Identifying Normal vs. Excessive Snapshot Stun Times in Virtual Machines

Products

• VMware vSAN

Issue/Introduction

Virtual machine snapshot operations can cause brief pauses in execution, commonly referred to as "stun time." Distinguishing between expected operational stun times and those indicative of underlying storage or application issues is critical for troubleshooting performance degradation during backup operations.

Environment

ESXi 8.0

Resolution

1. Collect the vmware.log file from the affected virtual machine directory.
2. Search the log for the "Checkpoint_Unstun" entry to identify the specific stun duration.
   • Example: vcpu-0 - Checkpoint_Unstun: vm stopped for 680000 us (0.68 seconds)
3. Compare the recorded stun time against known operational thresholds.
   • Stun times under 1 second are generally considered within normal operational parameters for snapshot commit or consolidation on standard storage backends.
4. If stun times are within the expected threshold but application-level timeouts persist, review the following:
   • Application-layer timeout configurations (e.g., database connection settings).
   • Backup software proxy performance and throughput logs.
   • Storage array latency metrics using esxtop during the backup window.
5. If stun times consistently exceed expected thresholds, investigate the storage subsystem for I/O latency or high contention as documented in the VMware Storage Performance Troubleshooting Guide.