Forum Discussion

CarloMG's avatar
CarloMG
Level 1
6 years ago

Exchange GRT backup duplication failure

We have been running granular restore technology (GRT) backups of an Exchange 2016 cluster for many months now.  The backups are to a disk on a Veritas appliance, and once a month our vaulting rules duplicate a one set of backups (for 6 Exchange mailbox databases) from disk to tape for long-term-retention.  We have the option to "enable message-level cataloging when duplicating Exchange images" enabled, so these duplication jobs take a long time (many hours per database) and generate a heavy workload on our Exchange servers and we accept that.

This pas weekend the duplication jobs generated by vaulting repeatedly failed with "Error bpduplicate (pid=xxxxxx) db_IMAGE() failed: database system error (220)" and then "no images were successfully processed (191).

I am aware of the articles saying this is usually due to a timeout in BPDBM which defaults to 5 minutes, but I don't believe that is issue here, for two reasons:

1) We do not have IGNORE_BPCD_TIMEOUT  set in bp.conf but these duplication jobs have alwasy worked for us here

2) These failures are occuring 2 or 3 seconds after the start of each duplication attempt.

Any ideas?

 

8
Linux
  • Lowell_Palecek's avatar
    Lowell_Palecek
    Level 6
    5 years ago

    The log path you want to investigate is this:

    bpduplicate in the admin log
    bpdbm on the master server
    ncflbc on the Exchange server

    The nbfsd log on the media server may be helpful, but it's needed only for particular narrow circumstances. When it is, you may also need bptm logs.