I'm trying to understand the best or most efficient setting for "Number of Host Failures allowed" setting in HA on a cluster of servers. If I have 5 hosts in my cluster, what should this be set to?
Talk about timing. I am also in the middle of configuring three DRS/HA clusters and was not sure about this number. I have seven hosts. Three of the hosts are DL580 G2's and four are DL580 G3's. Two of them are also on a different subnet. I have three seperate clusters because of processor incompability, different subnets and resources available. If I have two host servers in the cluster then I would think that one host failure would be the number allowed. Thoughts? I would think that with any failure you would want the guests to move and restart on another host. Is that what this setting is asking before I leave it at "1".
