Windows Clustering in SQL Server - SQL Server 2008

Another way to achieve HA with SQL Server is to use Windows clustering. The main reason for discussing clustering here is to compare it to database mirroring. You can use mirroring with clustering. However, you must mirror between clusters, not within a cluster. Therefore, you could have a principal running on a two-node cluster that communicates with a mirror that runs on a separate two-node cluster.

The only thing to watch out for when using mirroring and clustering is the failover times and conflicts. For example, if you set the timeout of mirroring to 1 minute and clusters usually failover after 30 seconds, you may have clusters that failover before your mirror or vice versa. Also, you could get into some weird situations where the cluster and the mirror failover at about the same time. You definitely will want to test any configuration on which you have mirroring and clustering running together.

The main difference between mirroring and clustering is that clustering is at an instance level, while mirroring is at a database level. Mirroring does not failover other services of SQL Server or other services of the operating system. Therefore, f you use mirroring, you will need to make sure that any ancillary services that you expect to be running on the server, such as SQL Server Agent or others, must be started and running on the mirror server.

One of the benefits of mirroring as compared with clustering is that mirroring does not require special hardware, nor does it require a shared disk subsystem. Your mirrored database servers can exist in separate locations.


All rights reserved © 2018 Wisdom IT Services India Pvt. Ltd DMCA.com Protection Status

SQL Server 2008 Topics