Collation and Win2K3 Clustered Environment.
From: MikeH (MikeH_at_discussions.microsoft.com)
Date: 02/16/05
- Next message: Geoff N. Hiten: "Re: 4 Queries about DB clustering"
- Previous message: Kamal Hassan: "Re: Need Advice: Multinode Cluster w/ SQL2K on 4 Nodes and SAN"
- Next in thread: Bart Duncan [MSFT]: "RE: Collation and Win2K3 Clustered Environment."
- Reply: Bart Duncan [MSFT]: "RE: Collation and Win2K3 Clustered Environment."
- Messages sorted by: [ date ] [ thread ]
Date: Wed, 16 Feb 2005 06:39:08 -0800
This past weekend I was challenged a little in resolving what I hoped was a
simple collation issue. We are running Win2K3 Ent. Clustering Services, SQL
Server 2000 Enterprise, b.8.00.760.
On Friday, 2.11, a user migrated an application to this new clustered setup
and immediately received the infamous 446 collation error.
I spent most of Friday evening and all day Saturday reviewing support topics
and news groups. I found that there was 1 issue that could be affecting us -
the build of SQL Server for Win2K3 Ent. Clustering is subtly different than
for Win2KAS Clustering - and all of my research came back to one thing: Run
rebuildm.exe and set the collation of this instance to be the same as the
existing development/production environment. The default setup on the cluster
was different than the previous production setup, as well as different than
the current development environment.
So, I backed everyone's database up, then the master objects, shut the
instance down, executed rebuild and set the collation to what I needed it to
be.
When I brought this user's database online - the same exact error occurs.
Prior to doing this, here is what we were faced with:
Development / Previous Production: Server=Win2K, SQL=SQL Server 2000
Enterprise w/Collate=SQL_Latin1_General_CP1_CI_AS.
The new clustered SQL=SQL Server Enterprise, w/Collate=Latin1_General_CI_AI.
The database/application in question, at the procedure where we receive the
collation error, basically calls a function that creates a temp table, then
another that pulls data for a report - however, the data is collected as a
JOIN on this tempdb temp object and a series of Views the programmer
previously setup. These views in turn hit various other static tables as well.
We were thinking that because of the collation difference on the new Win2K3
setup, as well as Collation/Locale settings differences with Win2K3
Clustering, we should at a minimum change the instance to match the
SQL_Latin1, etc. collation of development. This however, has not solved the
problem.
At this juncture, the Win2K3 Clustered server has the collation
SQL_Latin1_General_CP1_CI_AI, and the development environment (as well as
their current production environment, which is 2KAS nonclustered) is
SQL_Latin1_General_CP1_CI_AS... The only difference in the two right now is
the accent sensitivity - but this should not be the issue.
If anyone else out there has any feedback, I'd be grateful for your time.
Thanks....
mhamilton"AT"nusoftsolutions"DOT"com
- Next message: Geoff N. Hiten: "Re: 4 Queries about DB clustering"
- Previous message: Kamal Hassan: "Re: Need Advice: Multinode Cluster w/ SQL2K on 4 Nodes and SAN"
- Next in thread: Bart Duncan [MSFT]: "RE: Collation and Win2K3 Clustered Environment."
- Reply: Bart Duncan [MSFT]: "RE: Collation and Win2K3 Clustered Environment."
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
