Solved

Mssql always on Failover Cluster Instances configuration

  • 14 January 2022
  • 11 replies
  • 2100 views

Userlevel 1
Badge +5

Hello all,

 

Some doubt makes me ask some questions to community before acting please.

 

I’ve never set such mssql backup.

 

As i’m not a mssql expert, i’ll try to explain and understand as well as possible, so please be kind!

 

I’ve a mssql failover cluster composed with 2 nodes.

Only one node is active at the same time, and has the network share hosting the database mapped, not double mapped on passive during this.

There is no special availability group (that’s mostly my point)

 

I’ve experience in backuping an exchange dag, so i would suppose same backup configuration philosophy:

  • Install mssql agent on both nodes, no config
  • Create a virtual client “mssql ag client”

But this is my doubt, when i select this, i’m asked for sql clients (ok, nodes), sql instances (ok, instances), but asked for availability group.

I don’t have done now, i try to understand this before, so maybe i can just let availability group list empty?

 

Is someone able to tell me if i’m in the right way? Or else?

 

Thanks in advance Commvault people!

 

Regards

 

 

icon

Best answer by f-red 20 January 2022, 19:35

View original

11 replies

Userlevel 4
Badge +7

Hello,

It sounds like you may be confusing SQL Clustering + Availability Groups (Which is a form of clustering - but at a database level, not instance level).

You have said you have a 2 node failover cluster - this is a cluster of the entire SQL instance right? (If you open failover cluster manager, you see the SQL instance as a clustered app).

Check this out (Watch the entire video but the link is set to show you specifically the clustering config): 

Assumption is made here that you’ve already installed Commvault on both SQL nodes + the SQL agent (And don’t add the clustered SQL instances on these physical nodes).

Userlevel 7
Badge +15

Hi @f-red 

What you’re seeing is the configuration expectations for SQL Always On Availability Groups for SQL Server, pretty similar in principle to Exchange DAG.

Always On Availability Groups for SQL Server

There is an expectation for a pseudo client representing the AG group and you would add the individual SQL server nodes to the group.

However, if you aren’t using the AG structure for SQL, then you will need to consider these as 2 separate independent clients.

Thanks,

Stuart

Userlevel 6
Badge +14

@f-red Adding here, if you are running 11.21 or later and have not disabled it the software will auto discover and create the AG client once the SQL agent is installed on the nodes.

https://documentation.commvault.com/11.21/expert/129021_automatic_discovery_of_sql_availability_groups.html

Userlevel 1
Badge +5

Hello,

 

Thanks for your replies.

 

We finally did it, i’m writing here what was my solution.

 

At first, i didn’t knew mssql failover clusters:

  • Mssql cluster groups, with its vip mssql service
  • Failover by switching db network share mapping

 

I didn’t knew yet commvault “windows cluster” clients.

 

So, what i’ve done:

  • Installed fs and sql agents on the 2 nodes:
    • Configured system fs backup excluding quorum and network share pathes, with system state = enables os disaster restore
    • Nothing configured in local sql agents
  • Created new “windows cluster” client:
    • Added the 2 member nodes
    • Selecting clustering for sql agent only (my need)
    • Created an ad service user (mandatory for mssql clustering, no local users)
    • Made this user os local admin + mssql sysadmin
    • Added impersonate user at the sql agent level (because common for sub servers and subclients configuration)
    • Added sql server/instance: mssql vip \ cluster group
    • Then created subclients as needed

 

I don’t explain scheduling and transaction logs concerns here.

I’ve tested with manual failover, all is ok.

 

Thanks Commvault people.

 

Regards

 

Userlevel 7
Badge +23

Awesome, thanks for coming back and sharing (and so thoroughly!).

Userlevel 2
Badge +6

Hello,

I have an issue with the AutoDiscovery of SQL Instances in my Windows Failover Cluster configuration. I have installed the SQL Agent on all two Cluster Nodes and created a Server Group in the Command Center like mentioned here Configuring a SQL Failover Cluster (commvault.com).

The Server Group looks fine now but no SQL Instances are automatically added and a manual adding is not possible. I tried it with the Local System Account and a imporsonate User but no luck here.

Kind Regards

Florian

Userlevel 6
Badge +14

@flokaiser  Are you certain that you added the cluster role and not the cluster group itself? 

If you check out this video around the 3:00 mark note cluster roles setion

 

Userlevel 2
Badge +6

Hi,

 

I added only a Server Group with the two Nodes. Is there any autodiscovery of the Failover Cluster Configuration? I did this like in this Documentation Configuring a SQL Failover Cluster (commvault.com)

I thought the way which is shown in the Video is the old way to configure it.

 

Kind Regards

 

Userlevel 6
Badge +14

@flokaiser That should discover the clusters, etc Also this is traditional SQL cluster and not an AG?

Assuming agent is installed on all nodes, etc might be best to open a case to review deeper.

Userlevel 2
Badge +6

This is a traditional Cluster and not an AG ;) 

The Agents are installed on all nodes.

Userlevel 6
Badge +14

@flokaiser  Okay thanks for confirming. Might be best to open a case with support for further investigation.

Reply