SRM .psf File Capacity Usage Discrepancy on vSAN
Article ID: 375312
Updated On:
Products
Issue/Introduction
vSAN skyline capacity display screen shows psf file (vSphere Replication persistent state file) capacity value for "used capacity" is equal to the size of the base vmdk even though the actual usage is very less
Environment
vSphere vSAN 7.x
vSphere vSAN 8.x
Cause
The display in vCenter is correct as vCenter is picking up the total "thick provisioned" value. When using a VMware vSAN datastore, once the VM is selected for Site Recovery Manager protection, the PSF files are created per vmdk by vSphere Replication. For VMware vSAN datastores, the PSF file provisioned type depends on the configured VM namespace storage policy. If the namespace storage policy is set to Thick Provisioning, the PSF files are required and cannot be removed.
Example output:
structtype:
ObjectInfo
Health:
healthy
Object UUID:
e669b361-8a91-7200-1ced-e4434b931380 <---------------
Version:
15
Owner:
esxi-003.server.test.net
Policy:
stripeWidth: 1
cacheReservation: 0
proportionalCapacity: 100 <---------------- meaning all of the capacity assigned to the object is allocated (thick)
hostFailuresToTolerate: 1
forceProvisioning: 1
spbmProfileId: c9eb54d4-c849-4af0-b7b6-7a06c417f9f1
spbmProfileGenerationNumber: 0
replicaPreference: Performance
iopsLimit: 0
checksumDisabled: 0
CSN: 20183
SCSN: 17893
spbmProfileName: Thick_Policy <--------------------
Resolution
This behavior is changed in Esxi 9.0 onwards, vSphere Replication uses PSF/HBR PERSIST files/objects to store the replication state. Depending on the vSAN policy used, either thin or thick provisioned HBRPERSIST objects are created, regardless of the fact that hbr_fiter always specifies 'thin' allocation type
Additional Information
Using vSphere Replication Optimized Reprotect:
Feedback
