Hello,
Our ESXi 5.1 environment contains three HP DL380 G6 hosts with 72Gb RAM each. We have had HA configured so one of the hosts is completely dedicated to failover. We now have had enough VM sprawl were we need to upgrade our RAM to provide to all our new VMs. Instead of purchasing more RAM, we want to be able to use the CPU/RAM/NIC resources on that HA assigned host by placing VMs on that host but still recover if we loose any host.
So this is how we have it configured under our Cluster Settings:
HA is ON.
DRS is OFF.
Host Monitoring is ON.
Admission control is OFF.
Admission Control Policy is configured for - Percentage of cluster resources reservered as failover spare capacity: 30% CPU & 30% Memory.
now we have 0 hosts specified as a failover host.
Virtual Machine Options have the default Cluster settings as "Medium" for VM restart priority & "Leave Powered On" for Host Isolation response. Under the Virtual Machine Settings I do have low priority/Development VMs configured to "Shut Down" under the Host Isolation Response column.
Under VM Monitoring I have the Status as Disabled. Default Cluster Settings are set at "High".
Datastore Heartbeating is set at "Select any of the cluster datastores".
The idea is to use only 2/3 or 66% of the resources maximum on each host so if one host fails, vSphere will take those VMs and spread them across the two remaining good hosts. Is that possible? Do I need to enable DRS? Is this config correct? Any comments or suggestions would be appreciated.
Thanks!!!