When creating a VSAN policy for erasure coding in a stretched cluster, VSAN datastore is not shown or is incompatible.
The precise error message would be Datastore does not match the current VM policy. Storage Fault has occurred. However, vSAN health was all green, no physical disk or component had failed and there is an adequate number of nodes to satisfy a RAID 5/6.
One of the pre-requisites for an erasure code( RAID 5/6) is to have all the disk components in the VSAN to be set “All flash”. But, Sometimes when deploying the witness node, disks(Capacity disks) will be set as type “HDD,” or the witness node gets deployed on a non-SSD host. In some other scenarios, when the witness node is deployed capacity disk is chosen as HDD.
A fix for this problem is to set all the disk group in the cluster ” All flash”, including that of the witness.
Below steps can be followed:
1) Put witness node in MM
2) unmount witness disk group
3) mark all disks on witness node as All Flash
4) mount the DG back on witness
5) Recreate DG & verify status of disk group as All Flash
Finally, create a RAID 6 vsan storage policy. That should solve the problem
Cloud and Security 