Hi All -
I am attempting to set up our ESXi5 hosts to vmotion our guests in the event of a power failure on a host or 2 in our 6 host clusters (we have 4x 6 host cluster environment). My purpose is to let the VMs move hosts in case of 2 power failures in each cluster.
I've ensured we have enough resources to handle 2 host failures, I've then tested the response by just unplugging the hosts to see what happens. However, we lost all the VMs power, but they did power back up on another host within the cluster. I'd like to be able to not lose the VMs like that.
Is this a benefit of DRS? We do not have this turned on, so that would make sense to me.
Thoughts/advice are appreciated.
That is expected behavior as HA will only restart the guests after they have gone down due to the power failure. In order to not have any downtime, the virtual machine(s) will need to use VMware Fault Tolerance which runs the VM simultaneously on two hosts.
DRS will move VMs around based on resources, but it will not vMotion in the event of a power failure since there will be no way to keep the VM up as the current state will be lost when the host goes down.
DRS = load balancing solution
FT = fault tolerance
HA = availability
You are looking for FT. This creates 2 instances of your virtual machine and when a host fails with the "primary" instance the "secondary" will pick up where the primary left.