Unity pool full due to excessive preallocated space within luns

hi all,

One of my systems came in a DU situation due to a full pool.

Situation

• Unity XT 480F with advanced dedup active
• Large luns for Hyper-Vcluster shared volumes (can be compared to VMware VMFS)
• Hyper-V VM's are in a process to being move from lun A to lun B
• Hyper-V gives the free space back to the luns (unmap is working).
• During HyperV VM migration from lun A to lun B all VM's on the target lun freezed and paniced due to pool full condition on the Unity
• Source lun shows 4,7TB preallocated space and does not return this space back to the pool fast enough, hence causing a Unity pool-full condition
• All VM writes to fail against the target lun failed and production went offline
• Pool was expanded with a new disk, so now the pool has air to breathe, but the preallocated space does not decline in lun A, it is still around 4,7TB.
• There is not enough free space in the pool to migrate all data from this lun and recreate it. So we are stuck at this time.

Questions: (also asked to support, but reply not very helpful yet at this time)

• How to get the preallocated space back to the pool as quickly as possible
• Why does the system itself not return the preallocated space back to the pool as described? See this document, page 54 where is stated "Each pool may be able to reclaim preallocated space from storage resources if the storage resources are not using the space, and the pool space is running low".

Thanks for any help!

Jan-Pieter