We have been seeing quite a lot of alerts of high CPU in one of our vCenters. We have 4 in total, 2 in each of our data centers, for VDI deployment. The vCenters serve up not quite exact VM counts across eachother, but, they're fairly close in terms of load. Especially the primary vCenters on each side. All the vCenters are VCSAs, version 6.0.0.3000. Each have 8 vCPUs, each have 32GB of memory. They all live on PowerEdge M630 hosts, with the host resources not even remotely close to max.
However, just this one vCenter (DC1 #1) gives us those alerts. When I SSHed into the vCenter and did a top, I saw that vmware-invsvc seems to be the service that is crushing it. As seen below. None of the other 3 vCenters looked anything like this. DC2 #1 is the closest to it, and while he has a similar load and high usage of the same service, it's not as drastic.
Has anyone come across this previously, or something similar? We're wondering if this has anything to do with some inconsistent recompose operations we see from this vCenter. I did migrate him to another host just to see if it made a difference, but it did nothing.
Below the first image is of our problem vCenter, the last 3 are the 3 remaining VDI vCenters.:
DC1-1 (Problem) 42 Hosts / 3877 VMs
DC1-2 (Fine) 39 Hosts / 2881 VMs
DC2-1 (Fine) 42 Hosts / 3381 VMs
DC2-2 (Fine) 39 Hosts / 2204 VMs
The guest VMs across them are identical in terms of resources. They are linked clones all off the same parent image types.
I do have a ticket open with support, but, thought I'd see if any forum members had some ideas.
Thanks in advance.
