Determining Progress of Long running Snapshot Operations and Cancelling Them

Products

• VMware vSphere ESXi

Issue/Introduction

Long running snapshot delete or consolidation operations can impact the use of virtual machines. In addition, these operations may take longer to complete if the virtual machine is powered on and is issuing many writes per second to its virtual disks. If such writes are occurring, the deletion/consolidation may be better done when the virtual machine is less active.

VCF 9 introduces three mechanisms that can help you mitigate these issues. These mechanisms are only available if the virtual machine is registered to a ESX 9 or later host.

Environment

ESX 9.0

Resolution

VCF 9 introduces progress and completion time reporting for snapshot deletions/consolidations, the ability to cancel snapshot/delete operations, and the ability to resume a canceled offline snapshot consolidation. The following provides more details about these three mechanisms. 

vCenter reports the progress and estimated completion time for snapshot delete and virtual disk consolidation tasks. Use this information to determine the estimated time remaining. If the VM is actively writing to its disks, you can access the impact these writes are having by comparing the progress during high write I/Ops and low I/Ops. 

If you do not want to wait for the operation to finish, you may cancel it. 

You may cancel an in-progress snapshot delete operation. If you do, the operation will be aborted, and you will need to retry later. You could later either invoke the consolidate operation or delete all the snapshots.

You may also cancel a disk consolidation. If the VM is powered on, the operation is aborted, and if it is powered off, the operation is suspended. A suspended consolidation can be resumed later from the point at which the consolidation was suspended. Resumable offline consolidations can be used to consolidate large snapshots across multiple maintenance windows. 

Note, while canceling a snapshot deletion or consolidation allows you to retry/continue the operation later, there are three disadvantages to consider. First, until the deletion and consolidation is completed, the storage used by the to-be-deleted snapshots will not be entirely freed. Second, if you are using VMFS SEsparse, NFS without VAAI, and vsanSparse snapshots, until the deletion or consolidation is completed, the virtual machine read performance could be impacted. Generally, the impact on virtual-disk reads is greater for VMFS seSparse and NFS without VAAI than other storage types. vSANsparse has much lower read overhead due to the use of a metadata cache that optimizes reads. vVol, NFS with VAAI, and vSAN ESA all offer native snapshots and don’t incur a significant read penalty. Finally, again, if non native snapshots are used and your intent is to consolidate the running point of the VM, delaying the consolidation may increase the amount of data that must be consolidated in the end. Until all snapshots are deleted, any writes issued by the VM will be to a snapshot that itself needs to be consolidated. If you are not using native snapshots, additional time may be required to complete this consolidation.