August 7, 2009 at 6:08 pm
I am facing an problem with my replication. after creating transactional replication and reinitialising the snapshots i had an error that the concurrent snapshot cannot be applied by the logreader agent. so i went ahead and dropped the publication which was the source of contention and recreated it. i even changed the profile of distributuion agent to large timeout changing the query timeout to 3000000. now after that everything went smooth until i had a table(each publication as one table as article) which had massive records and it took around 2 hours for the distribution agent to load it into the subscriber. afterwards i was getting waiting for backend connection (it was applying indexes) and after 10 minutes it showed agent as suspect(I know nothing to worry). after 3 hours it started to apply snapshot of the same table again. this has been the process goin on.
any ideas why? because there are few other publications reinitialised after this table they are waiting for the distribution agent to get free. one more thing i have checked for locks at the subscriber(sp_who2) and my spid hich was BCP the table was blocking few other spid's and once it applies BCP it is changing its status to rollover(@@transcount>0 rollback tran).
any inputs would be appreciated.
August 27, 2009 at 2:07 am
Hey..
How big is the table that is taking a long time ?
What is your connection like to your subscribers ??
Have you thought of applying the data via BCP (a data file can be compressed to quite small) and then applying nosynch replication ?
Regards
Graeme
August 27, 2009 at 2:47 pm
Right...
I would backup and restore on subsriber and set up the replication with nosync( subsriber is already having data)...
my one cent.
Thanx.
Vinay
Thanx.
Vinay
http://rdbmsexperts.com/Blogs/
http://vinay-thakur.spaces.live.com/
http://twitter.com/ThakurVinay
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply