Installing SQL 2016r1 Cluster on Server 2016
(VMDK/Windows/SQL Clustering Part 3)
Other Articles in this Series:
Currently, I have a one-node Windows 2016 Failover Cluster set up, and in this exercise, I’m going to install and set up SQL2016SP1 Standard as a one-node SQL Cluster. Don’t bother Snapshotting this baby before you do this, because you can’t snapshot servers with shared VMDK’s in VMware. You will have to find an alternate way to backup it up first if you desire to roll back from failure. Veeam Works well if yours is up to date, but I’m sure there are other solutions out there. As of 9.5u3 you can use the Veeam Windows Agent with a Protection Group to accomplish this task (List Price for Agent is $127.50 per server per year, it is subscription based).
First things first. You will want to give the Windows Failover Cluster Computer Object permissions to create computer objects in the Computer OU in Active Directory. Open Active Directory Right Click on the “Computer” Object Click Permissions.
Next, run the SQL setup -Select the Installation heading and select New SQL Server Failover Cluster Installation.
I like to have Microsoft Update see what it has available now.
For me it was KB 3164398 and I chose to install it.
This will bring up the Install a SQL Server Failover Cluster Rules Wizard. I had a warning, but it was from the verification report having a warning. Nothing shocking.
You can see my feature selections in the picture below:
Create your SQL Server Network Name, this is separate from your already created Windows Server Failover Cluster Name vBoB here.
Name the Cluster Resource Group. Note it tells you that you can’t use Available Storage or Cluster Group as a name. That window can be very confusing. You could also just leave the default in place.
Adding all my Shared Disks to the Cluster.
Give a unique IP address to the SQL cluster.
You are going to have to create and add some accounts here. I created two domain users in Active Directory and followed the best practice of separate accounts. NOTE: I had to do this exercise from the start a second time because I forgot to do the OU permission change I mentioned earlier. The installation will fail the last second, and can kinda leave it in a “broken state” since instead of rolling back when you hit “cancel”, it leaves it with a broken DB Engine install. I highly suggest if you have an issue, keep attempting fixes and hit “retry”, never “cancel”.
Setup your SQL Authentication as needed for your organization. I prefer mixed mode.
I setup multiple drives, so I changed information on the Data Directories and Temp DB headings.
That should be everything you need to set up a one node SQL failover cluster. In Part 4 we will add a second server to SQL Cluster, look for that soon.
Also be on the look out for backing up the cluster with Veeam, because we got to figure out how to backup this bad boy!