Of course as soon as I ask about it, I find it.

In VMware you can run this command:

esxcli storage vmfs unmap -l MyDatastore

Article: Using the esxcli storage vmfs unmap command to reclaim VMFS deleted blocks on thin-provisioned LUNs (2057513) | VMware K…

start a space reclaim from the VMware side

if blocks are reclaimed on the source with an ESX reclaim and the LUNs are in a replication session, should the expectation on the target LUN be that they will be reduced as well? Or does the unmap need to be run on the target array as well?

thanks!

When you use UNMAP, you're removing slices from the Pool LUN, but you're not changing the configured capacity of the LUN. While I can not find a definitive statement, I believe that you should not have to do anything to the mirrored LUN on the other array. As the LUN is allocated based on the configured capacity, it should adjust automatically. But I'm still trying to find any documentation to for that.

glen

Hi,

I have reached this post searching on the internet. I have this problem, and I have opened a SR and the support has told me that the replica does not release the space.

In order to release, you need to do a failover, present the replica to a host and do a UNMAP :-(

Regards,

Enrique

try if you can do a LUN migration on your "replica"

Good point :-) but it is not possible:

"The Datastore is participating in a replication session as a destination resource. The Datastore cannot be moved. Delete the replication session and rename or delete the existing destination resource before reconfiguring replication with a new destination Datastore."