Publisher: SQL 2005 standard
Subscriber: SQL 2005 Standard
Merge replication (Pull)
DB size at Publisher is about 6GB. Snapshot file is about 400MB.
Hi, We have a total of 8 subscribers, 7 of which are all currently synchronizing fine every 30mins. 1 of the sites is on a slow connection and I cannot get the tables initialized. About 1 hour into the replication there will be a very slight (<3sec) network outage, and the replication job fails.
Error Message: The Merge Agent failed to enumerate retry information. This can occur when there is too much activity on the tempdb system database. Increase the -QueryTimeOut parameter and restart the synchronization.
I have set the Query Timeout to '0', but the situation is still the same. Is there anyway, I can make the initialization persistent to overcome bad network connection.
asked Feb 08, 2012 at 03:08 AM in Default
The slow network is really hard to deal with. There are some general guidelines, but what combination would suit you depend upon your environment.
Consider setting lower values for the following Merge Agent parameters: -DownloadGenerationsPerBatch and -UploadGenerationsPerBatch, -SrcThreads and -DestThreads. Lower values for -DownloadGenerationsPerBatch and -UploadGenerationsPerBatch can decrease throughput, but can result in a smaller number of retries over an unreliable network. These parameters are specified in the agent profile titled slow link and can also be specified on the agent command line. In addition to setting these parameters, consider using Web synchronization, which typically uses fewer network roundtrips. For more information, see [Web Synchronization for Merge Replication].
If the network connection is highly latent but is reliable, consider using the default values for -DownloadGenerationsPerBatch and -UploadGenerationsPerBatch, and specifying higher values for -LoginTimeOut and -QueryTimeOut.: http://technet.microsoft.com/en-us/library/ms151763(SQL.90).aspx
answered Feb 08, 2012 at 05:44 AM