With that info I would start
looking at your drives.
You know that if the drives are
configured wrong what you said happed will happen.
If nb load the tape on drive 2 but
the server is looking for it in drive 3 then they think there is an issue
and will freeze the tape.
So do some checking on the drives
on that server and see if they are still configured correctly.
From: Jeff Cleverley
[mailto:jeff.cleverley AT avagotech DOT com]
Sent: Tuesday, December 08, 2009 1:56 PM
To: Judy Hinchcliffe
Cc: VERITAS-BU AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] Client freezing media.
I've thought about it but it
for the media server since it doesn't show any valid media. I don't think
that it would be anything on the master since it and the 5 other media servers
have no problems using these tapes.
One other thing I noticed the other day. The media server decided not to
backup to itself. It backed up over the network to the master. For
those backups, they ran fine and used the tapes. Some of the backups
within the same session decided to backup to the media server and all failed
with the error 96 after wiping out all my scratch tapes.
If the media database on the media server is corrupt, how can I try and clean
it up? I'm guessing if I wipe it out or move it out of the way i may have
problems doing restores.
Thanks,
Jeff
On Tue, Dec 8, 2009 at 12:18 PM, <judy_hinchcliffe AT administaff DOT com>
wrote:
My first thought is to run the database consistency check.
Greetings,
I'm going nuts trying to figure out what is happening with one of my SAN media
servers (NBU 5.1, hpux 11.11 servers). Everything was fine on Saturday
morning. Backups went as planned. Saturday night, one of the media
servers wiped out all 65 scratch tapes in the library. On the master I
have taken those media and run the following against each media:
/usr/openv/netbackup/bin/admincmd/bpmedia -unfreeze -m $TAPENAME
echo y | /usr/openv/netbackup/bin/admincmd/bpexpdate -d 0 -m $TAPENAME >
/dev/null
echo y | /usr/openv/netbackup/bin/admincmd/bpexpdate -deassignempty -m
$TAPENAME > /dev/null
/usr/openv/volmgr/bin/vmquery -deassignbyid $TAPENAME $POOLNAME 0
/usr/openv/volmgr/bin/vmdelete -m $TAPENAME
I've then done a vmquery -m $TAPENAME and nothing shows up in a database on
either the master or the media server. I then re-scan the library and the
tapes go back to the scratch pool. When I start a new backup, I find this
in the media server bptm log file:
11:10:22.859 [12387] <2> vmdb_query_scratch_bypool2: server
returned: 1 BK7028 ------ 6 BK7028L2 -------- 8 21 coda 00_021_TLD - 41 0
0 0 0 root root 2 backup scratch 1260295150 1260295814 0 0 0 0 0 0 0 - 0 0 50 0
0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - ---
11:10:22.860 [12387] <2> db_byid: search for media id BK7028
11:10:22.860 [12387] <2> db_byid: BK7028 found at offset 70
11:10:22.860 [12387] <2> select_media: media id BK7028 just obtained from
Media Manager is already in the media database, requesting a new one
11:10:22.860 [12387] <2> select_media: getting new media id for retention
level 5
It then marks the tape as frozen and wipes everything out again. Where is
the process querying the database at that the vmquery doesn't find? I've
stopped and started all processes on both the master and media server multiple
times.
One thing that does seem odd is the media server vmquery does not show any
valid media. Everything was old. I did a vmdelete of all of
them. After wiping out all the scratch tapes the media manager still
shows no media.
Any help figuring out what bit of corruption it is looking at would be greatly
appreciated.
Thanks,
Jeff
--
Jeff Cleverley
Unix Systems Administrator
4380 Ziegler Road
Fort Collins, Colorado 80525
970-288-4611
--
Jeff Cleverley
Unix Systems Administrator
4380 Ziegler Road
Fort Collins, Colorado 80525
970-288-4611