janakors, the main point is the minimum host/node requirements for the Failures To Tolerate(FTT) with the Fault Tolerance Method (FTM) in use - if you want/need to use FTT=1,FTM=RAID5 this is 4 nodes, if using FTT=1,FTM=RAID1 this is 3 nodes, FTT=0,FTM=RAID0 it is technically 1 node (though of course this is not a good idea nor supported).
Thus you can't create or retain availability of RAID5 data with just 2 nodes - if this were a bigger cluster (6+ nodes) then provided there is the capacity and compute resources available to store and run all VMs, taking 2 nodes out of it shouldn't pose issues.
But if you only have 4 nodes (and are not tight on space and good at managing it) then going FTT=1,FTM=RAID1 would allow removing a node in future, or if the workload has relatively low space requirements you could even do a 2-node and run the witness on other infrastructure.
"so i can proactively counter such situation and have confidence to tell the management that yes we can remove or NO we can;t remove"
The main takeaway should be that you should not be removing nodes from a cluster beyond what the Storage Policies in use require to function and retain data availability and redundancy.
There are plenty of resources available for further information, but what I would advise starting with is if you want information about specific areas or aspects (e.g. basics cluster-designs and types) then storagehub should be your go-to:
Cluster Types | VMware vSAN 6.7 Technical Overview | VMware
Bob
