vMotion connectivity loss on vSphere Distributed Switch with 'Route based on IP hash' load-balancing policy

Products

• VMware vCenter Server

Issue/Introduction

• Virtual machines (VMs) experience significant network connectivity loss, lasting approximately 10 to 13 minutes, immediately following a successful vMotion migration within a vSphere cluster.
• This issue occurs specifically on virtual machines connected to a vSphere Distributed Switch (vDS) port group.
• Other VMs in the same cluster connected to different port groups and uplinks remain unaffected.
• This symptom can be verified by observing continuous ping packet loss to the VM's IP address following the migration.

Environment

VMware vCenter Server

Cause

The root cause is a Distributed Port Group (dvPG) configuration mismatch. The dvPG is configured with the Route based on IP hash load-balancing policy; however, the corresponding physical switch ports are not configured with an active LACP or EtherChannel group for the associated VLAN.

When a VM migrates to a destination host, the physical switch drops traffic until the MAC address table updates or the traffic hashing shifts to match the switch port configuration.

Resolution

To resolve this connectivity delay, update the teaming and failover policy to match the physical switch port configuration.

Prerequisites:

• Take a snapshot or backup of the vCenter Server.
• Ensure you have administrative privileges in the vSphere Client.

Steps:

1. Log in to the vSphere Client.
2. Navigate to the Networking view.
3. Expand the Datacenter and locate the vSphere Distributed Switch.
4. Right-click the affected Distributed Port Group and select Edit Settings.
5. In the Edit Settings dialog, select Teaming and failover from the navigation menu.
6. Locate the Load balancing policy.
7. Change the policy from Route based on IP hash to Route based on originating virtual port.
8. Click OK to apply the configuration.
9. Perform a test vMotion of a virtual machine to verify that network connectivity is restored immediately upon migration.

Additional Information

Additional Information:

• If the issue persists, verify physical switch uplink configurations for the affected VLAN.
• For general network redundancy best practices, refer to Best Practices for Configuring Network Redundancy in VMware vSphere and NSX.
• For software downloads and release information, visit the Broadcom Support Portal.