From: Earl Aboytes (earl@xxxxxxxxxxxx)
Date: Thu Jul 13 2000 - 02:57:31 GMT-3
Are there any other routing protocols running on this router? I would bet
something like RIP or IGRP is running there.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Earl Aboytes
Senior Technical Conultant
GTE Managed Solutions
805-381-8817
earl.aboytes@telops.gte.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of Vijay
Venkatesh
Sent: Wednesday, July 12, 2000 10:38 PM
To: Keith Kruepke
Cc: ccielab@groupstudy.com
Subject: Re: Painful OSPF demand circuit
Hi Keith,
Thank you for the quick response. Here are a few items
indented -
> As far as your configuration, based on your description, it sounds like
you are right on. You should NOT exclude the OSPF traffic from the dialer
list, as the interface should still come up for a topology change. All I
did was use 'ip ospf demand-circuit' on one side of the connection, and it
worked.
>
> To get an idea of when the traffic is being sent, try 'debug dialer
packet'. Yes, the 'show dialer' command will give you the info about the
packet that brought up the line, but the debug command will show all
packets, even when the line is up. It may help to figure out how often the
traffic is being sent. Also, debug both sides, to see if traffic is coming
from both.
****
The debug dialer packet gives no output. However I see that the call
diconnects from one channel and the same router redials out using
another channel. I did a debug on ip ospf pkt and saw nothing.I also
debugged both sides. The ip ospf int bri 0 shows that this is an
on-demand circuit and the ospf DB shows that there are DNAs. However,
when you do a sh ip ospf nei the dead timer is counting down. I have
got this working before at the office but I am not sure if it is the
IOS or
just me being rusty.
*****
> I would try 'show ip ospf interface', 'show ip ospf database', and 'show
ip ospf neighbor'. The interface on BOTH sides should say that it is
running as a demand circuit. The database should show the remote Router LSA
as being DNA (Do Not Age). The neighbor should have a blank dead time. If
any of these conditions are not met, there is a chance that the routers are
not recognizing the demand circuit.
*****
Then I think the routers are not recognizing on-demand circuits. I am
curious what exact version of the IOS are you using. I have a 8M flash
any recommendations on an IOS that works. AT this point I am willing
to try anything.
****
> I looked up when demand circuit support was added to the IOS, and it is
reported as of 11.2. Since you are running an 11.2, I am wondering if there
could be a problem with that implementation of it... I know this is a
cop-out response, but if possible, an IOS upgrade would not be a bad idea.
>
****
Yes, I think it is the IOS playing tricks again.
****
> Finally, just to make sure...you mentioned that both routers have the same
router ID. I am assuming you mean the process ID (router ospf X), which
does not have to match but does no harm. If you do actually mean router
IDs, that would be a problem, because router IDs must be unique. (This is
the highest IP address on a loopback or the highest IP on a physical
interface if no loopbacks exist.) If you do 'show ip ospf', the local
router ID should be indicated at the top. These should be unique.
****
My humble apologies for the confusion. The router ids are indeed diff.
The ospf process id is the same.
>
> Well, you may have tried many or all of the ideas here, but hopefully
something I said will help you find the problem. Good luck.
****
You most certainly did. If you could email the exact verson of the IOS
or even the IOS itself I will be much obliged. I have both routers
with
8M flashes. So hopefully we can find one that fits without going
through any compress routines. Thank you.
Regards,
Vijay.
> ----- Original Message -----
> From: "Vijay Venkatesh" <vijay.venkatesh@usa.net>
> To: "Earl Aboytes" <earl@linkline.com>
> Cc: "Stephens, Paul [Prof.Serv]" <Paul.Andrew.Stephens@compaq.com>;
<ccielab@groupstudy.com>
> Sent: Wednesday, July 12, 2000 11:09 PM
> Subject: Painful OSPF demand circuit
>
> Hello,
> I am beating my head against a brick wall trying to get the ospf
> demand circuit working. I am posting this in the hope someone will
> look
> at it from a different perspective. This is my situation -
>
> R1 is connected to R2 via a BRI interface. The Bris are on one
> subnet and are configured with spids for both channels and do not have
> dialer profiles. However, they have dialer map statements. This is a
> point to point bri. Nothing hangs of R1 and nothing hangs off R2. I am
> running both routers with the same ospf router id and both are in area
> 0. Now, when I bring up the bri and configure 'ip ospf demand-circuit'
> on side (either R1 or R2) the bri line comes and exchanges ospf info
> across the isdn line. Then it hangs up after the idle-timer of 60
> seconds. Either R1 or R2 keep redialing and keep the BRI up. Please
> not there are no topological changes that are causing this to happen.
> When I do a 'sh dialer' it keeps saying that one of the routers
> initiated the dial because it had to communicate to 224.0.0.5 which
> is the ospf multicast address. What do I do to make sure that the
> line stays down using nothing but the 'ip ospf demand-circuit'
> statement. My dialer list is protocol ip permit. The dialer maps
> are configured with the broadcast option. The version of code I am
> running is 11.2(19a) on both routers. Any suggestions are invited.
> Thank you for your insights and time.
>
> Regards,
> Vijay.
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:23:53 GMT-3