Citrix VAD 2203 CU3 on Nutanix Cluster - Power State Unknown

[Ken Z]

Hi Everyone

 

I've got a Nutanix Cluster running Citrix VAD 2203 CU3 and a combination of Windows 10 and Windows 11 VDI Desktops deployed using MCS. This has been running fine for well over a year now. It was upgraded to CU3 for security reasons.

Recently some of the VDI Desktops has started reporting their Power State as unknown. Nothing has changed (other than upgrading to CU3) in the environment that I'm aware of.

I found  a nice script for checking the PowerState issues here... https://www.citrix.com/blogs/2023/02/13/how-to-fix-vms-that-continuously-show-an-unknown-power-state

 

This script is reporting the following 

 

Cause : The hypervisor connection is unavailable due to SiteServiceStopped.

Suggestion : Check the hypervisor connection named NutanixVDI.

 

Running a health check on the Nutanix cluster reports all tests Passed

running a test on the hosting connection in Citrix Studio reports all tests successful.

 

Anyone used this script before and know what "SiteServiceStopped" refers to? The Citrix Broker Service on the delivery controller is running.

 

Thanks and  regards

 

Ken Z

[Jeff Riechers]

This usually comes from the Hosting connection from Controllers to your hypervisor.  I would try doing a test on that hosting connection to nutanix.  And if that test fine, do a full reboot of your controllers as they may have gotten some bad data that they are applying.

[Ken Z]

Hi Jeff

 

thanks for the response. 

I'd already ran a test of the hosting connection as mentioned in my OP, which passed all tests.

Having said that, I rebooted the controllers last night and it seemed to fix it, so even though the tests said they passed, something was obviously not well.

 

So, goes to show, don't assume because the tests passed, all is ok ?

 

Regards

 

Ken

[Jeff Riechers]

When in doubt, reboot.  

 

Hopefully it doesn't happen again.

[Ken Z]

Hi Jeff

 

well, it's happened again, and brought all the VDI desktops in the Nutanix cluster into an unknown state, meaning the entire farm was unusable.

Quick fix was to reboot one of the delivery controllers (there are two DDCs). All Citrix services were running so i would assume it some type of bug in 2203 CU3 interacting with the Nutanix API as this has happened twice in the space of less than a week.

 

I'm logging this with Citrix support this morning.

 

Regards

 

Ken

[Ray Davis]

KAZIMIERZ ZYGMUNT what was the outcome here?  Was it a bug in the plugins?

[Andreas Zuegler]

Hi,

 

was there any outcome with the citrix support? We are facing the same issue.