Reduced replication throughput due to network forwarding path aging

Products

• VMware Cloud Director
• VMware Live Recovery

Issue/Introduction

• Sustained throughput reduction (e.g., 2–3 MB/s per VM) regardless of available WAN capacity.
• Widespread RPO violations occurring simultaneously across different tenants and VM workloads.
• Throughput spikes and dips appear at regular, synchronized intervals (e.g., every 1–5 minutes) across all active replications.
• High variance in replication traffic metrics, preventing the maintenance of a stable baseline.

Environment

VMware Cloud Director Availability  4.7.x

Cause

This issue occurs when intermediate network devices along the replication path age out forwarding entries for replication connections during operation.

Resolution

This issue stems from the underlying network environment rather than VMware Cloud Director Availability, investigate and adjust the network infrastructure configuration along the replication path:

• Implement an automated connectivity probe (such as a TCP ping or heartbeat) between the source and destination endpoints (e.g., every 30 seconds) to prevent intermediate network forwarding tables from aging out the connection.

• Ensure MTU configuration is consistent across all intermediate network devices along the path (e.g., standard 1500 bytes).

• Confirm that intermediate devices are not performing unexpected Deep Packet Inspection (DPI) or payload scanning on the replication traffic.

Additional Information

• For general troubleshooting of replication performance, refer to the VMware Cloud Director Availability Documentation.
• For assistance with log retrieval, see the article Retrieving Logs in VMware Cloud Director Availability.