Modifying storage policy rules on Virtual Machine running on snapshot in vSAN Datastore
search cancel

Modifying storage policy rules on Virtual Machine running on snapshot in vSAN Datastore

book

Article ID: 326578

calendar_today

Updated On:

Products

VMware vSAN

Issue/Introduction

Symptoms:
  • When the spbm policy rule set is changed on a vDisk which has snapshots, the modified policy rules doesn't get applied to the base disk.
  • The newly changed policy gets applied only to the snapshot disk on which VM is currently running in case there are more than one snapshot the other snapshot disk ( object ) will continue with same policy


Environment

VMware vSAN 6.x
VMware vSAN 7.x
VMware vSAN 8.x

Resolution

This is working as per design, upon creating a virtual machine snapshot, separate Virtual SAN objects are created The newly created objects will store the new writes (new content) for each object (delta disk.) 
The benefit of this approach is specific for the delta disks in VIEW environment (For example, for each linked clone in Horizon View) may have a different policy than that of the base disk, as the requirements and workload for each are different

The user should consolidate snapshots and change the policy on the Virtual Machine unless otherwise specified for a specific reason.

Workaround:
  • A virtual machine should not be running on snapshot as the snapshot are only for temporary periods and point-in-time copies of data. The sole purpose of the snapshot is to enable the Administrator to revert back changes in case it does not bring desirable results after applying them (ex. Taking a snapshot of VM before applying a patch inside the guest OS)
  • It is very important to consolidate or delete all snapshots before you make changes in the Storage policy for objects located on a vSAN cluster (VM, VM DISK, Namespace etc - SPBM policy)

Additional Information

If it has been already tried to change the storage policy of a VM that has snapshots, the consolidation task triggered afterwards may fail because the snapshot chain may get compromised. To confirm this is the case and see how to correct it, please refer to: