MSSQL on Microsoft Cluster [ Edited ]

[#1]

Are there any recommendations/guides on how to use Snap Creator Framework to backup a MSSQL database running on a Microsoft Cluster?

regards, Johan Gislén

[#2]

Hi Johan,

Thanks for the question

Can you please give me some more background so I can better assist you?

What problems are you trying to solve and what are your use cases? Why is SnapManager for SQL not a solution and what are you expecting out of Snap Creator?

Regards,

Keith

[#3]

Hi,

SnapManager for SQL is installed on both nodes of the cluster, so I'm using SMSQL - but would prefer to use Snap Creator to administrate/schedule the jobs instead of having to do that on each and every SQL server.

On SQL servers that are standalone, Snap Creator works perfectly, but when adding a MSSQL server that's running on a cluster, I can only get backups to work if I specify one of the nodes on the "Agent" tab. However, if that node is down, no backups will occur (which is only natural since Snap Creator can't call the agent on the down node...).

Can I add the Snap Creator Agent service as a clustered resource to the database, or are Microsoft Clusters simply not supported?

regards, Johan

[#4 · Accepted solution]

Thanks for the details

Can you use a cluster IP? MS Cluster should be using a cluster IP or hostnam so that when for example the MSSQL instance fails over the hostname and IP doesnt change. If this is possible you would just use this for Agent and things should work.

If you cant use a clustered IP for the cluster resources such as MSSQL you could add agent service as a resource. You would still need to install agent on all nodes and I think tell MCS how to start and stop it but this should work assuming the IP / name changes when a failover occurs.

Let me know?

Keith

[#5]

If I try to use the cluster IP/hostname on the "Agent" tab and test the connection, SnapCreator doesn't receive a response from the agent - so the agent is probably not listening on that IP. I will try to add the agent service as a resource to the cluster.

regards, Johan

[#6]

The agent binds to all IPs so this could be network or firewall issue. Otherwise maybe the issue could be that if cluster IP wasnt present when agent started it doest get bound but I am pretty sure this should work. Then again I cant confirm, hopefully adding agent as resource works better.

Please keep us posted

Keith

[#7]

Using the cluster IP for the agent worked perfectly - turns out there was a firewall between the SnapCreator server and the SQL server that was blocking the traffic :-)