I was asked a question in my previous post on my High Availability Options for SharePoint. I thought it would be good to share this as a seperate post.
The question was:
How can I implement both solutions? (Database Mirroring & SQL Clustering)
Clustering (active\passive) – For prevention of hardware failure.
[Mirroring] – For prevention of clustered shared storage failure.
There are a few different scenarios or ‘modes’ that you can consider if you plan on implementing both solutions at the same time.
1. High-Safety mode Session with Automatic Failover
– Principal on active node in Cluster A.
– Mirror on another clustered node on Cluster B or standalone box.
– Requires a witness server- Synchronous database mirroring
– When cluster failover takes place, the mirror becomes the principal in Cluster B.
– When the passive node in Cluster A becomes active, it becomes the new mirror.
– You can fail back and forth.
2. High-Safety Mode Session Without Automatic Failover
– Same scenario above except not automatic failover with database mirroring.
– No witness server required
– Requires manual intervention to failover database mirroring.
– Synchronous database mirroring
3. High-Performance Mode Session
– Database mirroring Principal will be on the failover clustered instance.
– If cluster fails, the mirror can be activated.
You also must consider the following:
– Assess your environment. Will it support Clustering?
– Assess your shared Storage capabilities – may consider SAN replication or virtual storage.
– These solutions WILL require additional storage, so check on your storage capacity.
– Clustering over geographical distances.. i.e. two data centres – latency, bandwidth?
– Mirroring over geographical distances – latency?
For more information: http://msdn.microsoft.com/en-us/library/ms191309(v=sql.100).aspx