I have a windows host that is behaving very slowly. All the counters (CPU, Disk, etc) appear to be normal.

The virtual host itself is ESX, and is on a VNX Cluster with 13GB free disk space on a 2 TB disk.

I am assuming that ESX doesn't like the low disk space on that LUN and as a result is having issues with all VMs resident on there.

Right now I'm VMotioning 120 GB off that LUN, and it will take some time. Is there any analysis I can do while my host is "acting slow"?

My goal is to identify a Windows (or VMWare) metric that will allow me to predict this performance degradation.


ESXTOP is probably the best tool to get real-time performance statistics from the ESX host. I would start there and consult it when you're experiencing the slow down. You can also set up Perfmon on Windows and start collecting statistics on CPU, Memory and Disk I/O. I've attached two KB articles to help you configure and understand the output.

ESXTOP: https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1008205

Perfmon: https://technet.microsoft.com/en-us/library/2008.08.pulse.aspx


From inside the Windows VM, open task manager and, in the performance tab, select "show kernel times". If the red line representing kernel wait time show high peak, this means that some low-level operation (ie: I/O, IRQ handling, and the like) is taking too much time. On a VM, this usually means that the host system is being overloaded.

Obviously, from the hypervisor itself you should be able to get more (or more accurate) usage information and statistics.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.