[quagga-users 13332] Re: Q: deleting routes from the zebra_rib

Charlet, Ricky ricky.charlet at hp.com
Fri Jul 12 19:22:57 BST 2013

	Sorry for going silent for a couple of days. I'm gonna get back to trying some suggestions by Christian and hopefully post new info soon...

Ricky Charlet
Software Dev / Routing Dude: Aries team, Roseville CA
ricky.charlet at hp.com
USA: 916.785.2090

-----Original Message-----
From: Christian Franke [mailto:nobody at nowhere.ws] 
Sent: Wednesday, July 10, 2013 1:15 PM
To: Charlet, Ricky
Cc: quagga-users at lists.quagga.net
Subject: Re: [quagga-users 13318] Q: deleting routes from the zebra_rib


As far as my understanding of Zebra's RIB model goes, it is behaving as expected here. It will not remove anything on its own from the RIB, that is up to the protocol that installed it there. Whenever there's an update for a prefix, Zebra will reevaluate which route is the one with the smallest administrative distance and make sure that the FIB is updated to contain that route. (There might be exceptions to that when there are routes that are unusable e.g. because the nexthop is currently unreachable, but that's the general idea) This is what happens in the trace you attached. Zebra notices that there is a connected route and removes the RIP route from the FIB.

So I would say the question here is more about RIPd then about Zebra.

The following paragraph should be taken with some care as I am not that familiar with the operation of RIPd and I cannot judge on whether RIPd's bevaviour is correct or not.

My best bet would be that why it doesn't behave as you expect is that the interface where you have the prefix connected is not active in RIP and therefore RIPd doesn't know about the connected prefix. Can you check that RIPd does actually run on the interface where the prefix got connected or that it is configured in a way that it should redistribute that route? Enabling debugging in RIPd might give some information.


More information about the Quagga-users mailing list