We are migrating from an NFS SAN to an iSCSI SAN and i'm noticing that when migrating VMs the swap file is reporting full RAM size provisioned on the VMFS datastore but on the NFS datastore is reporting 0 bytes and i'm getting low space alerts as a result. We are not even close to being over committed on RAM and the hosts are in the same cluster, VMs remaining in the same resource pool. If I browse both datastores I can see that the underlying file is getting created (or reporting file size anyway) as full RAM size on both.
For example, a VM with 2GB RAM on a host with the swap datastore set to the NFS share creates a "2GB" file but reports 0 bytes provisioned, but when the VM is migrated to a host with the VMFS datastore, the swap file is getting created as 2GB but reporting 2GB provisioned. Is this just an artifact of how the different types of datastore are handled and should I just ignore the alerts or is there somehow a different problem at play? I know if the memory reservation is getting met (and it is), the swap file "should" be 0 bytes, which is why i'm confused. I'm running 5.5U1 with VMFS5 datastores FWIW