ExpletiveDelete
Enthusiast
Enthusiast

another option could be to set a percentage of resources to set aside fro HA instead of # of hosts (i.e. 15 or 20% - something LESS than 1/3, just to test). another option would be to temporarily break out 2 hosts into a test cluster and test the single VM on that cluster. It 'sounds' more like an OS/Memory leak issue more than a cluster issue (rather, an OS issue CAUSING an HA problem).

What is the error that prevents you from reserving all of the VM's memory?