VMware just released a new KB article. The article is on VMware HA and the log files it can possibly generate when restarting of a VM occurs.

The article contains an extensive description of the sympoms but the most important one is this one:

For any virtual machine that VMware HA is wrongly trying to start, it also generates thousands or tens of thousands of vmware-x.log files (one created every few minutes). Contents of each log file shows the virtual machine starting up then failing to start. Despite these logs, the virtual machines are actually pingable and running, and vm-support -x shows them as available running virtual machines on the host.

The resolution for the problem is the following:

1. Stop and start HA on the cluster. (This stops the flooding)
2. # find . -name ‘-*.log’ | xargs rm
   This commands removes all log files. Make sure you run this from /vmfs/volumes!