From: Rob Webber (rwebber@xxxxxxxxxxxx)
Date: Mon Feb 05 2001 - 15:09:53 GMT-3
I ran into this very same thing once not long ago. I had ospf demand-circuit
configured on one end, the neighbor and LSA would show up with the "DNA"
flag, but debug dialer packet still showed OSPF Hellos kept bringing the
link up. I think I fixed it by rebooting both boxes...either that or I had
to delete and reconfigure OSPF. It appeared to be some type of bug.
Rob.
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
Chris Mott
Sent: Monday, February 05, 2001 12:42 AM
To: CCIE
Subject: RE: OSPF/IGRP redistribution and DDR dial backup
OK I'm really a dumbass .. must be tired ... putting the access-list that
denied ospf killed the entire backup circuit process, which I must have
understood - prob why I didn't like the idea ... the issue is the IGRP
process is injecting Type 5 LSA'a into OSPF which, I think, in turn is
bringing up the link .. again, I do not like where this is leading ... I
also stepped in it by adding (mistakenly) the D0 interfaces into the OSPF
routing process ... once I removed that, it worked like it should be, but it
is making a call almost every minute or so, prob because of the LSA's from
the redistributed IGRP process ... the literature I have is not full on
this, so if anyone has a link or a page to point me to, I would appreciate
it ... thx for patience!
-----Original Message-----
From: Greg Ferro [mailto:gferro@netstarnetworks.com]
Sent: Sunday, February 04, 2001 4:37 PM
To: Chris Mott
Subject: Re: OSPF/IGRP redistribution and DDR dial backup
Do a "debug ip ospf lsa-generation" and watch what happens. I think you
should see something interesting.
At 03:51 PM 4/02/2001 -0600, you wrote:
>I have a scenario whereby R1 runs OSPF to S0.1 (multipoint, Area 0) and
S0.2
>(p-t-p, Area 33), and IGRP to S1 (FRSW). It also has a dial backup (using
>Dialer Profiles) via ISDN to R2 (which is also connected via it's own S0.1
>to R1 via FR, Area 0). OSPF is tuned and running well, and IGRP/OSPF
>redistribution is also functional - pings fly all around. ISDN is fully
>functional. I would give some ASCII art here, but I have never done so and
>would probably screw it up.
>
>Addressing:
>R1 - S0.1 - 172.16.1.1/24
> S0.2 - 172.16.65.1/24
> S1 - 182.16.1.1/24
> D0 - 172.16.10.2/24
>R2 - S0.1 - 172.16.1.2/24
> D0 - 172.16.10.1/24
>FRSW - S0 - 182.16.1.2/24
>
>I want to implement dial backup of the S0.1 interface. I am using the "ip
>ospf demand-circuit" command, but the issue is the redistribution of IGRP.
>I have passive-interface under IGRP for S0.1, S0.2, D0, and B0 (this just
to
>be sure - prob not needed here). I also have a redistribution filter via
>route-map under OSPF for the IGRP redistribution denying a few IGRP routes
>(which works fine). The Dialer-list has an extended access-list denying
>"ospf any any" and "ip any host 255.255.255.255" (broadcasts) with "permit
>ip any any" at the end. I thought I had covered all the bases.
>
>The jackpot issue is that the circuit keeps coming up and staying up, and
>debug dialer gives me no useful information. It DOES however tell me that
>broadcasts from 182.16.0.0 (the R1-FRSW circuit running IGRP) is
>uninteresting, as well as OSPF multicasts via 224.0.0.5. I cannot tell
what
>is bringing up the circuit, but I feel that it is a redistribution issue.
>
>I have scanned my e-mails, which contain a copy of the entire CCIE lab
>archives for the last three months (big file on my HD, but faster to access
>than the web site archives!), but there is nothing that points to a
>solution.
>
>My 1st (<- caveat here!) lab is in 10 days (Halifax, 14 Feb). I am close
to
>grasping the subtleties that will make or break my chances, and I thought I
>had this issue cold, but alas ... can y'all (I'm from Nashville, remember)
>think of anything I can do or check, or some fundamental I've missed?
>
>I've enjoyed reading the various posts, and I learn something every day
from
>all of you. Mainly I'm learning that I do not know what I thought I did,
>but that's the idea. Congrats to all new #'s, and I hope to join your
ranks
>soon.
>
>TIA
>
>Chris Mott
>CCDP, CCNP+Voice
>Solarcom,LLC www.solarcom.net
>
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:28:37 GMT-3