I've had this problem before. While I can't say what causes it (nor
how to make it work like it's supposed to), here's how I've resolved
it in the past
1) Wait for a command to fail against that article
2) Stop the distribution agent
3) Go to the snapshot directory at the distributor. In there, you'll
find a bunch of files, but they break down into a handful of
types: .pre (T-SQL script to run first), .sch (T-SQL schema
files), .bcp (the data), and .idx (T-SQL indexes)
4) Apply the scripts to the subscriber in the following
order: .pre, .sch, .bcp, and .idx. A note about the .bcp files:
they're in bcp-native format. You can apply them with something like
the following: bcp database.schema.table in file.bcp -n -S server -T
(understand that command before you apply it)
5) Once all of the scripts are applied at the subscriber, start the
distribution agent again.
The distribution agent should get over the rough patch and apply
commands from that point forward.
Post by Gary
I am new to SQL replication and have been assigned to replicate a few
tables to another server. I have configured it so that the distributor
is on the subscriber server and all seemed well replicating a handful
I have run into a problem when I created another publication, this
time for a table with approximately 9 million records in. It seems to
create the snapshot okay but does not seem to put the data in the new
I am not really too sure where to look for possible causes of this. I
tried this yesterday but seemed to be getting timeouts upon doing an
insert into this table and then today I am getting the "The row was
not found..." error.
Has anybody got any advice? Do I need to start from scratch on all
publications? This error is stopping my distributor from updating the
tables which did initially work now.