After a double Fault on a 2 node stretched/ROBO cluster multiple objects are in an inaccessible state.
Fault 1. Witness host for the environment is down or partitioned from the cluster.
Fault 2. Cache disk failure on physical host resulting in the disk group being offline
VMware vSAN (All Versions)
vSAN requires at least three fault domains to support FTT=1. Each fault domain consists of one or more hosts. If you lose more then 1 fault domain the data will become inaccessible until that fault domain is brought back online.
Open a case with Broadcom Support for further assistance to prevent any permanent changes that could potentially lead to Data loss.
vSAN Skyline Health - Data Health - vSAN Object Health - inaccessible objects