Forum Discussion

nbutechie29's avatar
10 years ago

requested media id was not found in the EMM database

Hello All,

We have been having catalog backup issues for a long time and couldnt find the solution.

The problem began when tape NB34567 got stuck in a drive and was damaged upon removal from the drive. The tape was not physical  available in the robot but it was present in the media listing. I had removed the "phantom tape", but still didnt fixed the issue. The catalog backup job is still requesting the old tape.

The  tape is not longer showing  in nbemmcmd  or in vmquery output.

Catalogue backup jobs is still trying to use the old tape and still complains as seen in error log below:

7/8/2014 10:17:00 AM - begin Catalog Backup
7/8/2014 10:17:00 AM - requesting resource NB34567

7/8/2014 10:17:00 AM - Error nbjm(pid=57085) NBU status: 95, EMM status: Media does not exist   
7/8/2014 10:17:03 AM - Error bpbackupdb(pid=6534) jmcomm_RequestMultipleResources() failed with stat = 95      
7/8/2014 10:17:03 AM - end Catalog Backup; elapsed time: 00:00:03
requested media id is not assigned to this host in the EMM database(95)
7/8/2014 10:17:03 AM - Error bpbackupdb(pid=8448) catalog backup to media id NB34567 FAILED

I would like to find out where the reference still exists for that "phantom tape", if the tape is not physically available in the robot and its not listed by nbemmcmd or vmquere?  Please help me find a solution to this mistery.

Thanks

  • HMMMM....

    Interesting.   The media has to exist somewhere in order for NBU to request it.  If the media is already reserved - it would show up in the nbrbutil output (note that nbrbutil -dumptables -f /a/path/to/put/output.file is easier to read); but it wouldn't ask for it every time.

    You stated that the media no longer exists in the vmquery output - but I see that the media id you gave is two charcters and 5 numbers long (longer than the six character/number combination that is typical).  Is this just an example media id?

    I would suggest running vmquery -a to an output file, then checking the output for just the numbers or the first two letters and a few of the numbers (just to be sure).

     

    Additionally- you might want to check out the following technote - it has a lot of good suggestions:

    http://www.symantec.com/docs/TECH46467
     

    If none of those report anything - run nbdb_unload  to a temp directory (i.e. nbdb_unload /tmp/unload).  Run grep of the resulting .dat files to see if that media id exists in any of the tables.  You can figure out which table it is by searching for the file name (such as 743.dat) in the reload.sql file.  That may help with this mystery.

    Let me know if you find the media id in any of the .dat files and what the file name is associated with that .dat file from the reload.sql file.

    Thanks,

     

    Deb

9 Replies

  • Inventory your physical tape library and robot .Also check whether other media are availble in  the catalog volume pool 

  • Are you still using cold catalog backups? Right-click on Catalog in the GUI and check Catalog backup properties. If this media-id is specified, select another unassigned tape in NetBackup pool.
  • this is a hot catalog backup and each time,it starts the job fails with the same error message.

     

  • Maybe still reserved by resource broker?

    See if this tape shows up in 'nbrbutil -dump' output.

    Other than this, all I can suggest is to restart NBU on the master to clean up anything still stuck in memory or some process.

     

  • Also, I  had performed database consistency check, stop/restarted netbackup and the issue still persists.

  • As per previous post - have you updated Inventory?
    A tape that is not in the robot cannot and will not be requested.

    Have you checked nbrbutil output? 

  • What does vmquery of the tape show ?

    Have experienced that volumes show up in the old volume manager commands even though they are not in the EMM database.

    Another place the reference could exist is on the media server that had the tape when problem occured, under .../netbackupdb/media/tpreq

  • Even after inventory same issue re occurs , check for bptm process in task mgr of media server( server used for catalog backup ). Kill that process . Note :- this should be done when no backups/ restores are running

  • HMMMM....

    Interesting.   The media has to exist somewhere in order for NBU to request it.  If the media is already reserved - it would show up in the nbrbutil output (note that nbrbutil -dumptables -f /a/path/to/put/output.file is easier to read); but it wouldn't ask for it every time.

    You stated that the media no longer exists in the vmquery output - but I see that the media id you gave is two charcters and 5 numbers long (longer than the six character/number combination that is typical).  Is this just an example media id?

    I would suggest running vmquery -a to an output file, then checking the output for just the numbers or the first two letters and a few of the numbers (just to be sure).

     

    Additionally- you might want to check out the following technote - it has a lot of good suggestions:

    http://www.symantec.com/docs/TECH46467
     

    If none of those report anything - run nbdb_unload  to a temp directory (i.e. nbdb_unload /tmp/unload).  Run grep of the resulting .dat files to see if that media id exists in any of the tables.  You can figure out which table it is by searching for the file name (such as 743.dat) in the reload.sql file.  That may help with this mystery.

    Let me know if you find the media id in any of the .dat files and what the file name is associated with that .dat file from the reload.sql file.

    Thanks,

     

    Deb