The EMM database will not come up.

Symptoms:
The EMM server fails to startup due to a corrupt transaction log as seen in the server.log file. This may be a result of the master server's hard drive becoming corrupt or full.


Log files:
The NBDB log file located at /usr/openv/db/log/ server.log shows:

I. 06/22 08:56:15. Transaction log: /usr/openv/db/data/NBDB.log...
E. 06/22 08:56:15. *** ERROR *** Assertion failed: 201129 (9.0.2.3508)
E. 06/22 08:56:15. File is shorter than expected


Solution / Workaround:
To recover from this situation two options are available:

Option 1:

Recover the EMM database from a catalog backup.

The Catalog recovery procedure can be found in the NetBackup Troubleshooting Guide.

-  See "About recovering the NetBackup catalog" starting on page 199 of the NetBackup Troubleshooting Guide for 7.1 linked in the Related Documents Section below

-  See "About recovering the NetBackup catalog" starting on page 207 of the NetBackup Troubleshooting Guide for 7.5 linked in the Related Documents Section below

Option 2:

Remove the corrupt transaction log and recreate a new one.

Warning: There may be some transactions that could be lost.

To recreate the transaction log on UNIX/Linux platforms:

1.   mv /usr/openv/db/data/NBDB.log /usr/openv/db/data/NBDB.log_save

2. Load the db profile

         .  /usr/openv/db/vxdbms_env.sh  

Ensure you include the dot, space, then shell script or it will not work correctly.

3.   /usr/openv/db/bin/dbeng11 -f /usr/openv/db/data/NBDB.db  

This will force start that database and stop it right away.

(Use /usr/openv/db/bin/dbeng9 for Netbackup 6.x)

4.  Start the ASA service: nbdb_admin -start  or /usr/openv/db/bin/nbdbms_start_server    

This creates the new log

5.   /usr/openv/db/bin/nbdb_ping can be run to ensure the server is up.

6.  Restarting Netbackup should now work ok.

7.  It's recommended to suspend the scheduling of backups in order to verify the EMM database is intact/consistent.

To suspend scheduling run: /usr/openv/netbackup/bin/admincmd/nbpemreq -suspend_scheduling

8.  Then use the nbcc utility to verify catalog consistency. (See document 306883 in the Related Documents section below for more information regarding nbcc)

9.  Resume scheduling by running   /usr/openv/netbackup/bin/admincmd/nbpemreq -resume_scheduling

To recreate the transaction log on Windows platforms:

1.   ren Install_Path\VERITAS\NetBackupDB\data\NBDB.log Install_Path\VERITAS\NetBackupDB\data\NBDB.log_save

2.   Install_Path\VERITAS\NetBackupDB\win32\dbeng11 -f Install_Path\VERITAS\NetBackupDB\data\NBDB.db  

Note, on 64-bit platforms, the win32 directory may need to be changed to win64

This will force start that database and stop it right away.

3.  Start the ASA service: Install_Path\VERITAS\NetBackup\bin\bpup -e SQLANYs_VERITAS_NB -f

This creates the new log

4.   Install_Path\VERITAS\NetBackup\bin\nbdb_ping can be run to ensure the server is up.

5.  Restarting Netbackup should now work.

6.  It's recommended to suspend the scheduling of backups in order to verify the EMM database is intact/consistent.

To suspend scheduling run: Install_Path\VERITAS\NetBackup\bin\admincmd\nbpemreq -suspend_scheduling

7.  Then use the nbcc utility to verify catalog consistency. (See document 306883 in the Related Documents section below for more information regarding nbcc)

8.  Resume scheduling by running :  Install_Path\VERITAS\NetBackup\bin\admincmd\nbpemreq -resume_scheduling