My question is: is this expected behavior? Or should I open a ticket specifically for this issue?
We recently had an issue on our DR cluster where "the component utilization for the entire cluster and all hosts in it is 100%" (SR 22305097602, now closed). But we had 50TB free on the 225TB vsan datastore! And we had no issues at all on the production cluster.
One thing we discovered in the process of the SR is that the storage policies on the replica vms (on the DR cluster) were different than the policies of the prod vms. All storage policies on the production cluster are mimicked on the DR cluster (so if there's a "policy 1" on the prod cluster, there's a "policy 1" on the DR cluster, and those policies are identical). The hosts and vsan disks are also all identical between the two clusters.
What we found: let's say a prod vm has "prod cluster storage policy 1" set for the entire vm (not per disk). For the most part, the replicated vm is different: the storage policy is set to "configure per disk", the "vm home" object policy is set to "dr cluster storage policy 1" as expected, but "hard disk 1 / 2 / etc." are all set as "dr datastore default policy" (which is different than the policy that's set on the prod vm).
Which prompts my question: is this expected behavior?
(BTW, the fix was to change the storage policy of several of the largest replicas on the DR stack, and that brought the component usage down. Then, for now, i've configured the "default datastore" storage policy on the DR cluster to be the same as the "prod cluster policy 1" used on the majority of our prod vms.)
Oh, also: we use Veeam to replicate our prod vms to our dr cluster. Veeam has stated that this is not a "veeam" issue.
