sql server clustering: Unique Keys in a SQL Server failover cluster -- sql server clustering

Re: Unique Keys in a SQL Server failover cluster

Andrew J. Kelly
4/23/2007 12:00:00 AM

sql server clustering:

I think you have your notion of a Clustered environment with SQL Server.
Windows clustering is a hardware failover solution only. SQL Server is
simply an app that sits on top of clustering. Only one instance of SQL
Server if ever writing to the database at any one time and the dbs are held
on a shared disk with only one node having control at a time. If a node
fails over the second node writes to the same db as the first. So Identities
are the same as before.

--
Andrew J. Kelly SQL MVP

[quoted text, click to view]

Unique Keys in a SQL Server failover cluster

micklang NO[at]SPAM gmail.com
4/23/2007 12:01:15 AM

I'd like to know how SQL Server 2005 handles identity columns in a
database hosted in a clustered environment.

I have concerns that there may be a chance that concurrent requests
could result in one or more of the servers in the cluster attempting
to insert a new rows for different requests with the same primary key
and that when SQL Server attempts to merge these changes together
there could be a failure.

Is this something a developer needs to be concerned about? For
instance should a developer use GUIDs for primary keys instead of
identity columns in a clustered environment?

If anyone can shed some light on this topic or point me to information
on the web that could help me out here I would be very appreciative.

Thanks

Michael