Bird Users,

BIRD version 1.2.5
Uptime > 3 years, nice work guys!

Two identical servers running BIRD with identical config.

I ran into some routes sticking around in some of my route server tables from a 
neighbor that was down.  I did some testing and I had some success reloading 
the pipe protocols, but not 100%.  Even with that neighbor and all it's pipes 
completely removed from the config the routes remained, originating from the 
de-configured bgp peer.  

During my testing I decided to do a mass reload of all the pipes to/from that 
down neighbor and that actually crashed the bird process, goodbye uptime.  The 
crash did clear up the stale routes of course.  I decided to just go ahead and 
restart my secondary route server as well to clean up the routes there.

I am writing to see if others have had this experience with stale routes.  My 
version is really old, could this be a bug that has not surfaced?  Could the 
long uptime have lead to weirdness?  Lastly, is there a configuration issue 
that could lead to this?

I am not seeking a definitive response, or a dive into my configs, as there 
seem to be a lot of factors that might never point to a cause, just looking to 
see if others have had a similar experience.  I just did a preliminary test and 
the configs work fine on 1.4.3 so the plan for now is to simply upgrade in the 
hopes this doesn't resurface.

-Stephen

Reply via email to