Date: Thu, 15 Aug 1996 09:14:15 -0700 From: bmah@cs.berkeley.edu (Bruce A. Mah) To: Jaye Mathisen <mrcpu@cdsnet.net> Cc: hackers@freebsd.org Subject: Re: New routed weirdness. Message-ID: <199608151614.JAA32056@premise.CS.Berkeley.EDU> In-Reply-To: Your message of "Wed, 14 Aug 1996 18:17:12 PDT." <Pine.NEB.3.95.960814180648.708b-100000@mail.cdsnet.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Jaye Mathisen writes: > The PM's run RIP V1. But if I read this correctly, somehow I'm picking up > RIP packets where the host is advertised as having a netmask of 29 bits or > 31 bits, rather than 32 bits. > > However, I don't see how this is possible. I guess I need some > suggestions on how to track down specifically what's wrong. Jaye-- My memory is a bit hazy, but I don't think this is a FreeBSD problem. I seem to remember that Livingston PortMasters can aggregate route advertisements for their PPP/SLIP dialups. Instead of advertising a large number of host routes, they can send out an advertisement for a subnet prefix that contains all of the IP addresses for their set of PPP/SLIP lines. This option is configurable on the PM. The catch is that RIP V1 doesn't have any way to express the length of this routing prefix (RIP V2 does), so the PM actually ends up generating funny routing updates. How these gets interpreted seems to be implementation-dependent. This happened at an ISP I helped set up, and it confused the heck out of a Solaris fileserver. Eventually we decided to turn off aggregation (not Livingston's term for this concept but you get my drift). My suggestion would be to do a tcpdump for routing updates and look carefully at the IP address in each advertisement. You might need to reconfigure the PMs to advertise individual host routes for the dialups (buys you correctness at the expense of larger routing updates). Hope this helps... Bruce.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199608151614.JAA32056>