Bulk Migration replication queries

Products

• VMware HCX

Issue/Introduction

This article addresses common questions about optimizing HCX Bulk Migration switchover times by managing VM state and services prior to the switchover process.

Questions and Answers

Are there any benefits to powering off the VMs a few hours prior to the switchover?

Yes, there are benefits to reducing VM activity prior to switchover. Powering off the VM or stopping services within the VM prior to the switchover process would result in less delta data being created. This results in:

• Less final sync data to be processed during the switchover phase
• Shorter switchover time frame
• Reduced migration downtime window

Recommended Approach: Instead of powering off the entire VM, we recommend stopping high-churn services within the guest operating system. This approach:

• Reduces data change rate without disrupting VMware Tools functionality
• Maintains VM manageability and monitoring capabilities
• Prevents potential issues with VMware Tools heartbeat and guest operations

Important: Do NOT stop VMware Tools services - only stop application services that are causing heavy disk I/O or data churn within the VM itself (such as databases, web servers, application servers, batch processing jobs, etc.).

Would we be able to power off the VMs and "time it" before the next RPO cycle?

Yes, it is possible to power off the VM or stop services in between RPO cycles, which occur every two hours. The timing considerations are:

• Monitor the last successful RPO sync completion time
• Plan the power off or service shutdown after a successful RPO sync completes
• This provides up to 2 hours before the next RPO cycle attempts to run
• The switchover can then be initiated without waiting for another full RPO cycle

Would powering off the VMs prior to the switchover cause any issues when performing the switchover?

Powering off the VM prior to switchover should not cause any issues with performing the switchover itself. However, consider these factors:

If choosing to power off the entire VM:

• Ensure the VM is powered off gracefully through the guest OS
• Enable "Force Power Off" option in the migration settings as a fallback
• Be aware that HCX will still attempt to power off the VM during switchover, which could potentially cause a migration error or warning if the VM is already powered off. This is another reason why we recommend stopping services instead of powering off the entire VM

If choosing to stop services (recommended):

• Document which services were stopped for post-migration restart
• Ensure critical monitoring and management services remain running
• Keep VMware Tools services running to maintain guest-host communication
• This approach avoids potential power state conflicts during the switchover process

Resolution

For detailed information about HCX Bulk Migration operations and best practices, see HCX Bulk Migration operations and best practices.

Best Practices:

1. Service Management Approach (Recommended):
   • Stop high-churn application services within the guest OS several hours before switchover
   • Keep VMware Tools services running
   • Monitor the reduction in disk I/O and network activity
   • Schedule switchover when data change rate is minimal
   • This approach avoids potential power state errors during migration
2. Timing Considerations:
   • Review RPO cycle timing (every 2 hours)
   • Plan service stops or VM power off after a successful RPO completion
   • Allow sufficient time for any running transactions to complete
3. Migration Planning:
   • Enable "Force Power Off" option as a safety measure
   • Document all stopped services for post-migration verification
   • Consider business impact of stopping services early versus longer switchover time