May 1, 2002 at 7:25 am
At the end of my synch on pushing the data to the subscriber, I receive a message that says "The process is running and is waiting for a response from one of the backend connections." This part takes a very long time to finish. The database that is being replicated is over 250 gig. Should this be expected?
Rob DeMotsis
Sr. SQL Server DBA
Pier 1 Imports, Inc.
Rob DeMotsis
Sr. SQL Server DBA
May 1, 2002 at 7:30 am
First what type of replication and how much data replicated at a time, and how long are we talking?
"Don't roll your eyes at me. I will tape them in place." (Teacher on Boston Public)
May 1, 2002 at 7:52 am
Merge type. About 17 gig is the size of the replicated db. The snapshot takes about and 1 hour 15 min. The merge takes about 5 to 8 hours.
Rob DeMotsis
Sr. SQL Server DBA
Pier 1 Imports, Inc.
Rob DeMotsis
Sr. SQL Server DBA
May 1, 2002 at 9:04 am
Not sure there is any need for concern as I find no listing of an issue with this. May be waiting for the replication finished reposne from the subscriber. Have you had any issues you can relate to this?
"Don't roll your eyes at me. I will tape them in place." (Teacher on Boston Public)
May 1, 2002 at 9:11 am
Not really any issues. The replication works fine after all is finished. I checked my net I/O and Disk I/O at the time of replicating and all seems fine as well. I guess it just takes that long to finish.
Thanks for the help.
Rob DeMotsis
Sr. SQL Server DBA
Pier 1 Imports, Inc.
Rob DeMotsis
Sr. SQL Server DBA
May 1, 2002 at 10:11 am
If I come across anything different I will try to pass a long.
"Don't roll your eyes at me. I will tape them in place." (Teacher on Boston Public)
May 16, 2002 at 3:47 am
Hi Rob,
I had the same problem with a much smaller database.
Could you please check the mainprocessor usage on both machines? Is it on 100% durring the whole time (or at least most of the time)?
I found the solution in setting the Clustered indexes on the table on the right column.
I a push replikation, the clustered index must be set on the rowguid row.
The next thing that should help to is, if you reindex all of the msmerge tables and the really large user tables. Use for example: DBCC dbreindex(msmerge_contents).
Use this reindex on both sides (publisher, subscriber)
Maybe this helps.
Thorsten
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply