High availability cluster

[phillbl]

Hi,

we are testing the failover & were wondering if it were possible to set the preference on which node the VM machines would migrate to if the nodes were available. we have two nodes, we tested a failover, they all vmotioned ok but we would like them to move back to the other node when back online. Is this possible?

Thanks

[hstagner]

Hello pjudge,

There are two parts to this scenario. In order to understand what's happening it is important to distinguish between vmotion and HA.When they failed over, your VM's did not VMotion (live migration). They couldn't because that would require two functioning hosts. What happened was HA (High Availability) kicked in and restarted your VM's on the other node.

Where Vmotion comes into play is DRS (Distributed Resource Scheduler). If DRS is set to automatic and it sees that some VM's would run better (need more resources) on the other host when it comes back online, then it will VMotion those VM's to the appropriate host. If you have DRS set to manual, then you can generate suggestions for the cluster and manually take action on those suggestions if it is needed.

I hope this helps.

Regards,

Harley Stagner

[phillbl]

Hi Harley,

Thanks for the answer. So when HA kicks in & restarts the VM machines on a different node they will stay on that node as this has not failed. If ther other node comes back online i assume we have no way of setting preference of the nodes that we would like the VM machines to be on? They will just stay on the other node until a failure.

Thanks

Phill

[hstagner · Accepted solution]

Hello pjudge,

The VM's will stay on the node if you only have HA configured. For automatic load balancing after you recover your second node, you need to enable DRS (Distributed Resource Scheduler) and set it to automatic. When DRS looks at the cluster and sees that balancing needs to occur, it will VMotion your VM's to the appropriate nodes automatically.

Regards,

Harley Stagner