Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
Mike2265
Starting Member
6 Posts |
 Posted - 2004-08-31 : 05:47:00
|
| I've had some hassles when re rolling out replication recently. We're running a merge replication with 3 subscribers.Specifically, I pulled subscription off, made a few db changes, recreated the subscribers and then rolled it out again. I didn't however pull replication entirely off the publisher before rolling out again which may have created my problems which were:* the repl ranges set were not greater than the highest value in the tables. Some subscribers and publisher tables had ranges with existing values and with the identity seed set at the start of the range caused primary key errors.* the next_seed values were also too low in the distribution database.I've adjusted everything accordingly variously changing the next identity (dbcc check_ident...), resetting the constraint ranges in the actual subscriber and publisher dbs and resetting the next seed value in the distributer databases.And so my first question is really one to seek some peace of mind that what I've done is the right thing and that it isn't going to fall over somewhere unexpectedly when ranges are next rolled out ...and secondly can someone tell me what the difference between the msrepl_identity_range in the distributer and the publisher is. These tables have vastly differing values even before altering the distributer db table (i haven't changed the publisher table and if it uses the next_seed from there it will cause problems). I can't seem to find any literature on this and a brief glance at the system stored procs seems to show that publisher one isn't used????cheers and thanks,Mike |
|
|
|
|
|
msrepl_identity_range differing values