What does HRESULT 0xC05CFF04 (ERROR_SVHDX_UNIT_ATTENTION_RESERVATIONS_RELEASED) mean?

 
Could be also:
ConstantTypeOS
STATUS_SVHDX_UNIT_ATTENTION_RESERVATIONS_RELEASEDNTSTATUSWindows
Previous Next
ERROR_SVHDX_UNIT_ATTENTION_RESERVATIONS_PREEMPTED ERROR_SVHDX_UNIT_ATTENTION_REGISTRATIONS_PREEMPTED

ERROR_SVHDX_UNIT_ATTENTION_RESERVATIONS_RELEASED

The reservation was released, not necessarily preempted

ERROR_SVHDX_UNIT_ATTENTION_RESERVATIONS_RELEASED reports a Unit Attention condition caused by release of this initiator’s persistent reservation. A release removes the reservation that controlled access to the logical unit. It is different from a preempt event: release is normally an explicit state transition, whereas preemption transfers or removes rights through another initiator’s persistent-reservation action.

Because registration and reservation are separate parts of the SCSI Persistent Reservation model, a released reservation does not by itself prove that all registrations were removed. The current state must be read from the device or virtual-disk service. An initiator that assumes its old access mode remains valid can receive subsequent reservation conflicts or can make unsafe recovery decisions.

Where this can arise

  • A cluster resource was intentionally taken offline, moved, or failed over.
  • A reservation-management path released ownership during cleanup after a previous error.
  • A service restarted and rebuilt its shared-disk state, leaving a former client with stale assumptions.
  • A control-plane action changed the reservation without synchronizing all consumers.

Safe response

  • Query the current registrations and reservation holder before issuing I/O that relies on exclusive or registrant-only access.
  • Correlate the event with cluster resource state, host migration, maintenance actions, and guest shutdown or restart events.
  • Do not attempt to recreate the reservation from an arbitrary node. Use the cluster’s documented ownership and fencing workflow.
  • When the release was unplanned, preserve logs from the client, virtual-disk server, and cluster layer before performing corrective actions.

References

Looking for a different code? Search another status or error code.