Veritas NetBackup™ Troubleshooting Guide

To recover the catalog image files using bprecover -wizard

1. If recovering the catalog to a new NetBackup installation, such as at a disaster recovery site, do the following:

   • Install NetBackup.

   • Configure the devices that are required for the recovery.

   • Add the media that are required for the recovery to the devices.

   • Create symlinks to match those in the original environment.

     See About NetBackup catalog recovery and symbolic links.

2. If the EMM server is on a different host than the master server, start the NetBackup services on that host by entering the following command:

   • On Windows:

     install_path\NetBackup\bin\bpup

   • On UNIX and Linux:

     /usr/openv/netbackup/bin/bp.start_all

3. Start the NetBackup services on the master server by entering the following command:

   • On Windows:

     install_path\NetBackup\bin\bpup

   • On UNIX and Linux:

     /usr/openv/netbackup/bin/bp.start_all

4. Start the bprecover wizard by entering the following command:

   bprecover -wizard

   The following is displayed:

   Welcome to the NetBackup Catalog Recovery Wizard!
   Please make sure the devices and media that contain catalog
   disaster recovery data are available
   Are you ready to continue?(Y/N)

5. Enter Y to continue. You are prompted to enter the full path name of the disaster recovery file, as follows:

   Please specify the full pathname to the catalog disaster recovery file:

6. Enter the fully qualified path name to the disaster recovery file for the backup that you want to restore. For example:

   /mnt/hdd2/netbackup/dr-file/Backup-Catalog_1318222845_FULL

   If the most recent catalog backup was an incremental backup, use the disaster recovery file from the incremental backup. (There is no need to first restore the full backup and then follow with the incremental backup.) Alternately, you can recover from earlier version of the catalog.

   If you specified a DR file for a full backup, a message similar to the following appears:

   vm2.example.com_1318222845
   All media resources were located
   
   Do you want to recover the entire NetBackup catalog? (Y/N)

   If you specified a DR file for an incremental backup, a message similar to the following is displayed:

   vm2.example.com_1318309224
   All media resources were located
   
   The last catalog backup in the catalog disaster recovery file is 
   an incremental.
   If no catalog backup images exist in the catalog,
   a PARTIAL catalog recovery will only restore the NetBackup catalog
   files backed up in that incremental backup.
   
   However, all of the catalog backup images up to the last full catalog
   backup are restored.  Then you can restore the remaining NetBackup 
   catalog files from the Backup, Archive, and Restore user interface.
   If catalog backup images already exist, all files that were included
   in the related set of catalog backups are restored.
   
   Do you want to recover the entire NetBackup catalog? (Y/N)

7. Enter N to continue. The following is displayed:

   A PARTIAL catalog recovery includes the images directory
   containing the dotf files and staging of the NetBackup relational 
   database (NBDB) for further processing.
   
   Do you also want to include policy data?(Y/N)

8. Enter Y or N to continue. The following is displayed:

   Do you also want to include licensing data?(Y/N)

9. Enter Y or N to continue. The following is displayed:

   Catalog recovery is in progress. Please wait...
   
   Completed successful recovery of NBDB in staging directory on 
   vm2.example.com
   
   This portion of the catalog recovery has completed.
   Because this was a PARTIAL recovery of the NetBackup catalog,
   any remaining files included in the catalog backup can be restored
   using the Backup, Archive, and Restore user interface.
   
   The image metadata that is stored in NBDB in the staging directory
   can be exported using "cat_export -staging", and, imported using 
   "cat_import".
   
   The "nbdb_unload -staging" command can be used to unload one or more
   database tables from NBDB in the staging directory.
   
   The "nbdb_restore -recover -staging" command can be used to replace
   NBDB in the data directory with the contents from the staging 
   directory.
   
   WRN - NetBackup will not run scheduled backup jobs until NetBackup 
   is restarted.
   
   For more information, please review the log file:
   /usr/openv/netbackup/logs/user_ops/root/logs/Recover1318357550.log
   

10. When the recovery job is finished, each image file is restored to the proper image directory and the configuration files are restored. If you chose to recover the policy data and licensing data, it is restored also.
11. Export the image metadata from the relational database in the staging directory, as follows:

    cat_export -all -staging -source_master source-master-server-name

    The export is required so that the image metadata can be imported into the relational database. A catalog image file recovery does not recover the relational database.

12. Import the image metadata into the relational database, as follows:

    cat_import -all -replace_destination

13. If you recovered the catalog from a disk device, you may have to fix the disk media ID references in the image headers. The image headers were recovered from the catalog backup.

    See About NetBackup catalog recovery from disk devices.

    To fix the disk media IDs in the image headers, run the following command:

    nbcatsync -backupid image_id -prune_catalog

    Replace image_id with the ID of the catalog backup. The bprecover output contains the image ID of the catalog backup being restored. Alternatively, you can find the image ID of the catalog backup by examining the DR file.

14. Before you continue, be aware of the following points:

    • If you recovered the catalog from removable media, NetBackup freezes the catalog media.

    • Before you restart NetBackup, Veritas recommends that you freeze the media that contains the backups more recent than the date of the catalog from which you recovered.

    • NetBackup does not run scheduled backup jobs until you stop and then restart NetBackup.

      You can submit backup jobs manually before you stop and restart NetBackup. However, if you do not freeze the media that contains the backups more recent than the date of the catalog from which you recovered, NetBackup may overwrite that media.

    • Because this operation is a partial recovery, you must recover the relational database portion of the catalog.

15. Clean up whitelist cache for all hosts.

16. Stop and restart NetBackup services on the master server and other hosts, as follows:

    • On Windows:

      install_path\NetBackup\bin\bpdown
      install_path\NetBackup\bin\bpup

    • On UNIX:

      /usr/openv/netbackup/bin/bp.kill_all
      /usr/openv/netbackup/bin/bp.start_all

17. After the services are restarted, run the following command:

    • If NetBackup (or host ID-based) certificates are used in your NetBackup domain, do the following:

      On a non-clustered setup:

      Windows:

      install_path\netbackup\bin\nbcertcmd -renewcertificate

      UNIX:

      /usr/openv/netbackup/bin/nbcertcmd -renewcertificate

      On a clustered setup:

      Windows:

      install_path\netbackup\bin\nbcertcmd -renewcertificate -cluster

      UNIX:

      /usr/openv/netbackup/bin/nbcertcmd -renewcertificate -cluster

    • If external CA-signed certificates are used in your NetBackup domain, do the following:

      If external CA-signed certificates are used in your NetBackup domain, do the following On a non-clustered setup

      UNIX:

      /usr/openv/netbackup/bin/nbcertcmd -enrollCertificate

      Windows:

      install_path\netbackup\bin\nbcertcmd -enrollCertificate

      On a clustered setup:

      UNIX:

      /usr/openv/netbackup/bin/nbcertcmd -enrollCertificate -cluster

      Windows:

      install_path\netbackup\bin\nbcertcmd -enrollCertificate -cluster

    • If the command runs successfully, proceed with the next step.

    • If the command fails with the exist status 5988, refer to the following topic:

      See Steps to carry out when you see exit status 5988 during catalog recovery.

      Proceed with the next step.

18. If the catalog recovery is part of a server recovery procedure, complete the remaining steps in the appropriate recovery procedure.

    This procedure can include the following tasks:

    • Importing the backups from the backup media into the catalog

    • Write protecting the media

    • Ejecting the media and setting it aside

    • Freezing the media