Microsoft Cluster Snapshot Sync Question

[quantum10]

We have a SQL cluster with the nimble as the shared storage iscsi target. The NCM is installed.

When setting up the volume protection I can only enter one server as the sync server, but the drives are connected to all servers in the cluster and in the case of a failover from the preferred node the proper server may not hold the drives currently.

Is there a way to set this up so we can snapshot with out errors no matter which node the work load is on?

[david_tan2]

There are two ways I can think of....

1. Commvault can do it all for you with intellisnap

2. Depending on your network setup, create a new ip resource on the cluster that nimble can talk to, that sits in the same cluster group as sql. Configure nimble to use this ip for its vss snapshots.

[pfrisoli27]

Hi Christopher,

I hope you are well. I believe what you are attempting is to Cluster SQL using a Nimble volume. My thought is you will need to setup a volume that will serve as the quorum disk and on that you would need to enable "multiple initiator access" when you build the volume, and setup an initiator group containing both those server IQN's. This volume would serve as the quorum for the SQL cluster. You would also have a volume defined and available to both servers to service the database and log files, of which, you could then define snapshots with VSS enablement to ensure quiescence of the data. I would suggest a quick call with Support to check my thoughts on setup and configuration for this build. Hope this helps! Paul

PS: I saw the mention of Commvault within this thread and that is another way to service the SQL Cluster.

[quantum10]

We don't have Commvault.

I tried using a clustered IP with no luck.

I'm thinking of setting up a test cluster and trying to use the generic service role to create a cluster version of the Nimble VSS services and give them an IP to test, but I don't know when I'll have time. Any one else tried this?

[david_tan2]

The IP address will need to sit in the same subnet as the storage nics configured on each node (unless you can use the management subnet). Configure that IP in nimble as the app IP. Nimble will then talk to the node that owns the IP resource and do the vss snap. I'm pretty sure this will work. I can test it tomorrow.

[quantum10]

I added an IP on the storage subnet as a client access point and still received the same error. Let me know if you get it working.

[pfrisoli27]

Hi Chris,

I know this has gotten me in the past. Have you checked to ensure that a firewall is not blocking connectivity? Try disabling the Windows firewall if you have it enabled. Thanks P...

[david_tan2 · Accepted solution]

Hi Chris,

Confirmed this works. Here's what I did (Note: windows firewall is disabled - this might simply be your problem as Paul mentioned)

1. First, in cluster manager find the storage subnet under Networks and ensure that "Allow cluster network communication on this network" and "Allow clients to connect through this network" are enabled.

2. Next create a new IP resource from the storage subnet in the same cluster group as SQL. Set the storage subnet from he drop down list and enter an available IP. 3. Bring the resource online.

4. Log on to the array gui and create a new volume collection if it doesn't exist. Add the required volume(s) and set the application server to the IP address used in 2. Enable VSS integration of course.

5. Run the validate on the volume collection ( this should succeed)

6. Snap away!

[quantum10]

This worked. I added a client access point originally which didn't work, but the IP resource did. Thank you.