October 18, 2011 at 5:42 am
Hi Everyone,
I have a situation where I have a Storage Foundation cluster for MS SQL 2008 R2. I have mistakenly installed the Instance root directory on shared storage instead of local storage. Although the cluster functions without any issue, I wish to move the instance to local storage. The SQL binaries are on local storage (c:/program files).
Would anyone be able to help me with the easiest way to carry this out. I have researched a bit and have come to the conclusion that the easiest way to do this would be to install another named instance on local storage. Detach the DB from the original shared instance and attach it to the new named instance.
Any suggestions and ideas would be much appreciated. Thank you.
S\
October 18, 2011 at 8:04 am
I have researched a bit and have come to the conclusion that the easiest way to do this would be to install another named instance on local storage. Detach the DB from the original shared instance and attach it to the new named instance.
I guess you know the answer 😉
October 18, 2011 at 8:53 am
I guess I'm a bit confused as to why you would want an instance root to be on local storage instead of the shared storage. If you have it local and the server goes down, the instance root won't automatically fail over to the other server(s) in the cluster. That Would Be Bad, IMHO.
Could you explain your reasoning so I could add it to my knowledge base?
October 18, 2011 at 9:03 am
I guess I'm a bit confused as to why you would want an instance root to be on local storage instead of the shared storage. If you have it local and the server goes down, the instance root won't automatically fail over to the other server(s) in the cluster. That Would Be Bad, IMHO.
OP has mentioned Instance Root (not the shared data i.e. database files). I don't think we need to have shared Instance Root for failover.
October 18, 2011 at 9:03 am
As far as I know, even MS recommends the instance root directory to be local and just the DB files to be shared. For instance, in my case, I wanted to use SQL Reporting Services. By default, SQL reporting service will only install where ever the instance root directory is placed.
Reporting services require a unique decryption key (used to load config) which is placed in its directory. Each machine has its own key. So in the case of fail over, when the shared volume is mounted on the second node, the reporting service would be unable to load the config since the key will not match for the second machine.
In any case, for my solution, I definitely had to install the instance root directory on local disks (clustering software requirement).
Thanks.
October 18, 2011 at 9:08 am
Thank you for clarifying that for me. I always like to hear about issues I haven't encountered yet.
October 18, 2011 at 12:46 pm
@ Dev, do you think this is the best approach?
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply