Forum Discussion

Kevin_O_Connor's avatar
15 years ago
Solved

Restore Redirect

This is a little completcated.

We backed up an Exchange SG to VTL, the tape expired and we reimported it via our Master Server rather than the media server that backed it up.

Long story short, it was done this way because at the time we were configuring a new VTL and ATL environment but NDMP wasn't working correctly and other methods were used.

The image was then duplicated off VTL to tape.

When we try to run the restore now Netbackup picks up the correct tape but tries to run the restore via the master server.

I've changed the Media Owner to the Robot Control Host of the TLD the tape is in, configured the "General server" settings to point the master server to the RCH for restores but the job keeps trying to go through the master server.

The job output looks like this

22/11/2010 15:15:20 - begin Restore
22/11/2010 15:15:49 - 1 images required
22/11/2010 15:15:49 - media 15717S required
22/11/2010 15:16:22 - restoring image exc005_1267229051
22/11/2010 15:16:58 - requesting resource 15717S
22/11/2010 15:16:58 - awaiting resource 15717S Reason: Tape media server is not active, Media Server: app012s,
     Robot Number: 6, Robot Type: TLD, Media ID: N/A, Drive Name: N/A,
     Volume Pool: N/A, Storage Unit: N/A, Drive Scan Host: N/A
    
22/11/2010 15:17:00 - connecting
22/11/2010 15:17:35 - connected; connect time: 00:00:35

TLD6 is the correct ATL but app012s has no access to it.

Any suggestions ?

 

Thanks

Kevin

  • Ok, if you change the Media Ownership back to app012s master server first, using the process you initially ran then I suspect that will re-assign it to that host. Then run the movedb and it *should* work and correctly update the image record as well.

    That would be the cleanest way.

    Otherwise, if it's only this imported image that uses that media, edit the image record in the database and just replace app012s with app129s in each of the fragment lines.

    Hope this helps.

    Regards,

    max.

  • Hi Kevin,

    When you say under general settings you pointed the master to the robot control host for restores, I take it that you did this as the alternate restore server? (like using a FORCE_RESTORE_MEDIA_SERVER = <old_server> <new_server> in the bp.conf ?) If not, give that a try.

    Failing that you should be able to change it via bpmedia.

    # bpmedia -movedb -m <media_id> -newserver <hostname>

    This will update the image record (which is the important part, since it's the media server listed in the fragment information which it will reference as the server to mount the tape on).

    If that doesn't work, check the image file and see what media server it thinks it's on via bpimagelist. You can (depending on how brave you're feeling) edit the image record in the database and change the media server entry in the fragments to the desired media server, however the above movedb should do that as part of the process.

    Regards,

    max.

  • Thanks for the help, unfortunately I’ve already tried this.

    I used the "Change Media Owner" option within Netbackup to move it to the new media server, when I run bpmedialist -mlist -l -h app012s it doesn't appear in the list while if I run bpmedialist -mlist -l -h app129s it does.

     

    I know I'm missing something but I just can't figure out what.

  • Kevin,

    If you do a bpimagelist -backupid <backupid> on the image you want to restore, what entries do you have for the media server in the fragment information?

    i.e.

    FRAG 1 1 5231168 0 2 20 2 <tapeid> <mediaserver_id>

    Is it showing the correct media server?

    I'm not sure what process the 'Change Media Owner' goes through and if it actually modifies image records in the catalog or not.

    One thing that nags at the back of my mind, when you import a tape I think it freezes it, again I'm not sure what implications that *may* have with changing media ownership. It shouldn't make any difference, but I've learnt not to take these things for granted :)

    Regards,

    max.

  • Thanks Max.

     

    The Frag information actually references app012s instead of app129s.

    I tried running "bpmedia -movedb -m 15717S -newserver app129s" but got the following error "requested media id is not assigned to this host in the EMM database"

  • Ok, if you change the Media Ownership back to app012s master server first, using the process you initially ran then I suspect that will re-assign it to that host. Then run the movedb and it *should* work and correctly update the image record as well.

    That would be the cleanest way.

    Otherwise, if it's only this imported image that uses that media, edit the image record in the database and just replace app012s with app129s in each of the fragment lines.

    Hope this helps.

    Regards,

    max.

  • You can change the media server name in the Image database using :

    bpimage -newserver <name>  -id <media-id>

  • I finally managed to get it working, thanks for all of the help.

    In the end I changed the media owner within Netbackup back to app012s and then used the command line to change the media owner to app129s.

     

    The restore is working away.