An ESXi 5.x host is member of a cluster in a vSphere 5.x infrastructure.
HA is enabled for the cluster, the cluster settings say that the VM restart priority is High and critical VM are configured to use cluster setting for the VM Restart Priority.
So I presume that if a node fails the critical VMs are restarted on another node, is it correct?
Shoud I configure for each node the "Virtual Machine Startup and Shutdown Settings" so that critical VMs are restarted automatically or is it useless?
What is it supposed to happen if the whole of thecluster fails and restarts (let's say, for a general prower failure)?
How should I configure the restasrt priority (either at cluster and node level) so that all the ciritical VM restart in case of a node or cluster failure?
Regards
marius
Good morning Marius,
So, for the first Scenario:
Whenever a node goes down and your machines are located to a SAN Storage that is connected to the failed node and a working node, the machines would be restarted on the working node. When the failed node comes back and there is no DRS enabled, you have to move back the machines manually.
When a complete Cluster goes down your vm's are also down. As they have no host to start.
When you doing the setup on a cluster level, it will work for all included nodes into this cluster. If you have cluster-wide settings and then change a setting of a node of the cluster this setting will be used instead of the cluster-setting. But even only for this node!
Hope this helps a bit....
greetz,
Matthias
Thank you for your message.
So it's useless (and maybe dangerous) to set Virtual Machine Startup and Shutdown Settings at node level after setting at cluster level, am I right?
Regards
marius
It depends.. If you have enough tolerance ( ie. if your cluster can tolerate the VMs to be running in case of one host failure). If you dont have enough resources HA will not power on few machines. So if you set the priority for your production VMs, rest assured those will be powered on first