I am planning o install a VirtualCenter Server on a separate dedicated Windows based host machine and cluster 4 ESX Enterprise v3.5 hosts to it. I plan to run a few mission critical servers on the cluster and configre HA, VMotion and DRS on it.
I understand that the VC shall be monitorng the cluster and the VC shall be responsible for all the decisions of the HA. VMotion and DRS functions. In case the hardware of my VC server fails the entire faul-tolerence of the cluster is gone. So VC becomes the single point of failure in the entire solution. Please correct me if I am wrong. If so, what is the way out to make my entire solutio fail proof.
java is correct - HA functinality gets pushed down to the ESX Host - a running VC environment is required for vmotion so it will stop and since DRS relies on vmotion it will stop as well - now what about your running vms - they will continue to run unaffected - VC acts soley as a Proxy and should not impact the running vms or your ability to start vms directly on your esx host