November 29, 2014 at 4:15 pm
Looking for help -- I'm stuck.
Have a server running 14 production databases, 12 of which are replicated (transactional replication) to a second server for reporting. Distributor on same machine as publishers. All set to 'SyncWithBackup' = true, distribution set to Full recovery mode, log backups every 30 minutes, full backup nightly. This generally runs just fine.
Occasionally, the process 'hangs' indefinitely (has gone 12 hours or more before being caught) and I need to stop the backup job, stop one or more of the log reader agents, and restart everything, and it proceeds just fine. Annoying, but not fatal, and not very often.
This time, no matter what, the backup job hangs when it runs. This is true whether it is the FULL backup or just a Transaction Log backup. It hangs on the stored procedure sp_msrepl_backup_start, at the point where it is attempting to update the table 'MSrepl_backup_lsns'. When it is hung like this, all of the log reader agent jobs are also hung, blocked by this stored proc. I've tried stopping ALL of the log reader agents prior to starting the backup, but the backup process still hangs up at the same spot and never ends.
I can run the select statement in that SP that gathers the data for the databases 'manually' in a query window and it finishes in about 10 seconds. It actually seems to be hung up on the 'UPDATE' statement. When it is hung, I cannot SELECT from MSrepl_backup_lsns unless I append WITH (NOLOCK) to the statement. Nothing else I can find indicates that there is anything else locking that table. DBCC OPENTRAN shows that there is a lock on that table held by that stored proc -- but I can't see any reason why it won't update the table (17 total records) and move on.
As I said, normally this runs just fine. Totally baffled by what may be causing this at this time.
Any ideas?
Rob Schripsema
Propack, Inc.
February 14, 2015 at 8:28 am
I'm having the exact same problem, except it is occurring on SQL 2008 r2. We have only 1 replicated production database. The job will hang and then fail reporting the TEMPDB is out of space.
- Tony C
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply