Hi BIRD team,

I hope you're doing well.

While investigating BGP route state behavior, we came across a potential bug in BIRD 3.3.2.


We observed that the kernel protocol can install a route with a higher total AIGP metric as an ECMP next hop, even though BIRD selects the route with the lower total AIGP metric.


I've attached
report.md and attachments.zip with more details.


Please let us know if this is expected behavior or if we've overlooked anything in our analysis.


Thanks for your time!


Best regards,

Yann