Hyper-V Polling OK, but Could Not Poll

jack.vaughan

Greetings thwackers!

I have an issue that has come up often, but I usually just fix it and move on.  Thought maybe I should get some input this time...

Two Hyper-V hosts just had interfaces go unknown, and the status of associated elements are not matching up.  For instance, the Disk Volumes all show green in the pop-up hover display, but they aren't actually collecting data.  Same for RAM and CPU.  Further, the Hyper-V Host Details shows:

Management	Edit Host
Host Status	Could Not Poll
Host Status Details	Provider load failure Scope: <IP ADDRESS>\root\CIMV2[<USERNAME>] ErrorCode: 0x80041013
Host Polling Status	Hyper-V Polling OK
Product Name	Microsoft Windows Server 2012 Datacenter

On the box itself, I'm seeing successful logins, but then queries fail out.

My question here is this:  shouldn't "Could Not Poll" and "Hyper-V Polling OK" be mutually exclusive?  Or are these not actually related?

Thanks in advance!

jayd

I don't have an answer to your specific question, but I do seem to be experiencing the same thing. I don't know if the cause is related. I guess it depends on what you're doing to "fix it and move on". I'll explain what's going on in my environment and the corrective action I'm taking in case it helps put anyone else on the right track.

A couple of months ago, I started migrating our Hyper-V host management from SCVMM 2012 to SCVMM 2016. Ever since then, host servers will randomly start failing to refresh within SCVMM. When this happens, the VMs are no longer manageable using SCVMM. Sometimes, they're not responsive from Hyper-V Manager either. Then I started noticing that at the same time this condition existed in SCVMM, Orion was reporting that it could not poll the host server. It was reporting the exact same error and details you describe.

It took some research, but I found that the problem was apparently related to a failed state with WMI. Apart from rebooting the host server, the non-disruptive fix was to restart the Windows Management Instrumentation service along with its dependent services. Afterwards, Orion could successfully poll the host and SCVMM could refresh the VMs. Since this is only happening to servers I have moved to SCVMM 2016, I have to assume something in the SCVMM 2016 Agent service is causing a problem with WMI. Hopefully a patch or some other solution for the cause is forthcoming, because it's a real PITA.

Jay