'Failed to get raw report data from NetVault' error when running a report after upgrading to NVBU 10 (4036473)

'Failed to get raw report data from NetVault' error when running a report after upgrading to NVBU 10

The following messages and error occur when the Overnight Job Summary Report is launched:

1] Job Message   2014/06/06 05:00:00   4   Jobs   DCBCK03   Starting job 4 'Overnight Job Summary' (Phase 1 , Instance 809) for UID 0
2] Job Message   2014/06/06 05:00:00   4   System   DCBCK03   Start a new job component
3] Severe Error   2014/06/06 05:00:00   4   System   DCBCK03   Failed when producing report: Failed to get raw report data from NetVault
4] Error   2014/06/06 05:00:01   4   System   DCBCK03   Failed when processing report components
5] Error   2014/06/06 05:00:01   4   Jobs   DCBCK03   Job Status: Report Failed

This is strictly a NetVault 10 issue.  The problem occurs when a backup has a selection set that have nothing in them.  The sets are usually backups from NetVault 7.* before the database changes were made in NetVault 8

The trace file lists a problem with a "selection options set" for job/jobs, which used the "netapp dcst03 volumes daily /vol/oracle_generic_backups1_snapmirror - 6003670270b48e6bb4020000e41d00003a000000" selection set and had corrupted.
Example from the trace is below:

2 SCHEDDBA:08644 171 0 090304.374132 Looking up info for set defined by table 'backupselectionset'
4 DBCONNCT:08644 21 0 090304.374132 select autocreated from backupselectionset where lower(name)='selections for job netapp dcst03 volumes daily /vol/oracle_generic_backups1_snapmirror - 6003670270b48e6bb4020000e41d00003a000000'
4 DBCONNCT:08644 22 0 090304.374132 DbExecuteCommand: query returned 2
0 SCHEDDBA:08644 205 0 090304.374132 Failed to convert job to legacy format: selection options set is empty.

To resolve this issue identify the backups concerned by issuing the following queries to the database as follows.

cd <NetVault home>/pgsql/bin

./psql -p <postgres port - you can get it from pgdb.cfg from config folder> -U postgres
it will prompt for password used during installation/upgrade for postgres
\c netvault_scheduling;

Now execute following queries to find out which jobs have the problem selection sets

SELECT jobid, targetset FROM jobdescription WHERE targetset NOT IN (SELECT name FROM backuptargetset);

SELECT jobid, scheduleset FROM jobdescription WHERE scheduleset NOT IN (SELECT name FROM scheduleset);

SELECT jobid, sourceset FROM jobdescription WHERE sourceset NOT IN (SELECT name FROM deviceoptionsset);

SELECT jobid, seloptionsset FROM jobdescription WHERE seloptionsset NOT IN (SELECT name FROM backupoptionsset);

select jobid, jobtype, selectionsset from jobdescription where selectionsset not in (select name from backupselectionset) and selectionsset not in (select name from restoreselectionset);

select jobid, jobtype, advoptionsset from jobdescription where advoptionsset not in (select name from backupadvoptionsset) and advoptionsset not in (select name from restoreadvoptionsset);

select jobid, title, seloptionsset from jobdescription where seloptionsset is NULL AND jobtype=1;

One we have the list of jobs and there problem sets we can modify the problem set and save them which will overwrite the problem sets.  

Run the report to test f we have resolved the problem

If modifying the set and saving it does not work delete the job if no longer in use. 

If job is being used then create new job with alternate selection set, to replace the problem selection set.

This procedure will not work if oracle RMAN controlled backups are being run on the NetVault installation because the create spawned job with no backup option set and will cause this issue when reports run. So the report may run but as soon as the oracle RMAN controlled  backup job runs it will create the new jobs that will need deleting for the report to work again.