December 19, 2013 at 10:19 am
I'm seeing higher than normal latency in our transactional replication and it seems that all the log readers are sitting with wait types of Repl_Schema_Access...much more so than what we've seen in the past. I've tried to do some research on this wait type, but I've not had much luck yet. Is there anyone familiar with this? Any thoughts on what it might be pointing towards as a holdup?
Thank you,
Josh
December 20, 2013 at 2:02 am
Hi,
What is your setup of the replication
How many publisher
How many subscriber do you use
Did you use a distributor-server, or is the distributor on one of the publishers?
Do you use PUSH or PULL subscriptions?
I had high latency, because i used only one distribution database for many publications.
kind regards,
andreas
December 20, 2013 at 8:24 am
I've got 7 publishers sending to 2 different subscribers. There is a combination of push and pull subscriptions running. The distributor itself is a separate server. Currently the latency we see is only on the log reader....distributor to subscriber seems to run well.
Josh
December 20, 2013 at 8:28 am
Hi,
have you tested the latency with the "insert token" in the replication monitor?
December 20, 2013 at 8:44 am
Haven't used tokens. I've been checking numbers using sp_replcounters on the publisher and using fn_dblog as well. I've also used the DMV's on the distributor, but it seems to see log reader latency differently than the sp_replcounters does. We tend to be able to tie in latency of the log pretty accurately using fn_dblog and that matches up with the latency seconds in sp_replcounters.
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply