Hey,

we are using Storage Pods/Datastore Clusters. We have Compute profiles with two storage volumes, one small for the operating system, and one big for the data.

VM creation sometimes fails with this error message:

"Failed to create a compute NJA VSphere (VMware) instance tsseodb98nja.example.com: NoDiskSpace: Insufficient disk space on datastore 'va5_esx_t1_04'."

this happens even if some of the members of the storage pod have enough resources left.

so it seems like only the size of the first volume is used to select the storage device from a storage pod, and the selection is not repeated for the second device.