March 20, 2008 at 3:18 am
Apologies for posting a question that I dare say has been posted many times before, but i'm struggling to find the answer, or guidance really.
Does transactional replication slow down inserts, updates, deletes in a published database because the log writing process may somtimes have to wait for the log reader? Or are the internals of transactional replication written such that this sort of locking doesn't happen? What about general resource contention, a disk head reading the log file for replication purposes cannot be writing to the log. I haven't seen a reference to general performance considerations to clarify this.
In a nutshell, if I wanted to change a snapshot replication (SS2K) to transactional for the convenience of re-inititialising at the article level when new articles are defined (with perhaps one or two other benefits), presumably i would have to tell the relevant users that data entry would be potentially slowed down?
Many thanks.
James Manly
March 20, 2008 at 3:39 am
Hi James,
have a look at http://msdn2.microsoft.com/en-us/library/aa902656(SQL.80).aspx#sql2k_replperf_tran4_topic05a
and look at the section "Cost of Transactional Replication at the Publisher"
Regards,
Andras
March 20, 2008 at 4:06 am
Thanks for this, very useful!
March 25, 2008 at 8:30 pm
Very helpful link and very good question.
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply