Hi,
Let`s say we have an 3 nodes(node a,node b,node c)vcs.When i perform an failover(maintenance reason) of an service group from node a,the best practice is to failover to the most resourceful node(the node that has the less cpu usage,and more memory).Right?Thus first i verify on node b and c the cpu load,so on.right?
To me the nodes within vcs should have the cpus load at most 30% or 40%,because if one node has problems(and the service group failover),the node who take over the service group,should to has enough cpus capability to absorb also the workload of this new service group.So if the nodes in vcs have normally 30% cpus load,then in this case the takeover server will have 60% cpus load-and still runs the services.Right?
An single node,not in a cluster,could have higher figure as cpu load over an node who is part of the cluster.Right?
A single node-no cluster-could has 50% cpus load.
thanks so much,
marius