HyperMetro / Replication Paused / scsi not ready on replicated lun [Dorado 5000 V6 / 6.0.0.SPH1]

Hi,

Let's assume the following hypermetro setup :

* Linux node1 in datacenter1 issuing r/w iops on dorado1 preferred lun R1

* node1 put scsi3 persistent group reservation on each path before importing lvm vg on disk

* Linux node2 in datacenter2 doing nothing, and have access to replicated lun R2

* Alua enabled

* Each server is zoned/mapped to both arrays (preferred paths to datacenter1, non preferred to datacenter2)

Of course I understand that, as soon as replication is paused, write io must be blocked on the non preferred paths to avoid corruption.

The issue I have is that server node1 is not able to query/remove/set scsi reservation status on R2 (non preferred) paths.

If I want to deport my lvm vg, I can only remove scsi3 pgr in datacenter1, but not in datacenter2.

I think that putting the non preferred paths in "read-only" instead of "not-ready" would resolve the problem.

Do you know if it's possible ?

Thanks for reading !