October 31, 2011 at 12:35 am
OK, I posted two previous questions about my 2 node 2008 cluster and inability to install a SQL Server 2008 Cluster because I didn't realize shared storage is a requirement. We already popped for Win2K8 Ent. Edition too.
This is a low traffic application that needs high availability. A SAN solution for the shared storage with proper failover support is expensive. A cheaper NAS solution still poses a single point of failure.
My two nodes each have RAID 1 arrays. Would it be better to implement this application with synchronous database mirroring instead of a clustered install? If so, then there really is no need for clustering at all that I can tell. We'd have wasted the Enterprise OS license and now need two licenses for SQL Server 2008 to mirror.
By the way, in the non-clustered solution, can the SQL data files be stored on a DFS replicated volume? If so, could the other node be activated on the replicated files should the 1st node fail? I'm guessing this is a bad idea since there is no guarantee data replication will be in sync when the failure occurs.
Thanks again!
Larry
October 31, 2011 at 3:51 am
Mirroring is at the database level only unlike clustering. I guess you are already aware of this. Regarding the mirroring mode, if you are looking for automatic failover you would require 3 instances. How many databases would you be having on your instance by the way?
M&M
October 31, 2011 at 4:35 am
larrykl (10/31/2011)
I didn't realize shared storage is a requirement. We already popped for Win2K8 Ent. Edition too.
exactly why you should always research what is required before purchasing anything, even software!
larrykl (10/31/2011)
A SAN solution for the shared storage with proper failover support is expensive. A cheaper NAS solution still poses a single point of failure.
This is not entirely true, you can build just as much redundancy into NAS solutions as you can a SAN. Yes a replicated SAN is expensive, but you can also use replicated NAS devices as there are plenty of options for doing this.
The major difference is likely to be NAS over iSCSI rather than SAN fibre channel!
larrykl (10/31/2011)
My two nodes each have RAID 1 arrays. Would it be better to implement this application with synchronous database mirroring instead of a clustered install? If so, then there really is no need for clustering at all that I can tell. We'd have wasted the Enterprise OS license and now need two licenses for SQL Server 2008 to mirror.By the way, in the non-clustered solution, can the SQL data files be stored on a DFS replicated volume? If so, could the other node be activated on the replicated files should the 1st node fail? I'm guessing this is a bad idea since there is no guarantee data replication will be in sync when the failure occurs.
Thanks again!
Larry
My advice to you is to go back to the drawing board and research exactly what it is that your business requires.
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply