Ways to test the failover

[chrisguk]

My setup includes 2 storages with Windows 2016 install and to HyperV host servers also with server 2016 installed. I have the 2 heartbeat and Sync networks together with a seperate LAN for management and Hyper-V.

This whole platform sits in a nested VMware host by the way.

I suddenly found that I am unable to actually start any VMs with this setup although of course I was able to install Hyper-V.

What is the best way to actually test the Failover of the Compute and Storage layer though?

Should I have storage spaces direct also installed or not?

[Ivan (staff)]

Hello chrisguk,
Thanks for interesting in StarWind solution.
To check the failover you can turn off Storage node and double-check the disks migration to the different storage node. The same with Compute node (should migrate the roles to the different compute node)
Please note, I do not recommend to use any cache on StarWind vSAN to avoid any troubles in this particular configuration
I do recommend to provide with StarWind VMs thick provision eager zeroed hard disks which later will be presented by StarWind vSAN to the nested HyperV configuration
You can find all information about Compute and Storage separated scenario on HyperV environment by following next link: https://www.starwindsoftware.com/starwi ... -v-cluster

[chrisguk]

Hi Ivan,

Thank you for your reply.

Presently Starwind looks great to me as solution for us. Something that the documentation doesn't present is how you would store the VMs. I have to assume that once the CSV has been created you need to create a Scale out file server and then the shares for the VM storage? Someone else also suggested to me that you would need to install storage spaces direct, is this correct in your opinion?

[Ivan (staff)]

Hello Chrisguk,
When you will create StarWind device and connect them via iSCSI you will able to create a Failover Cluster and add this disks to the Cluster.
Then you can Add to this disk a Cluster Shared Volume. After this step, you are able to see the CSV Volume by this path: C:\ClusterStorage\volume1
To make possible failover all VMs should be located on C:\ClusterStorage\volume1 or you can create Scale-Out File server and locate all VMs on this SoFS.
You can find all information about StarWind in our resource library https://www.starwindsoftware.com/resource-library or StarWind Knowledgebase https://knowledgebase.starwindsoftware.com/