Restarting hostd and vpxa on ESXi in Horizon Environment

[any_guy]

Hello,

so we've had a ticket open with Broadcom, Wolken and Omnissa for a while now. I'll leave out the details, but it's basically being tossed like a hot potato.

The most recent steps that Wolken support wants us to undertake, is to restart the hostd and vpxa services on the hosts that are running our Horizon Environment.

Now they did confirm it will not cause any downtime, however, I'm very hesitant to believe that. I know for a fact, that when restarting hostd, the host will go into 'inaccessible' mode in vCenter for 2-5 minutes. I can't see this NOT causing issues with Horizon. Anyone have experience with restarting those two services in a Horizon Environment?

I'm not entirely sure why they are persisting in us restarting those services, when we already did a complete Horizon Environment sequential reboot, twice, without any impact on the issue.

Cheers,

[Mikhail]

if I got it right, they want you to restart hostd and vpxa

/etc/init.d/hostd restart

/etc/init.d/vpxa restart

this will cause the host management "freeze" for a couple of minutes indeed.

the VMs will keep running, but if Horizon will issue any management command in that timeframe it might fail, like for example reverting a VM in the IC desktop pool.

you'll have to resolve the errors if they will not be automagically solved.

[StephenWagner7]

One thing I want to add, is that if you're ever doing maintenance like this during production, it's always a good idea to make sure you provision some extra spares when running Instant Clones, and then disable provisioning for your desktop pools, just to stop the calls to vSphere/vCenter/ESXi from Horizon.