[vMode][Virtual Machines] - vMode navigation showing stale VM state

[omerbutt1478]

Gateway Version: 2.7.4.20
Extension Version that's been used if it applies

Issues

Title:
WAC vMode: Left nav tree shows stale VM count/state (4 VMs, "Running") while VM inventory grid correctly shows 3 VMs "Stopped" — confirmed via Failover Cluster Manager

Body:
The left navigation tree in WAC Virtualization Mode shows incorrect VM inventory for a host — 4 VMs listed, all displayed as running. The Virtual Machines grid for the same host correctly shows only 3 VMs, all stopped. Failover Cluster Manager confirms the grid (right panel) is accurate; the tree is stale, showing a phantom VM and wrong power state.

Reproduced on 2 independently onboarded clusters (Hyper-V Server 2022 and Hyper-V Server 2025) on WAC 2.7.4.20. Issue persists after multiple WAC service restarts and across different client machines/browsers, ruling out client-side caching. Screenshot attached.

Expected behavior
It should show correct VMs inventory and state

in right and left panel. Please see the screen shot attached.

Screenshots & Additional context
I have attached the screenshot

[HaraBria5]

If you right click the host and select Get agent health are you able to confirm its connected successfully and you see active collectors

[omerbutt1478]

If you right click the host and select Get agent health are you able to confirm its connected successfully and you see active collectors

yes sir

[HaraBria5]

The host should have a service running called WAC Agent which runs a process for the basic collector running at intervals to collect data which you confirmed as running via the web interface, this is the process that should be fetching the VM state and passing it to WAC. Do you anything on useful in the event logs on the host itself under Application and Service Logs in the WAC Agent section as you should see events mentioning that it is collecting data and what VMs it’s found etc as the collector runs. Do you also see any events on the WAC gateway itself also in the Application and Service Logs under WindowsAdminCenter that may suggest a problem on the WAC side

If you right click one of the Stopped VMs, in the left panel does it change the state or fire the VM up correctly and from there does it reflect as down if you power it back down. I do wonder and this is just a hunch does the time on the gateway exactly match the hosts.

It is also possible to enable logging on Postgres to see exactly what it is writing the database itself but for the time being I am reluctant to even suggest that at the risk of going too deep into the weeds and should the WAC team have better steps for you to try first.

One way to determine the data it should be fetching is by opening PowerShell on the host and running the script below, the state code should reflect what should be sent back to WAC as 2 will be Powered On and 3 will be Powered Off for example.

$virtualSystems = Get-CimInstance -Namespace "root/Virtualization/V2" -ClassName "Msvm_ComputerSystem"

foreach ($system in $virtualSystems) {
if ($system.Caption -like "Virtual Machine") {
[PSCustomObject]@{
Name = $system.ElementName
Id = $system.Name
StateCode = $system.EnabledState
CreatedTime = $system.InstallDate
}
}
}

[omerbutt1478]

can you reopen the the case by mistake I clicked close.

[omerbutt1478]

The host should have a service running called WAC Agent which runs a process for the basic collector running at intervals to collect data which you confirmed as running via the web interface, this is the process that should be fetching the VM state and passing it to WAC. Do you anything on useful in the event logs on the host itself under Application and Service Logs in the WAC Agent section as you should see events mentioning that it is collecting data and what VMs it’s found etc as the collector runs. Do you also see any events on the WAC gateway itself also in the Application and Service Logs under WindowsAdminCenter that may suggest a problem on the WAC side

If you right click one of the Stopped VMs, in the left panel does it change the state or fire the VM up correctly and from there does it reflect as down if you power it back down. I do wonder and this is just a hunch does the time on the gateway exactly match the hosts.

It is also possible to enable logging on Postgres to see exactly what it is writing the database itself but for the time being I am reluctant to even suggest that at the risk of going too deep into the weeds and should the WAC team have better steps for you to try first.

One way to determine the data it should be fetching is by opening PowerShell on the host and running the script below, the state code should reflect what should be sent back to WAC as 2 will be Powered On and 3 will be Powered Off for example.

$virtualSystems = Get-CimInstance -Namespace "root/Virtualization/V2" -ClassName "Msvm_ComputerSystem"

foreach ($system in $virtualSystems) { if ($system.Caption -like "Virtual Machine") { [PSCustomObject]@{ Name = $system.ElementName Id = $system.Name StateCode = $system.EnabledState CreatedTime = $system.InstallDate } } }

I can turn on VMs, migrate them one host to another etc comletely fine. The only issue is in the right hand panel info is correct and in left panel info is wrong. Please see my first comment and screen shot

[HaraBria5]

That I do understand, the reason I ask about if forcing starting and stopping via the right click menu on the left panel was to see if behavior changes just in case you were using the native tools directly and waiting for the collector to update after some time and so you are able to better marry this up with any times in the logs.

There is a few things in my previous message which your welcome to try which could help pinpoint where your problem may be as the panel to the left is stateful and stored in its database which it gets from the agent, the VM extension seems to still use Powershell remoting at least in this release which means it needs to ask for the information every time you open it

[omerbutt1478]

That I do understand, the reason I ask about if forcing starting and stopping via the right click menu on the left panel was to see if behavior changes just in case you were using the native tools directly and waiting for the collector to update after some time and so you are able to better marry this up with any times in the logs.

There is a few things in my previous message which your welcome to try which could help pinpoint where your problem may be as the panel to the left is stateful and stored in its database which it gets from the agent, the VM extension seems to still use Powershell remoting at least in this release which means it needs to ask for the information every time you open it

Hi Hara, I know you are trying to help but I will not wasteour time anymore and wait for Microsoft to give us a more stable release. There are other issues in this release as well but we don't have time to check each feature. I think it's not the right time to move to Hyper-V from VMware. We will try to re-visit in next 3 years because we need to renew our VMware subscription in next 2 months time