In our environment we have vSAN hosts configured with 2 Disk Groups and the VMs have a FTT1 (Raid 1) storage policy.
We have to remove both Disk Groups to recreate a single Disk Group on each host.
At the moment, I am placing a host in maintenance mode with 'Full data migration' before recreating its Disk Group. Everything is fine but it takes a lot of time to complete the maintenance with this mode.
I am wondering if it is safe to use 'Ensure accessibility' instead before removing the Disk Groups and recreate a single one ? VMs policy is Raid 1 so I believe there will always be a copy available on another host and once I recreate the Disk Group both copies will be available again, right?
"Ensure availability" could be an option, but keep in mind the side effects.
While you would speed up the phase until the first node enters MM, after that node reappears with it's new DG design it wouldn't contain any data.
So when you than try to bring the next node into MM, the probability that vSAN has to copy data increases and therefor that node might need more time to enter MM.
Since it is a cluster wide activity on all nodes (Deleting DG and recreating it), you may not save much time. May be the first node goes to MM quickly but not the other nodes. Because, when you delete a disk group on a node, all the data will be erased and when you put second node in MM , VSAN has to ensure all the objects are complying to RAID 1 policy for that it has to rebuild many objects at least the object which was residing (currently not available after DG recreation) on node1 has to be rebuilt, so eventually it is time-consuming task.
For this scenario, it is always recommended to go with full data migration as that is the safest method then using ensure accessibility mode.