3 drives, and the first 3
of the 27 jobs went active. 

The remainder 24 jobs will fail with a status 96 because of the following
sequence of events: The 4th job also went active at the same time as the
first 3, however, since there was not an available tape to fulfill the mount
request on drive 4, this job failed w/ a status 96. The 5th job went active,
and also failed due to the same reason, the 6th, 7th, and so on. The 1st 3
jobs finished successfully. 

Possible Solutions to this problem:
-Ensure there are enough tapes either in the scratch pool, or the pool that
the backup request retrieves tapes from to fill all the drives before a
backup request is initiated.
-Use Multiplexing/Multiple Data Streams to get more than one job active per
tape.
-Temporarily reduce the storage unit's available drive count to match the
number of available tapes.
-Reduce the "Max Jobs Per Class" class attribute to 1.
*/

i recall reading about this back in the netbackup 3.2 and 3.4 days, and that
4.5 was supposed to fix this.  we are on 4.5mp6 and still have this
happening if our scratch pool runs out of tapes, except we have 24 drives
and 100's of clients.  it seems a shame that when 23 drives still have tapes
in them that all of the other backups will fail.  is there a bp.conf or
vm.conf entry that i have missed, such as
AVOID_UNNECCESSARY_STATUS_96_ERRORS ?

thanks in advance,
jerald iverson

-----------------------------------------
Confidentiality Note:  The information contained in this message, and any 
attachments, may contain confidential and/or privileged material.  It is 
intended solely for the person or entity to which it is addressed.  Any review, 
retransmission, dissemination, or taking of any action in reliance upon this 
information by persons or entities other than the intended recipient is 
prohibited.  If you received this in error, please contact the sender and 
delete the material from any computer.


<Prev in Thread] Current Thread [Next in Thread>
  • [Veritas-bu] status 96 errors, Iverson, Jerald <=