Disable HA and vmotion for 2 vms on a cluster

[Virtguy]

Hi, i have a customer with vsphere enterprise out of Support/maintenance.

Due to budget, he´s thinking about purchase essentials plus kit for 3 host more than renew or purchase 3 new enterprise licenses.

Problem, he have 2 vm cluster nodes virtualized of a specific service..actually host affinity is configured to avoid the posibility that one host will  have the two node cluster vms ( have lots of  rdm disk ).

basically all the machines of the cluster ( except this two vms) uses basic funct. of ha cluster ( licensed to essentials+)

customer is not using FT at this time.

customer not have DRS configured more than Host affinity.

My question is.. if i:

- disable ha for this two specific vms ( restart priority disable )  === no automatic vmotion for this vms

- Create a port group on Host 1 called "host1" and in host 2 called "host"2 and attach this vms on those port groups. === no manual vmotion for this vms.

Never never i will have this two vms on the same host. ( includes power electrical loss.. accidental shutdown of hosts etc etc etc )???????

is this correct ?

Thanks in advance community !

[unsichtbare]

HA does not utilize vMotion at all. It re-starts VMs on remaining hosts when unplanned hardware failure causes one host to fail.

Your idea about creating port groups with slightly different names would probably work, but it seems like a very unclean way of accomplishing the goal.

Why would you want to exempt 2 VMs from HA again?

[RichardBush]

It sounds like you have a 3 node esx cluster ? is that correct ?

As, if you had a two node cluster you could just disable HA for the VMs and they wouldn't power on again on the remaining host ( includes power electrical loss.. accidental shutdown of hosts etc etc etc ). As you say the only reason you d want to two port groups would be vmotion, but now a days it will just prompt you to change the PG should it not exist on the destination, this is more to cover admin errors i guess ?

if it is indeed 3 hosts;

An easy way i could see this setup working would be to set a specific host for failover, and then enable HA. Its not clean, or exactly money saving as you d have a host sat for redundancy only, but would work.

The other options would be to try and  utilize admission control, and resource pool or vapp for your VM clusters, so that they would consume to many resources to allow the two to power on on the same host, this would take some planning / testing / playing around with but seems to be an option.

Rich

[Virtguy]

thanks for the replys.

Days ago, we decided to forgot the idea of purchase essentials plus , thats wy the customer is in production 24x7, and i am affraid my "manual cheap" host affinity would be problematic.

Thanks