At a customer I have been deploying a new System Center Virtual Machine Manager 2012 Sp1.

I was here and deployed their previous solution with VMM 2012 and BareMetal deploy of their hosts. They have been patching and updating the hosts with the integration of wsus in VMM. As I described in an earlier post my first intention was to upgrade the database but after some investigation and trials I decided to start fresh instead…

So with the new installation I wanted to add the old Hyper-V cluster and manage it from the VMM 2012 SP1 and also be able to do migration from the old bare metal deployed hosts and cluster to the new cluster.

After adding the 2008 r2 cluster I had two nodes that did not respond correctly and with the following error

I tried to restart the VMM-agent on the hosts and that did not help, also tried to reboot the servers, which did not help either.

On one node I also tried to uninstall the VMM agent on the host itself and then to get it back into VMM involved some VMM-database massage, and that did not help either.

So what did help?!

Rebuilding the performance counters as this KB2554336 describes

To rebuild all Performance counters including extensible and third-party counters, type the following commands at an Administrative command prompt. Press ENTER after each command.

Rebuilding the counters:
     cd c:\windows\system32
     lodctr /R
     cd c:\windows\sysWOW64
     lodctr /R

Resyncing the counters with Windows Management Instrumentation (WMI):
     WINMGMT.EXE /RESYNCPERF

Stop and restart the Performance Logs and Alerts service. 
Stop and restart the Windows Management Instrumentation service.

and then restart the VMM agent did solve the above issue and now the hosts reports nicely in VMM

The hosts had been reporting correctly in the old VMM 2012 so what caused the per counters to go bad, maybe the installation of the new VMM 2012 SP1 agent? If you have had the same issues, please comment this post and share your thoughts!