Tape is unloaded after every job

Disabling on-demand unloads
The NetBackup EMM service may ask the resource broker (nbrb) to unload
drives even though the media unload delay has not expired. This request is
called an on-demand unload. If allocating resources for a request is not possible
without unloading the drive, EMM may ask nbrb to unload the drive.
It may be helpful to disable on-demand unloads when a series of small related
backup jobs are scheduled (such as multiple NetBackup database agent jobs).
To disable on-demand unloads
◆ Create the following file:
On UNIX
/usr/openv/netbackup/db/config/RB_DISABLE_REAL_UNLOADS_ON_DEMAND
On Windows
install_path\Veritas\NetBackup\db\config\RB_DISABLE_REAL_UNLOADS_ON_DEMAND

Veritas NetBackup™
Backup Planning and
Performance Tuning Guide

Marek,

Thanks for the reply.  I had noticed that information, but it said it was for standalone drives, not autoloaders.  I will give it a try and repost the results.

Thanks,
Scott

http://seer.entsupport.symantec.com/docs/295336.htm

It's different parameter with standalone.

No Dice. 

I created the file as suggested and restarted all NB services.  The EMM waits specified seconds in Media unmount delay, and then tells the robot to unload the drive and put the tape back in the slot.
Next job comes right after and tells the robot to load the same tape in the drive...

Any other ideas?

while the first one is running or wating for the tape what does the "state details" say for the second job?

It says: "requesting resource vmware-cb.NBU_POLICY.MAXJOBS.SQLServers"

I forgot about this:

NBRB_MPX_GROUP_UNLOAD_DELAY
The NBRB_MPX_GROUP_UNLOAD_DELAY entry serves as a performance tuning option for the Intelligent Resource Manager. This entry indicates the number of seconds (default: 10) that the NetBackup Resource Broker (nbrb.exe) will wait for a new job to appear before unloading a tape. This setting can help avoid unnecessary reloading of tapes and applies to all backup jobs.

The following example demonstrates how to set a NetBackup configuration on a different system.
bpsetconfig -h master_server
NBRB_MPX_GROUP_UNLOAD_DELAY = 120
<ctl-Z>

It helped me once with Oracle databases.

bpsetconfig.exe responds with:
Cannot find the NetBackup configuration information

when I try that command.  I also don't find anything on that variable on the FAQs...  Do you have a seer id I can lookup that info from?  

If I understand this variable correctly, it's the same one I get from the GUI on the Media server properties under "Media request delay:"   Is that correct?

NBRB_MPX_GROUP_UNLOAD_DELAY is described in Admin Guide II for Windows and Unix

I've done it on UNIX, as new entry in bp.conf, in windows you should to use bpsetconfig or add it directly to registry
You will not find this option in GUI. It's not "Media Request Delay".

I have some more info.  I didn't notice before, but what is happening is each queued job is looking for the last media it used (which is full) and each job tries to request the same full media 186APZ and then has to request next media any....
Why doesn't NB know that 186APZ is already full and not try to load it again?  Seems ridiculous and it's going to prematurely wear out our robot making it load 2 tapes for each of our 25 databases each night...

Thanks for everyone's input so far.  Any ideas on how to make the Media server cache the availability of space on a tape?

3/25/2009 4:20:06 PM - requesting resource vmware-cb-hcart3-robot-tld-0
3/25/2009 4:20:06 PM - requesting resource vmware-cb.NBU_CLIENT.MAXJOBS.bozeman-sqlsrvr
3/25/2009 4:20:06 PM - requesting resource vmware-cb.NBU_POLICY.MAXJOBS.SQLServers
3/25/2009 4:32:34 PM - granted resource vmware-cb.NBU_CLIENT.MAXJOBS.bozeman-sqlsrvr
3/25/2009 4:32:34 PM - granted resource vmware-cb.NBU_POLICY.MAXJOBS.SQLServers
3/25/2009 4:32:34 PM - granted resource 186APZ
3/25/2009 4:32:34 PM - granted resource HP.ULTRIUM3-SCSI.000
3/25/2009 4:32:34 PM - granted resource vmware-cb-hcart3-robot-tld-0
3/25/2009 4:32:34 PM - estimated 0 kbytes needed
3/25/2009 4:32:34 PM - started process bpbrm (5392)
3/25/2009 4:32:38 PM - connecting
3/25/2009 4:32:58 PM - connected; connect time: 00:00:20
3/25/2009 4:33:06 PM - Error bptm(pid=2848) error requesting media, TpErrno = Robot operation failed    
3/25/2009 4:33:06 PM - Warning bptm(pid=2848) media id 186APZ load operation reported an error    
3/25/2009 4:33:02 PM - mounting 186APZ
3/25/2009 4:33:06 PM - current media 186APZ complete, requesting next media Any
3/25/2009 4:33:12 PM - granted resource 341APZ
3/25/2009 4:33:12 PM - granted resource HP.ULTRIUM3-SCSI.000
3/25/2009 4:33:12 PM - granted resource vmware-cb-hcart3-robot-tld-0
3/25/2009 4:33:14 PM - mounting 341APZ
3/25/2009 4:35:03 PM - mounted; mount time: 00:01:49
3/25/2009 4:35:05 PM - positioning 341APZ to file 14
3/25/2009 4:36:06 PM - positioned 341APZ; position time: 00:01:01
3/25/2009 4:36:06 PM - begin writing