There are 6 replies, with the last one on 08/11/2008 at 4:50:06 PM by deandownsouth Quote: “So I’m not sure if I should move the VM’s to the same host server and go with the “Cluster in a Box” solution.” I’m not sure why you would bother doing clustering if you have both nodes of the cluster on the same physical host. The only reason to do that would to test clustering, or learn clustering. Clustering is an High Availability configuration, so putting both nodes in such a way that a single system failure from the host or host OS will bring down the entire cluster is not providing anything close to HA. All you are really getting is HA from SQL Server’s standpoint for OS things like Windows patching. You say “…in case one VM fails.” Since you are removing all hardware and driver issues, a Windows VM is pretty stable in a virtual environment, so although it is possible to have problems, those are more likely to happen as result of patching than anything else-and using internal or external snapshots is a good way to mitigate that, or better yet, since you are running in virtual, it is just as easy to create a dev environment that mirrors your production system and test the patches there. As Arek says, if High Availability is what you are striving for, running your VM in a VI 3 HA/DRS cluster is a much better solution-unless you can’t tolerate the few minutes it takes to bring the VM back online in the (rare) case that an ESX host goes belly up. Many MSCS, Veritas HA, and Oracle RAC customers are moving from those solutions to an HA/DRS solution since it takes the complexity out of clustering the OS and application in the VM.

Read the rest here:
Cluster on VM’s

Related posts

• Is Vista a sequel to Windows Me?? (0)
• Windows Storage Server - Why/Why Not? (0)
• Vista Updates VS Linux VS Mac (0)
• Mailshot problem CRM 4.0 router (0)
• Hide User Account Name when Locked (0)