Occasionally, in-progress VM clones or relocations will cause all subsequent tasks to queue until the clone/relocation completes. This includes VM power on, reset, vMotions, etc. Connecting to the individual VM host (bypassing vCenter) will allow a VM to be manipulated, so it seems that it's a vCenter Server issue.
This used to happen maybe once a month to us, but lately it seems we have users notice the problem at least once a week.
Our vCenter Server lives on a standalone host with 4x2GHz CPU cores and 8GB of RAM. The system runs Windows 2008 x64 and we are using a local SQL 2005 Express instance to hold the vCenter database. This server handles a single datacenter with 1 cluster of 8 VM hosts, (currently) 754 total VM objects. We typically have ~40 concurrent vSphere client sessions. I realize our environment exceeds the published maximums for the express database, but the system is basically idle. No resource pressures at all.
Is this a common situation? If so, is there a known workaround or configuration setting I should be looking at?