November 19, 2007 at 9:55 am
I have a 250GB database publishing articles to another server. During maintenance (re-index or purge) of the publishing db, the log reader fails. I know the trx log gets very large, but even after a trx log backup, I could not get the log reader to start back up. I eneded up having to re-create the publication. What can I do to keep replication responsive when either server, publisher or subscriber, is in a maintenace cycle? I have increased the number of times for retry to 50, with increments of 2 min. Any other suggestions?
November 29, 2007 at 11:33 am
I assume the log reader is getting blocked by a reindex job. As far as purge, if you are purging millions of records - watch your distributor - can it handle the load? Watch your distribution db grow - does it have enough hd space?
w/ a db of that size you probably have some pretty big purge jobs - you may consider removing article from trans repl, purge, then add article back in and bcp any remaining data to subscriber w/ snapshot. Transactional replication does not handle big batch jobs very well... hence the name transactional...
ChrisB MCDBA
MSSQLConsutling.com
Chris Becker bcsdata.net
November 29, 2007 at 11:38 am
Thanks for responding. I was afraid of that answer.....
December 5, 2007 at 1:26 am
Hmmm,
That does sound likely.
You might be able to use a specific profile for your logreader but that still does not get away from the large number of transactions for Trans rep to deal with. Is Trans Rep the right solution for these articles ?
Regards
Graeme
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply