ESXi 5.0u2
The host p410i Raid controller memory cache module had failed. The host console became unresponsive(could not enable ssh or restart mgmt agents), turned (disconnected) state in vCenter, the now unmanageable VMs were responsive and running. When the host was shutdown hard to initiate an HA event, HA did not kick in.
To resolve I wrote a script in conjunction with vNugglets.com for implementation with vDSwitch. The Host was hosed but this enabled me to get the VMs re-registered, portgroups assigned, and back online in a timely fashion.
http://www.knightusn.com/home/re-registeringvmsandreconfigurevds-portgroupsremovedfrominventory
I'm trying to find the root cause of why HA did not kick in when the host failed. I submitted the logs to VMware and got a 
shrug response.
Is this a common/known issue with HP's p410i RaidController?
Would an attempt to reconnect the host remove the VMs from the protected list and disable HA for that host?
-vFantastic