Duplicate jobs from policies not running

[Peter_Cant]

I realize that there have been other posts on this subject, but the answers do not seem entirely satisfactory and we have tried all the things suggested.

We have four media servers attached to a 4-drive ADIC tape library via a fiber SAN. We have the SSO option loaded and one of the media servers is designated the primary SSO server, whilst the others are secondary. We have installed all the BE 11D hotfixes up to, and including, 11. The jobs (policy based) we have created are designed to do a backup to disk, followed immediately by a duplicate to tape. This usually succeeds on the primary SSO media server's job, but the secondary servers' jobs usually fail during the duplicate job. We have read the error codes, searched the forums and tech support pages and followed the advice therein. This has led to the following being tried:-

Deleting catalogs, recreating catalog folders
Repairing the database (plus compact, age etc etc as per instructions)
Recreating selection lists
Recreating policies
Recreating selection lists and policies at same time
reinstalling BE11D completely

Still we get "The requested source duplicate backup sets catalog record could not be found. Perhaps the media containing the source backup sets was previously deleted". How was it deleted? This is a duplicate job set to run immediately after the B2D that created the catalog.

We are also getting get lots of "ODBC Access error. Possible lost connection to database." messages. We have looked this up and found no support for this error under 11D.

Just for good measure, we are experiencing some instances of "The query for media sequence number 0 of this media family was unsuccessful. Ensure that all media in the family have been inventoried and cataloged" which is also tripping up the duplicate job.

We have tried differing B2D folder settings - our B2D file size is currently 50GB. B2D media set is set to allow no append (0 hours) and allow overwrite after 24 hours (to allow enough time for daily duplicate to tape).

We only use a B2D in order to overcome slow write to tape which drags the tape drive performance to virtual standstill, whereas B2D doesn't mind how fast the writing is. When it works, the subsequent duplicate to tape is faster, but still not as fast as we would like. A straightforward backup to tape of the bkf files created by the initial B2D is very fast, but this can neither be automated nor does it allow us to restore direct from tape.

Any ideas to overcome the problem or adopt a different strategy would be gratefully received.

Thanks

Peter

[Sandy_Sandifer]

Hi Peter,

I'll try and address these one by one:

1. For this issue: "The requested source duplicate backup sets ..." please see the following technote:
http://support.veritas.com/docs/278589
At the bottom of it you'll notice the most common scenario (due to failed backups) and the resolutions for it. With that said, this error is probably not the root cause of your issues, just part of the symptom due to whatever causes it to fail.
2. For the ODBC errors, try the following:
a. Make sure the logon credentials for the MSSQL$BKUPEXEC instance on the SSO secondary servers are set to LocalSystem.
b. Make sure the Backup Exec Job Engine service has a dependency on the
MSSQL$BKUPEXEC service on the SSO secondary servers.
c. http://support.veritas.com/docs/287529
3. For the "query for media sequence number 0..." error, try the following:
a. http://support.veritas.com/docs/276753. Even though this has been verified as fixed in version 11d, I'd still like to make sure it hasn't resurfaced.
b. Check Event Viewer's System log for any event id 7, 9, 11, 15's that correspond to when the backups are occurring. These would indicate hardware problems that we would need to address.

Please let me know if this helps resolve your issue, and if not, what errors you are seeing along with any additional steps you've taken.

Thanks.