April 1, 2005 at 2:48 am
Hi
Just to refresh I have a transactional replication running with a Publisher
and one subscriber. The distributor is on the subscriber. I have got 4
publication and hence 4 distribution agents. all are running continuously. I
just activated my subscriptions as my synchronisation was inactive during the
retention period set. After completing the snapshots for each of the
publication, my logreader finished delivering the transactions and at the end
made the snapshots active, so that my distribution agents can continue.
Now, 2 of the distribution agent finished their job successfully and now
says no replicated transaction. (thats because there is almost no data in the
articles in these publication at the moment). My problem is the other two
agents are failing each time after its applied all the scripts for articles.
Next it just says "Launching osql to apply the scipt
<PostSnapshotCommand_0.sql>" and after about 25 minutes the agent fails
saying the "the agent is suspect. No response for the past 20 minutes" (i
have sent to inactivity check on the replication to 20 minutes)
I would like to know why doesnt it successfully apply my post snapshot
script. what are the parameters in the agent profile that i can set, I
already have the querytimeout to a high value.
Help much appreciated.
Cheers
Vikram
April 1, 2005 at 6:26 am
What does the postsnapshot do?
I have a postsnap that creates some indexes on a large table and takes some time to complete. After 20 minutes I get the no response message, however once the script has completed everything is ok and shows as such under the replication monitor.
When you get the message, check to see if the script is still executing.
April 1, 2005 at 6:50 am
thanks Nicholas,
I have got it working now. I increased the inactivity threshold for the replication monitor to about 45 minutes instead of 10 or 20 minutes. and aparently it has delivered on the transaction and the publisher and the subscriber are in sync.
However, the transaction log usage at the publisher is way to high. so would it be OK if i dump the transactions? or is it going to come down on its own, if so how and when.
Thanks very much
Vikram
April 1, 2005 at 6:58 am
I would set up regular trans log backups on the publishing server. It will help retain the db recoverability, without the log getting uncontrollable in size.
April 1, 2005 at 7:26 am
As soon as the distribution agent starts delivering all transactions the log will be clearing but the size won't be reduced.
* Noel
April 1, 2005 at 8:01 am
Thanks so much...
But i ran a checkpoint and cleared off the junk transaction from the log which allowed me to shrink the fiile as well without much problems.
I have run some test data inserts/updates on the publisher and the replication now seems to be working fine. with the logreader and the distribution agent delivering the transaction
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply