Snapshot Operation limitations directly correlating to number of Supported VMDKs

Products

• VMware vSphere ESXi

Issue/Introduction

Snapshot Operation limitations directly correlating to number of Supported VMDKs

Symptoms:
Remote Console to Guest Operating System such as RDP or SSH session disconnects during Snapshot creation operation.
High number of Network ping drops are also observed.

Environment

VMware vSphere ESXi 6.7

VMware vSphere ESXi 7.0

Cause

The snapshot workflow is linear in nature. The unstun values to complete Snapshot creation operation directly correlates with number of VMDKs associated to a Virtual Machine. 

For Example: If the unstun values stays at 0.xx seconds, then multiplying with number of VMDKs sums up the total Checkpoint_Unstun required to complete Snapshot creation operation.

Virtual Machine "Test" has 249 VMDKs and has a unstun values of 0.55 seconds. This leads Checkpoint_Unstun to be 132490356 us or 132.5 secs!" This might not be an acceptable time frame by session management set at Operating System or at resident application.

Resolution

Connect with your Application or Guest Operating System vendor to know more about set limitations in regards to Image Based Snapshot Operations. Follow the best practice guidelines suggested, before implying them in a Production Environment.

Workaround:
File-based backup solution.
Agent based backup solution

Reducing the number of VMDKs per Virtual Machine to an acceptable unstun time frame.

Additional Information

Impact/Risks:
Impacts users connected to an application due to larger unstun time frame taken to complete the Snapshot Operation.