Friday, February 17, 2012

Concurrent

I want to use what I believe would be called concurrent snapshot replication
and by that I mean that any changes to one of the articles will immediately
(or very shortly) be pushed over to any subscribers. Here are a few
"feasibility" type of questions:
1. Do you have to have a key on all articles/tables that need to be
replicated? (I read that having a key on a clustered index is recommended,
but I want to know if a key is a hard requirement.)
2. Can you do this from one 7.0 server to another 7.0 server? (Something
in BOL led me to believe it can only be done 2000 to 7.0 for concurrent
replication but I want to veryify.)
3. Is there a definitive list of "gotchas" somewhere, i.e things that I
should check first to see if it will work? (For example, I read that
triggers on the subscriber I believe could cause issues.)
I think you mean transactional replication rather than "concurrent
snapshot". Concurrent in the context or replication means that for
transactional replication the snapshot is generated by holding brief locks
instead of locking all the tables.
To answer your questions
1) yes
2) yes. The concurrent snapshot option for SQL Server is only available with
transactional replication in SQL 2000
3) No. With defaults the publisher will set up the tables on the subscriber
they way they want them. IE the triggers will be whacked.
Hilary Cotter
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
"CLM" <CLM@.discussions.microsoft.com> wrote in message
news:106827F6-D03B-4A2F-B745-32F7FADE665E@.microsoft.com...
> I want to use what I believe would be called concurrent snapshot
replication
> and by that I mean that any changes to one of the articles will
immediately
> (or very shortly) be pushed over to any subscribers. Here are a few
> "feasibility" type of questions:
> 1. Do you have to have a key on all articles/tables that need to be
> replicated? (I read that having a key on a clustered index is
recommended,
> but I want to know if a key is a hard requirement.)
> 2. Can you do this from one 7.0 server to another 7.0 server? (Something
> in BOL led me to believe it can only be done 2000 to 7.0 for concurrent
> replication but I want to veryify.)
> 3. Is there a definitive list of "gotchas" somewhere, i.e things that I
> should check first to see if it will work? (For example, I read that
> triggers on the subscriber I believe could cause issues.)

No comments:

Post a Comment