I have a small cluster of mixed 4.0 and 4.1 servers (3 each) governed by a VCenter running on one of the newer servers. Up until just after Christmas, it seems, VCenter has been pretty snappy.
Now, though, launching the vSphere client when pointed to the VCenter server ("Zorin") takes quite some time, and once it's up, VS takes its sweet time in updating the interface, showing options, etc. It just took 30-45 seconds to show me the perf graphs on one of my VMs, for example.
This problem doesn't exist if I point my vSphere client at one of the hosts directly -- even the host running Zorin. When I do that, everything's quite snappy.
Zorin itself, when I remote desktop into it, is also quite slow. (Other VMs are not.) However, there's no obvious reason for this -- it does not appear to be CPU bound, nor does RAM appear to be a problem, and it has plenty of disk space free. The VM *is* modestly spec'd -- 1 vCPU, 4GB of RAM, Windows Server 2008 32-bit -- but if it's not yelling, I figured that was fine.
My VCenter is backed by a local instance of SQL Express 2005. After googling about yesterday, I found that DB index fragmentation can produce slowness, so I set about fixing that. My indices are no longer fragmented, but that did not address the problem. I have also, I should not, rebooted, but that didn't help either. Neither do I see anything immediately in the Windows event viewer, though I'm also not entirely sure what I'd look for there.
The logs for vpxd DO seem to be quite active -- theyr'e cycling over every 12 or 15 hours. That strikes me as a lot, but I don't see anything in the vpxd-xxx.log that seems really out of place, unless it's a problem that it's doing these logged sync operations so frequently.
Where else shoudl I look for trouble here? I'm kind of at a loss, and fear this slowness is an indicator of some larger, more serious problem that would have production implications.
Thanks.