Forum Discussion

Denda's avatar
Denda
Level 5
14 years ago

Unable to import 6.5.4 tapes into 7.0.1

Master/Media Server environment - VM's W2K8 / 7.0.1

Tape Media Server environment - Physical W2K3 / 7.0.1

I need to figure out why I cannot import my 6.5.4 backup tapes to my 7.0.1 environment. 

When I created the 7.0.1 environment, I created it from scratch, new hardware, clean install of 7.0.1.  I thought it would be a simple process to just import any tapes from the old 6.5.4 environment to the new, however I am coming against some challenges. 

Was under the gun a couple weeks ago and had to boot up our old 6.5.4 master server to restore some critical data which messed up our current environment and I just got that fixed, now another request has come in for data from last year... So.. now I am forced to figure this issue out once and for all.

I can successfully perform Phase 1 with the tapes (backup spans over 3 tapes).

Import phase 1 started 05/17/11 09:44:03
09:44:03 INF - Create DB information for media id NPZ311.
09:44:03 INF - Initiation of bptm process to phase 1 import media id NPZ311 was successful.
09:44:04 INF - Waiting for mount of media id NPZ311 on server hcinetbkup01 for reading.

09:44:32 INF - Creation of database information for import of media id NPZ311 complete.

Import phase 1 started 05/17/11 09:47:36
09:47:36 INF - Create DB information for media id NPS790.
09:47:36 INF - Initiation of bptm process to phase 1 import media id NPS790 was successful.
09:47:37 INF - Waiting for mount of media id NPS790 on server hcinetbkup01 for reading.

09:48:07 INF - Found existing image information for client hcisqlprod05, policy Tape_Daily_Backup_DMBI, and schedule Daily_Backup, done on 05/15/2011 04:00:05.
09:48:07 WRN - Cannot import client hcisqlprod05, policy Tape_Daily_Backup_DMBI, and schedule Daily_Backup, copy 1 fragment 1 already exists. Skipping this image.

09:48:07 INF - Creation of database information for import of media id NPS790 complete.

Import phase 1 started 05/17/11 09:48:41
09:48:41 INF - Create DB information for media id NPK779.
09:48:41 INF - Initiation of bptm process to phase 1 import media id NPK779 was successful.
09:48:41 INF - Waiting for mount of media id NPK779 on server hcinetbkup01 for reading.

09:49:11 INF - Creation of database information for import of media id NPK779 complete. 

Once I try to implement Phase 2, it states there are no images ready for Phase 2.  I have tried via the GUI as well as command line.

Differences?  Different media and master server than the original backup.

I can boot up the old environment again (6.5.4) and the restore works fine, so I know the tapes are OK.  Any ideas?

Thanks All!

20 Replies

  • I have Symantec support working on this now, I will update the post when/IF we figure it out.

  • If you have done the phase 1 via the GUI, please check the detailed output in the Results tab. If you copy the text and post it here we might be able to see why phase 2 is giving you grief.
  • That info is in the orginal post (very first one).  It tells me that Phase 1 is done successfully but I am not seeing the tmp file that support is telling me should be there at Phase 1 so I think this 'success' message is incorrect.

  • can you give result of this commands :

    bpimmedia -d 01/01/1970 -mediaid NPZ311

    bpimmedia -d 01/01/1970 -mediaid NPS790

    bpimmedia -d 01/01/1970 -mediaid NPK779

    bpimport -PM -v -s 01/01/1970

     

     

    Now you can test this methode (CLI)

    Expired all images associated with these medias on NBU DB IMAGE

    bpexpdate -m NPZ311 -d 0 -force

    bpexpdate -m NPS790 -d 0 -force

    bpexpdate -m NPK779 -d 0 -force

     

    Manually import PHASE 1

    bpimport -creat_db_info -id NPZ311 -v

    bpimport -creat_db_info -id NPS790 -v

    bpimport -creat_db_info -id NPS779 -v

     

    Find all backupid available for Import phase 2

    bpimport -PM -v -s 01/01/1970

     

    Manually import PHASE 2 backupid

    bpimport -v -s 01/01/1970 -backpuid <BID_find>

     

     

    Stéphane COLIN

    C-Storage

  • Unfortunately, I did a stupid thing and now I need to go grab a different backup.  As soon as I expired the tape earlier today, a backup job grabbed the tape before I could get it into another group and started the primary tape. I did not check to make sure my tape backups were done prior to messing with this.  I will post later... ugh...

  • ok,

    By default NetBackup does not overwrite a media with TAR write by another master.

    Have you disabled this protection (ALLOW_MEDIA_OVERWRITE = TAR on bp.conf) ?

    I ask you this question to be sure that we do not have a new problem with this protection on NetBackup.

     

     

    Stéphane COLIN

    C-Storage

  • Really?  There is nothing in my Media Host Override, if that is what you are referring to... No BP.conf that I know of in windows.  The tape is currently being written to, are you saying it just may be adding data to the tape and not overwriting?

    How would netbackup know that it was written by another master (completely different environment) since it never was fully imported, the tape was just added to the inventory which looks as though it is scratch?

  • I don't quite agree with Stephane's last post - a tape in the Scratch pool is available to be overwritten. You have seen the proof.

    I would like to see increased logging for future imports. It does not seems as if ANYTHING happened for tape  NPZ311:

    09:44:04 INF - Waiting for mount of media id NPZ311 on server hcinetbkup01 for reading.
    09:44:32 INF - Creation of database information for import of media id NPZ311 complete.

    Same for NPK779:
    09:48:41 INF - Waiting for mount of media id NPK779 on server hcinetbkup01 for reading.
    09:49:11 INF - Creation of database information for import of media id NPK779 complete.

    Phase 1 import for tape NPS790 found existing database info (imported previously??)

     

    For future imports - PLEASE write protect your restore tapes. Also, do not put the tapes in the robot and use normal inventory to add tape to EMM - if your default is Scratch pool for new tapes, you do NOT want your restore tapes in Scratch, even if you change the pool. As you have found, expiring them will return them to Scratch.
    Rather create a separate RESTORE pool and use 'vmadd' to add the tape and put it in RESTORE pool. Then write-protect the tape and put it in the robot. Robot inventory will move it from standalone to robot.

     

    Another idea, since you seem to be needing regular restores: use Recovery Without Import method to copy images from old master to new master.

    Original 3-page white paper: http://eval.symantec.com/mktginfo/products/White_P...

    Updated detailed doc:
    http://eval.symantec.com/mktginfo/enterprise/white...

  • Correct, if the tape has been written to using for example os tar outside of NBU.

    If there is a NBU header on the tape, NBU will overwrite it if there are no valid images in the catalog for the tape - in this case, it just believed it was scratch.

    There is no issue with NBU, to put your mind at rest.

  • OK.. so here is the rest of the story..

    grabbed another set of backup tapes, same result, could not implement Phase 2..

    grabbed a third set of backup tapes, everything is importing in correctly..

    What's the difference?  The original set of backup tapes + the second set were compressed... The third set is not.  I think something is wacky with compression on my old system.  Has anyone else ever experienced something like this?

    Gives me one more variable to test DR on...