Hey Checkyourlogs Fans,
Today, I wanted to write a very quick post showing how you can use Windows Admin Center to quickly diagnose issues with poorly performing Virtual Machines on a Storage Spaces Direct Cluster. We had a call come in from a customer today that their SQL Server in production was running a bit sluggish on our all flash (NVME) S2D Cluster.
Step 1 – Open Windows Admin Center and browse the Virtual Machine Summary Tab.
What I noticed right away is that my CPU looks fine only 16 % utilized and none of my VM’s appear to be running away.
Note: We do have one issue which is the Percent of RAM 74.2% is not good for this 2 x Node S2D Configuration. What it means is that we are oversubscribed and cannot perform automated maintenance like Cluster Aware Updating because there isn’t enough memory to move over to the other nodes. We have ordered more RAM to solve this issue.
So, if none of my VM’s are running away and my CPU looks fine what’s next.
Step 2 – Open the Inventory Tab
The highlighted instance is our SQL Server, and we can see that it is maxed out on memory.
Step 3 – Logon to the VM to identify the issues.
A glance at the Virtual Machine we can see that the SQL Server Service is consuming almost all the available RAM.
Step 4 – Open SQL Server Management Studio and check the Memory Settings.
As you can see, the maximum memory is set at exactly what we have available for the Virtual Machine. So, we have two choices à Increase the RAM or drop this setting back to something like 12GB to free up some room for the OS.
In our case, we decided to add 8 GB of RAM to the SQL Server, which did solve our problem.
I hope this helps you out today
John