Unsolved
This post is more than 5 years old
26 Posts
0
1212
Networker backup job - pool spanning jukeboxes
Hello
I do not have means to make a simulation to confirm / clear my suspicion, so I'll ask here.
It seems to me that DB2 module backup does not know how to take a tape from a pool for a running job if that tape resides in different jukebox.
At the same time, no media alerts are generated and from few cases that I saw it happening, job is contend waiting for next available volume IN THE SAME jukebox, while there are free volumes in another.
Or is that behavior common to all jobs? I seem to notice tendency for any group/job to focus on one jukebox until completion..
(Server is Win2003R2, client AIX, jukeboxes are within same VTL)
mridul_singh
57 Posts
0
January 23rd, 2012 20:00
Is there any restrcition on device selection in the media pool?
CroWiz
26 Posts
0
January 24th, 2012 07:00
There is indeed a restriction.
But it covers ALL drives in first AND second jukebox!
mridul_singh
57 Posts
0
January 24th, 2012 20:00
There has to be some restriction on the drives selection. Try a backup by removing all the device restriction from the media pool.
Have a look at the DB2 configuration file for any such specifications. As you have mentioned that the backup is not taking the tape from another jukebox, though volumes are available, are the volumes labelled in the required media pool. If not enable "auto media management", "recycle to other pool" and "recycle from other pools" attributes in jukebox and media pool properties respectively and try the backup.
The other way around is to check if the file system backup works. This will help you to isolate the problem.
CroWiz
26 Posts
0
January 25th, 2012 01:00
Backup will readily take volumes from the SAME jukebox-spanning pool from either jukebox.
It just wont jump to use devices/volumes in another jukebox, once it starts ..
DB2 configuration defines only the pool, not the devices.
coganb
736 Posts
0
January 25th, 2012 03:00
Are these two jukeboxes connected to the same storage node? NetWorker will use the Storage Node affinity as defined in the client resource to decide which storage node to send the data to.
-Bobby
CroWiz
26 Posts
0
January 25th, 2012 06:00
Same storage node. Same physical VTL. Same type of media (LTO3).
coganb
736 Posts
0
January 26th, 2012 01:00
I'm not sure what's going on here Zlatko. I agree with Mrdiul, you should experiment with different parameters of the situation to try to isolate what could be causing this, starting with seeing if it behaves the same way for a filesystem backup.
-Bobby
CroWiz
26 Posts
0
January 26th, 2012 06:00
I agree that the first step should be to determine whether this is limited to DB2 backup or is it valid for all (filesystem) backups. However I do not have means to play with it at the moment.
ble1
14.3K Posts
0
January 27th, 2012 05:00
Do you see this pattern only with DB2? Can you dump here configuration for pool, jukebox and client affected?