We are moving our guests from our current 5.5 cluster to a new, separate 6.5 cluster. The two clusters see the same back-end storage. Our guests typically use 2 datastores, a primary one for C:|/ and a swap datastore. To move a guest, I power off and unregister it from the old cluster, then browse the primary datastore from the new cluster and register the guest. They fail to power on in the new cluster.
Looking at the guest's settings, disks on the primary datastore are fine (showing the human-readable datastore name); disks on the swap datastore are highlighted in red and there is a reference to /vmfs/volumes/<uuid> which is the swap datastore on the old cluster (which does not match the <uuid> for the swap datastore on the new cluster).
We can edit the vmx and change the path to the vmdk to match the new <uuid> of the swap datastore, but this seems kludgey. We could also sVmotion the swap disks to the primary datastore on the old cluster, move the guest to the new cluster and then sVmotion the swap disks to the swap datastore, but again, this is kludgey and if we can avoid it, we'd rather not have swap disks in our array snapshots (the primary motivation for having separate swap disks in the first place). A third option we've ruled out is creating a symlink from the old <uuid> in /vmfs/volumes to the new <uuid>.
Are there any other options? To be clear these are separate clusters (by design) so we cannot easily vmotion the guests. Thanks!
