Testing Cluster

[kevin_lane]

Without going and powering yanking the server to test, I was wondering if there was another way of testing my S2D cluster, 

 

if i did this Suspend-ClusterNode -Name "NodeName" -Drain and then Resume-ClusterNode -Name "NodeName" -Failback Immediate

 

would you say this is a safe option 

[CrootUK]

Whilst i’ve done this endless times on a traditional hyper-v cluster, not s2d so take this advice with a pinch of salt

 

I’d personally do it via failover cluster manager; pause a node let all VMs drain which you can watch in vm/hyper-v view ensure the node is “paused” 

effectively then your running on how ever many nodes you have left. At that stage you should feel comfortable enough that a that remaining nodes can run the your environment. 

 

If you wanted to you could then power down the host and go as far as pulling the power out (imo not needed)

 

Once ready power the node up and give it a few minutes, then run “get-storagejob” in powershell to check the storage sync is good (which does run even on paused nodes to my knowledge) then resume the node in the node view in failover cluster manager. 

 

If your using Windows Admin Center advice is similar but a little different. 

 

your powershell is fine, pro with GUI method is you get to see it doing stuff, vms moving, storage moving etc.  

Edited May 6, 2025 by CrootUK

[kevin_lane]

Due to this not being in production i decided to run the powershell script 

 

just to give you some back ground

 

its a 5 node S2D cluster that is on a 3 mirror,

 

the nodes (host) evicted the vms as predicted, and then when brought back online the cluster then moved them back 

 

whilst vm was moving back and forth access was to the vms was maintained and no errors occurred.