January 28, 2010 at 11:24 pm
we have seen few times after rebooting the publisher server Transactional replication gets stuck (not fail) and latency keeps on going and after we restart sql services on subscribers ,it starts catching up.
what is the concept/logic behind for this behavior?
version:sql2008 Enterprise sp1
Thanks!!
January 29, 2010 at 12:37 am
Is your publisher also your distributor? Do you know where the transactions are stuck? (ie are the log reader agents working correctly and commands getting posted into the distribution database ok)
January 29, 2010 at 10:10 am
we are using remote distributor and we verified Transactions were not able to go to distributor as well...
January 29, 2010 at 9:49 pm
Do you have entries in the msrepl_errors table in the distribution database for the corresponding times?
January 29, 2010 at 10:58 pm
Yes below are the errors in sequence time desc
Time source_name error_text
2010-01-28 18:26:50.833MSSQL_ENGCommunication link failure
2010-01-28 18:26:50.833MSSQL_REPLThe process could not execute 'sp_replcmds' on 'Publisher'.
2010-01-28 18:26:50.830MSSQL_ENGTCP Provider: An existing connection was forcibly closed by the remote host.
2010-01-28 18:26:49.553MSSQL_REPLThe process could not execute 'sp_replcmds' on 'Publisher'.
2010-01-28 18:26:49.550MSSQL_ENGCommunication link failure
2010-01-28 18:26:49.547MSSQL_ENGTCP Provider: An existing connection was forcibly closed by the remote host.
also forgot to add we are using "PULL" subscription...
January 29, 2010 at 11:05 pm
Do you have retries enabled on your replication jobs at all? Hadn't thought of that until this second. If you pull up the SQL Agent Jobs on both the distributor and subscriber do you show the jobs as errored (and have you any notifications on these jobs for when they fail)?
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply