April 15, 2002 at 11:38 pm
Dear sirs
I have one SQL Server 2000 instance with two databases (X and Y) running on 2 node (A and B) failover cluster. Both nodes are active, but SQL Server runs currently only on node A. Both databases X and Y include stored procedures that do cross database queries (eg. SELECT * FROM X.dbo.Companies, Y.dbo.Messages WHERE ...).
I want to split databases X and Y onto different SQL Server instances and put them running on separate clusted nodes. Theoretically this should double the overall performance.
I probably need to update every single query that uses both databases to look something like this:
SELECT * FROM A.X.dbo.Companies, B.Y.dbo.Messages WHERE ...
Is there a way to "map" databases so database X running on SQL Server A could see database Y running on Server B on Server A? This way I wouldn't need to update all my queries (a lot).
Has anyone done this and could give me some insight on this? What about cross database query performance?
Any help or comments is greatly appreciated. Thank you.
__
BR,
Lauri Hölttä
April 16, 2002 at 5:03 am
Sorry to get the effect you are requesting would reuire they be on the same instance of SQL. When they are not then you no long can do this directly. Now indirectly you could probably cheat by creating a local DB name Y with views the same name as the tables they reference and inside the views they do a SELECT * FROM B.Y.dbo.Messages but I don't know for sure what kind of performance hit this may have, but it should keep you from rewritting the queries you refer to.
"Don't roll your eyes at me. I will tape them in place." (Teacher on Boston Public)
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply