My Desktop is running RedHat Desktop 5.4 64-bit. On that I'm running VMWare Workstation 6.5, and I have a Vista Virtual Machine.
In the Vista VM, I was running the vSphere client and connected to a HP ProLiant DL380 G4 server running ESXi 4, downloaded from HP. The ESXi server has two local datastores and one NFS data store served by FreeNAS.
I was logged into the Vista VM as a user, not administrator. I was converting several VMWare Server VMs to this ESXi server. I had converted 3 CentOS and 1 W2K VMs. The Windows VM got a new adapter and got a DHCP address. I fixed that but the Vista VM had cached the DHCP address and not the static.
And here we go; as a regular user I can't ipconfig /flushdns, so I did a little trick, in VMWare Workstation I disconnected/ reconnect the network interface for a couple of seconds. A couple of seconds later, the ESXi 4 server reboots!
After it rebooted, the W2K machine did not appear in the inventory. It was on the datastore so I added it, and it booted ok.
I have not been able to reproduce this. Is there a log I can pull from anywhere that I might be able to submit to support? I want to know what the cause was, if I can trust ESXi 4 on this machine, or if I might have a hardware problem. I might have gotten a different IP address during the disconnect/connect stun, which isn't happening in my tests now.
tl;dr I pulled the network cable of a desktop with an active vSphere client connection and the ESXi server rebooted.