I have 3 ESX 3.5 hosts in a virtual datacenter within a cluster, which has HA and DRS enabled. If I boot all machines (set to 2 CPU + 2048GB mem), one on each host, everything looks good. To start with. After a few minutes I start getting Tasks/Events notifications of each host 'Configuring HA'. If I am luck it will then start to swap round the guests. If I am unlucky the hosts become unresponsive, with some having the errors of: 'An error occured while communicating with the remote host.' I am then prevented from connecting to the hosts at all and the guest VMs become 'orphaned' and greyed out. I am unabled to connect to any of the host servers in any other format other than the DRAC card (remote hardware access to the server). Once connected this way, I find the console to be completely unresponsive, on key press can change it. Consequently I cannot sign into the cosole and have to restart the whole host OS. Once this has rebooted the host acts as if nothing has happened and starts operating as normal.
I have taken HA and DRS off the cluster configuration now in an effort to see if this has any effect. I am currently running on EVAL licenses, waiting for management to pull their finger out for the full licenses after a bank buyout (not HBOS!). I know I am therefore limited to the basic 2 CPUs per guest, but I cannot see any other limitation that would cause this behaviour. I am running on DELL M1000 chassis with M600 blades and Brocade SAN switches, all of which appear on the supported client list. VCS is V2.5, straight install and the hosts are ESX3.5 straight install. I have not applied any patching to the VCS or ESX servers yet, although if this is a patch fix htat would be brilliant.
I would appreciate any insights on this, I am a Windows chimp, but capable of following LINUX commands.
