Has anyone noticed any oddities with performance counters and vSphere 5? Specifically, I'm seeing two problmes:
The weirdness is that this behavior is NOT consistent. Sometimes the above queries work just fine for a period of time (except for the 5 minute interval stuff. That is just broken altogether). And then later things go haywire, then they go back to valid values. This is all running on a GA version of vSphere running in an ESXi 5 Cluster with two reasonably strong hosts (a lab setup of course). This cluster only has 2 running VMs and they are very small WIndows 2K3 systems. So there is almost ZERO load on the hosts at any given time. But, I get the same behavior with an ESXi 4.1 Cluster as well with a loaded environment. So, it doesn't appear that ESXi version is a problem (this didn't happen on a vCenter 4.1 at all). Just curious if anyone is seeing this behavior as well. Thanks.
Hi gilesm,
Even I have noticed this very wierd behaviour of VC5 Cluster. I have also observed that if we batch the performance queries of Host and Cluster together, the data is not returned for cluster. The same has been for a batched operation consisting of query for Resource pool and VM. It would be great if i can get any leads to resolve this issue.
Thanks
Nithin
I have the same issues while retrieving cpu.usagemhz.average and other perf counter for clusters configured on vSphere 5. The data returned seem to be staggered, with a lot of missing points. Can anyone from vmware confirm is this is a bug.
I think it could be a bug. Workload on the ESXi is not the issue, but vCenter/DB is. You may want to file a bug with VMware tech support.