.Rebooting the Object Gateway no longer sees the `tcmu_rbd_lock_break` and `tcmu_notify_lock_lost` errors During failover/failback and initial device discovery with a multipath setup, `tcmu-runner` may need to take the lock away from another iSCSI gateway. When this happens the initiator may be sending IO to multiple paths at the same time, so multiple gateways will try to take the lock from each other. This can result in errors like: `Could not break lock from $locker_id (Err -16)` These errors are expected temporarily while the failover/failback or device setup process is executing. However, `tcmu-runner` would sometimes not fully clean itself up after losing the lock, so the `-16` error would repeat and the gateway could not be used until it was rebooted. The `tcmu-runner` daemon has been changed to always reopen the device after losing the lock so the internal state is reinitialized before use. With this update to {product}, failover/failback and device set up complete. Lock breaking errors with the error code `-16` should only temporarily be logged to the target logs.