From: Simon Baxter (Simon.Baxter@xxxxxxxxxxxxxx)
Date: Wed Dec 06 2000 - 21:59:12 GMT-3
How would you do that?
-----Original Message-----
From: gferro@netstarnetworks.com [mailto:gferro@netstarnetworks.com]
Sent: Thursday, December 07, 2000 11:01 AM
To: Simon Baxter; Justin Menga
Cc: ccielab@groupstudy.com
Subject: RE: OSPF Demand Circuit
This does not work for OSPF. A distribute list for OSPF does not stop the
formation of neighbor adjacency and transmission of the link state
database. What it does do is stop routes from populating the ip route
table. (Note the landmine here). SO the link state updates are still being
propagated and the demand circuit will still come up to ensure that the
database is consistent for all routers in the area.
You need to ensure that OSPF route going into IGRP do not return from IGRP
to OSPF (whereupon they become Type5 External LSA's). Look in the ospf
database and you'll see them there.
At 09:22 AM 7/12/2000, Simon Baxter wrote:
>But aren't you saying the problem is in the IGRP-OSPF direction?
>
>how would
>
>router os 10
>redist igrp 10 subn
>distribute-list 100 in igrp 10
>!
>access 100 per ip any x.x.x.x m.m.m.m
>
>
>Doesn't "Solution: filter IGRP to OSPF redist to ensure no OSPF routes are
>coming
>back from the IGRP area." imply you need to stop routes going back in
>mutual redistribution?
>
>Simon
>
>-----Original Message-----
>From: gferro@netstarnetworks.com [mailto:gferro@netstarnetworks.com]
>Sent: Thursday, December 07, 2000 12:49 AM
>To: Justin Menga
>Cc: ccielab@groupstudy.com
>Subject: Re: OSPF Demand Circuit
>
>
>IGRP is putting routes into the OSPF process as Type 5 External LSA summary
>routes. Every time IGRP updates, it updates the routes vs the OSPF internal
>routes, OSPF database sequence number increments, ospf on demand comes up
>to synchronise.
>
>to see this debug ip os lsa (as other have said).
>
>Solution: filter IGRP to OSPF redist to ensure no OSPF routes are coming
>back from the IGRP area.
>
>
>
>At 08:15 PM 6/12/2000, you wrote:
> >Hi,
> >
> >I am having real trouble with these.......
> >
> >If I have a point-to-point link, and on one side I have dialer map
> >statements and on the other side I do NOT have dialer map statements
(i.e.
> >dialer string xxxxxx), what I have found:
> >
> >1. If I leave the network type as POINT_TO_POINT I must include the
> >broadcast keyword in the dialer map statement, otherwise an adjacency
never
> >forms - this is logical as point-to-point uses multicast.
> >
> >2. If I include the broadcast keyword in the dialer map statement and
have
> >the dialer interface configured as ip ospf demand-circuit, HELLO's are
>still
> >sent over the dialer interface - however NOT while the line is up, only
>once
> >it disconnects.
> >
> >3. Demand-circuit doesn't work if you use NON_BROADCAST or BROADCAST -
> >Hellos continue to be sent - I thought this might get around using the
> >broadcast in the dialer map (at least for NON_BROADCAST).
> >
> >This 'seems' to be a simple technology, but it won't work properly for
me!!
> >Any ideas as to why OSPF multicasts to BRI0 once the line goes down??
> >
> >I HAVE JUST DONE SOME MORE TESTING AND FOUND THAT THIS BEHAVIOUR ONLY
>OCCURS
> >IF YOU ARE REDISTRIBUTING INTO OSPF. ONCE THE LINE GOES DOWN, THE ROUTER
> >REDISTRIBUTING BRINGS UP THE LINE. THE REASON FOR CALL IS AN OSPF
>MULTICAST
> >TO 224.0.0.5. IN THE SCENARIO (BOOTCAMP 17), ONE OF THE ROUTERS IS
> >REDISTRIBUTING IGRP. THE OTHER END IS REDISTRIBUTING IS-IS AND DOES NOT
> >BRING UP THE LINE WITH THIS REDISTRIBUTION.
> >
> >IS THIS NORMAL??
> >
> >Regards,
> >
> >Justin Menga MCSE+I CCNP CCSE ASE
> >WAN Specialist
> >Computerland New Zealand
> >PO Box 3631, Auckland
> >DDI: (+64) 9 360 4864 Mobile: (+64) 25 349 599
> >mailto: justin.menga@computerland.co.nz
> >
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:25:59 GMT-3