06-05-2015 05:32 PM - edited 03-05-2019 01:37 AM
I have a router which learns a subnet via ibgp (preferred) but also has a floating static route with AD 220 (static route is not redistributed to BGP). When primary path fails, floating static route gets installed on routing table as expected. But when primary path recovers, ibgp route does not get installed on routing table unless route is manually cleared (clear ip bgp x.x.x.x). Once manually cleared, ibgp route for that subnet gets installed on the routing table.
Any input is appreciated.
06-06-2015 05:59 AM
Hi,
you are saying "When primary path fails, floating static route gets installed on routing table as expected."
Isn't it also redistributed to BGP at that moment?
If yes, it gets the BGP weight attribute value 32768 by default and when the same prefix comes via iBGP again, the redistributed one is still preferred within BGP which prevents the iBGP one to get into the routing table.
If my idea is correct, you need to set the weight attribute to 0 when redistributing the static route to BGP. Depending on other attributes the iBGP prefix might win then. If not, you might still need to change other attributes, to increase the Local Preference for the iBGP prefix, e.g.
Best regards,
Milan
06-07-2015 05:25 AM
Hello
I am wondering if bgp synchronisation has anything to do with this.
Is this subnet learned in Ibgp is an external route?
do you have full NLRI between the Ibgp sessions and synchronisation turned off.
sh IP bgp summary
sh ip bgp
sh IP route bgp
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide