After BGP peer configuration, active and standby edge nodes continuously failover

Products

• VMware NSX

Issue/Introduction

After configuration of a BGP neighbor, edge nodes hosting the active and standby Service Router (SR) of the configured Tier-0 will continuously switch-over.

Environment

VMware NSX-T Data Center
VMware NSX

Cause

NSX-T BGP stack expects the BGP keep alive timer to less or equal to one third (1/3) of the BGP Hold timer value.

Resolution

To workaround the issue, configure keep alive and hold timers to satisfy the below relation.

1. Valid Configuration: Keep Alive Timer value 99, Hold Timer Value 300
2. Valid Configuration: Keep Alive Timer value 100, Hold Timer Value 300
3. Invalid configuration: Keep Alive Timer value 101, Hold Timer Value 300

(1) and (2) satisfy the relation BGP keep alive timer value <= 1/3 BGP Hold timer value, where as (3) doesn’t satisfy the relation.