RecoverPoint SCSI Unmap

Two unity 600F arrays (4.1.2.9257522).

RecoverPoint 4.4.1

We moved a bunch of VMs off some data stores that are being replicated (sync with the dynamic option) with RecoverPoint. After, I manually ran the SCSI unmap command from a vm host for the old data stores. (esxcli storage vmfs unmap -l DatastoreName)

This worked as expected. The thin pool on the primary array gained a bunch of free space. But the thin pool on the second array hasn't changed.

It seems RecoverPoint isn't splitting the unmap operations to both LUNs.

Has anyone else seen this behavior? Any ideas how I could reclaim the the free space on the replicated luns? Short of stopping replicating, mounting them and running the unmap command directly on them also.

Also we are moving to VMFS 6, which has native support for the scsi unmap operation. So I would see this same behavior there when that is in place?

Thanks in advance.