Oh well, the intervention solved nothing.

New situation is this

Cluster is now only 2 nodes

1 NIC each only for Public, iSCSI and Heartbeat connections (no more NIC Teaming)

MPIO uninstalled on all nodes

Any existing Persistent Reservations cleared on all disks.

The two disks attached to the HP MSA filer still lose ownership around the same time each day and throw 1038 events.

This is starting to annoy me now 😉