Hello,

this is somewhat related to:

Unfortunately, I do not have the resources to set up a couple of hosts and test this. Has anyone experienced a host failure in a cluster running instant clone VMs? With vSphere HA enabled, if the VMs on the failed host are powered on a working host, will that work with Instant Clones? Specifically, will Horizon be aware that the location of those VMs has changed, or will they just go into the 'missing' state?

Cheers,

4 minutes ago, any_guy said:

Hello,

this is somewhat related to:

Unfortunately, I do not have the resources to set up a couple of hosts and test this. Has anyone experienced a host failure in a cluster running instant clone VMs? With vSphere HA enabled, if the VMs on the failed host are powered on a working host, will that work with Instant Clones? Specifically, will Horizon be aware that the location of those VMs has changed, or will they just go into the 'missing' state?

Cheers,

I've never seen any official documentation on this matter, but I've had the thought as well.

Technically, with HA, the VMs would get restarted, however since the user isn't logged on they wouldn't be resumable, I'm pretty sure these VMs wouldn't be marked as available, and would either be orphaned and/or shutdown/destroyed.

Without HA, if the user attempted to re-connect, I believe they would just get whatever VM is available in their associated pool, therefore you wouldn't be wasting resources on starting HA VMs only to be orphaned and/or destroyed.

Would love for an official response though, or a reference to some documentation.

In a persistent environment, you'd most definitely want HA enabled if you can support it.

1 hour ago, StephenWagner7 said:

I've never seen any official documentation on this matter, but I've had the thought as well.

Technically, with HA, the VMs would get restarted, however since the user isn't logged on they wouldn't be resumable, I'm pretty sure these VMs wouldn't be marked as available, and would either be orphaned and/or shutdown/destroyed.

Without HA, if the user attempted to re-connect, I believe they would just get whatever VM is available in their associated pool, therefore you wouldn't be wasting resources on starting HA VMs only to be orphaned and/or destroyed.

Would love for an official response though, or a reference to some documentation.

In a persistent environment, you'd most definitely want HA enabled if you can support it.

For the described reasons, we have HA currently turned off in our Horizon Cluster. We are mostly using Instant Clones and I'm afraid HA would cause more problems when turned on.

I would also appreciate an official response on this.
 

ic can't be restarted keeping the session so it simply doesn't make sense. It's better to just recreate than to reboot imho