Right - this has been solved - Many thanks to Stephen Wagner and his brilliant site :
This is what I found.
ESXi 6.7 does not like it if you change the configuration of a virtual pool after it has been presented (presented and mounted by ESXi)
For example, if you are using Virtual Disk Pools which contain volumes that are presented to ESX 6.7 as Datastores and at a later date you change that pool by adding a disk group to it or a read cache to it - you will probably come across the problem that I have doucmented here (datastore will not mount on reboot of esxi 6.7 after virtual disk pool changes)
The only way you can avoid this problem is if all your disks have the same Sector Format. All the disks in the MSA should be 512n or 512e (is there any other format?)
Remember this only applies to ESXi 6.7 (In my testing 6.5 is not affected)
