January 10, 2005 at 8:35 am
I'm using SQL Server 2000, SP3, on Server 2000. Merge replication model. Publisher is local and subscriber is remote over equivalent of a 56K modem connection. I have created a new profile for the agent and modified the timeout, bcp batch size, keepmessagealiveinterval, and a couple of others. I have added the output file and verboseoutputlevel so I can see what's happening.
In my log file I can see that it applies all the scripts and starts bcping data. It starts out fine with all tables entries looking like:
Percent Complete: 50
Bulk copying data into table 'Table1'
Percent Complete: 50
select * from "dbo"."Table2" where 1 = 2
Bulk copied data into table 'Table1' (98 rows)
[1/7/2005 8:16:04 AM]Subscriber.DB: select * from "dbo"."Table2" where 1 = 2
and then it looks like it 'steps on itself' and the entries change to:
Percent Complete: 50
Bulk copying data into table 'Table3'
Percent Complete: 50
select * from "dbo"."Table4" where 1 = 2
Bulk copying data into table 'Table4'
[1/7/2005 8:16:12 AM]Subscriber.DB: select * from "dbo"."Table4" where 1 = 2
Like it timed out on bcping the data into Table3. I search the rest of the file but the Bulk copied data into table 'Table3' (150 rows) entry never appears. It does this for about 14 or 15 tables and then it errors out with The process could not deliver the snapshot to the Subscriber.
It doesn't happen consistently either, sometimes the table that 'errored' out before works the next time and tables that worked before error out on subsequent tries. Also, I generate the snapshot (which take <20 minutes) 30 minutes before I run the replication agent.
Any help here would be appreciated. I've been trying to get this to work for the last 6 weeks and am having no luck. I have a very small window of time when I can actually test due to bandwidth issues.
Angela
January 13, 2005 at 8:00 am
This was removed by the editor as SPAM
January 14, 2005 at 11:24 am
Well, I have a couple of thoughts here, but the direction I'm leaning towards is that maybe your problem is caused by connection issues. You mentioned that the snapshot push seems to bomb out in different spots on various attempts.
Next time you try pushing out, open a command shell on the server and run a ping -t to the box you're pushing the data to. Watch this window and see if you're experiencing packet loss. If so, that may be your answer.
The other thought I had is that maybe replication is bombing out because you had a table or query that's referencing an object in a foreign database that's not being replicated. I've had a snapshot push in merge replication die because of this before, but that theory is not consistent with your push bombing out in different locations.
Good luck in tracking this down.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply