Hi. We have 2 Hyper-V clusters (one 2008 R2, one 2012). Both are working fine and we manage them from VMM 2012 SP1.
However we have issues with VMM showing incorrect values for various properties like dynamic memory demand, assigned memory, disk i/o, network i/o, disk allocated etc.
Virtually every VM has one or more of these values reported badly. For example, VM that has constant 50-100% CPU usage shows as 0% in VMM, another VM has 30GB dynamic disk with about 17GB used, yet VMM shows the disk allocated size of 16kB, another VMs with DM enabled show 0 assigned memory etc. There are also times when DM counters (demand, allocated etc.) are displayed, but they're completely wrong.
I've tried refreshing the clusters, refreshing the hosts, refreshing the VMs, nothing helps and the hosts status is all green (OK). I even manually (via PSH script) collected the VM info from KVP (where applicable), yet the values are still incorrect.
Before you suggest that, I've tried the obvious things like resetting the WMI performance counters on hosts (more than once), reinstalling VMM agents, removing/adding the clusters from/back to VMM etc.
And ofcourse the values exposed to Hyper-V manager are displayed correctly in Hyper-V manager.
It's like if VMM is constantly polling bad or none WMI data from the hosts and yet there's no info/error about this.
This makes VMM practically useless when it comes to basic monitoring and capacity management. For example the most painful are the incorrect reported disk allocated sizes (this even shows in the VM properties) and assigned memory.
How to diagnose from here? Any ideas?