-It's important for customers to be aware of what happens if objects are created in a vSAN stretch cluster when there are not enough fault domains in the cluster. This unavailability of enough fault domains can be due to any reason such as site/host failure, witness failure, or network instability in general.
-The behavior of forceProvisioning of an object exists to support the use case of site recovery. When a site is failed and is being recovered, users may want to revert snapshots and power on VMs. These operations require new object creations and such creations should go through in such circumstances.
-Objects which are force provisioned because of this behavior can be identified by looking at the nonComplianceAcceptable attribute in the object config. This attribute will be set to 1. To see this run cmmds-tool find -f python -t DOM_OBJECT <uuid of non-compliant object>
For example cmmds-tool find -f python -t DOM_OBJECT bff18362-ded2-d6c7-300f-e4434baf4ec0
{
"uuid": "bff18362-ded2-d6c7-300f-e4434baf4ec0",.
"owner": "5fc78698-b704-3616-914a-e4434baf9f90",.
"health": "Healthy",.
"revision": "10376",
"type": "DOM_OBJECT",
"flag": "2",
"minHostVersion": "3",
"md5sum": "43ea20213a497a925998d4e9afd74be6",
"valueLen": "4512",
"content": "{\"type\": \"Configuration\", \"attributes\": {\"CSN\": 33694, \"SCSN\": 23351, \"addressSpace\": 2199023255552, \"nonComplianceAcceptable\": 1, \"scrubStartTime\": 1652814272031426, \"objectVersion\": 14,...
-Such objects should also be reported as non-compliant in vCenter.
-Once all 3 sites are available with their respective hosts in the cluster, the objects will be automatically reconfigured to fix the compliance.