Virtual Machine 'stun' or unresponsiveness during snapshot or maintenance tasks in vSphere ESXi 8.0.x.

Products

• VMware vSphere ESXi

Issue/Introduction

• VM becomes frozen for several seconds (e.g., >4 seconds).
• vmware.log contains: vcpu-0 - CPT: vm was stunned for ######## us.
• Storage performance warnings in vmkernel.log.
• BIOS settings indicate missing P-States or Power Management optimizations.

Environment

• VMware vSphere ESXi 8.0.x
• VCF environment

Cause

The stun process is required to maintain data consistency during snapshot/plugin tasks. Extended durations are typically caused by:

1. High I/O load during the operation.
2. Suboptimal host power management (missing P-States).
3. Outdated host BIOS/firmware.

Resolution

1. Verify host BIOS settings and ensure "OS Controlled" power management is enabled to allow P-States.
2. Update ESXi hosts to the latest patch level to include storage driver optimizations.
3. Perform high-impact operations (backups/upgrades) during low I/O windows.
4. Configure Guest OS SCSI timeouts to 60 seconds to tolerate brief stuns.

Additional Information: For more information on snapshot behavior, see Best practices for using VMware snapshots (KB 318825).