March 22, 2005 at 4:13 am
Hi All,
We have merge replication running to our site in India. Actually we have recently made India the Publisher and we are the subscriber.
I recently had an error on the merge agent which said that the Merge Agent was suspect (No activity for 10 mins). So I increased the Inactivity Threshold to 30 mins (not the default 10) and this seemed to correct the error.
I've been looking at the agent profiles for the merge agent.
There is one for High Volume between two servers.
HAs anyone had experience with these and would you suggest that I use this to prevent any 'suspect' issues.
Many thanks
Graeme
March 23, 2005 at 5:20 am
I've had no activity for hours and not had a "suspect" message. We run merge replication with "continuous" synchronization. We are also running merge replication with synchronization set at 20 minutes.
I've seen "suspect" problems when there have been connectivity issues and the # of retries have been met without successfully re-connecting to the subscriber. The default retries is 10 and occurs every 1 minute. Increasing this setting would help. It's under the "Run Agent" step of the Agent Properties.
Where's the inactivty threshhold?
March 23, 2005 at 5:51 am
Hi,
It's under Refresh Rate and settings...when you right click on an agent.
We have quite a bit of activity when the agent is started. It is scheduled to run for 12 hours each day every 2 mins. This is what caused the error before I reset the threshold....I think.
Also as an extra query.....
When the merge agent first starts and stops as per the schedule, the next run date (The Merge Agent Job) is reset to the next day instead of 2 mins ahead (as per the schedule). So effectivley the job does not run anymore...anyone seen this before ???
here is how it goes.
Merge Agent starts at say 8:00
Runs for about 20 minutes because of the initial high load of changes.
the job stops the Agent at 8:21...the next run date should be 8:23 but instead it is 8:00 the following day.
Very strange ??
Thing is..if I view the agent schedule, make no changes but click on OK..it resets back to what it should and runs ok for the rest of the day
Confused ?...so am I
Thanks..Graeme
March 23, 2005 at 8:28 am
That's a "trait" of the SQL Agent. It updates the next run time AFTER it finishes. When you schedule a job to run periodically and is still running at the time of the next run you get that "effect". When it start and finishes within the scheduled period you can simply refresh the job and it will correct itself
hth
* Noel
March 23, 2005 at 8:44 am
Mmmm...interesting point Noel,
We have a junior DBA who has been starting the merge agent before it's scheduled start....why?...I'm not sure after thinking about it. I think the thought was to get the merge running as soon as possible after our overnight calc jobs had finished.
Could this be the 'trait' you discussed.
What do you think?
Thanks for that Noel
Graeme
March 23, 2005 at 9:47 am
yup, everytime I've gotten an error like this, it has been because of network connectivity issues. Adjusting your threshholds and number of retries should do the trick. Good luck on this.
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply