I have a SQL 2005 cluster running, now the client would like to have the
dataset replicated to another SQL server offsite for further redundancy.
For the time being, the offsite server will be a stand-alone, however in the
future it could become another clustered solution.

The network connectivity between the two sites should be sufficient (as far
as we can tell at this point!).

Would SQL mirroring be a good solution in this scenario? If so, does the
target server need to be configured the same as the source (directory
structure of SQL, etc)? Any other considerations?

Thanks,
--
Shawn P Bolan
MCT, MCSE (NT4, Win2k, Win2003), MCDST, Security+, A+, Network+
sbolan@yahoo.com

You have hit on a perfect example to delineate the two technologies.

Failover Clustering is to protect High Availability due to server failure
(or system component single point of failure).

Mirroring on the other hand is to reduce the risk of site/local failure. It
is similar to log shipping, more like real-time transaction shipping.

All that is needed is a database on the target end to receive the
transactions, and a witness server to handle automatic failover.

Since this solution is to reduce risks against site issues, it is best to
host the witness at the remote site.

Also, since it is real-time transaction shipping, the off-site host will
need to remain online.

If your remote site remains offline or the number of databases per instance
that need to be mirrored is high, then you will need to look at something
like a SAN copy/replication technology.

There is a good white paper in the high availability series that discusses
these kinds of configurations: Remote Mirroring and Stretch Clustering
http://www.microsoft.com/technet/prodtechnol/sql/2000/deploy/hasog05.mspx.

Sincerely,


Anthony Thomas


--

[quoted text, click to view]