From: Nigel Taylor (nigel_taylor@hotmail.com)
Date: Wed Sep 04 2002 - 08:36:30 GMT-3
Khalid,
Well this one escaped me and I must admit I didn't notice the
original post by keith that makes use of OSPF process ID 100, versus
Chris's post that uses OSPF Process ID 1. You note a good point which is
what was discussed by both Peter and I which is referenced by rfc1587.
Peter, also notes that Pat Murphy has a number of nssa related drafts
which attempts to standardize the current implementation of a NSSA area
and the LSA's used to construct it.
Yes, you're correct in what you note that the second ospf process on
keith's R1, may explain why the P-bit is being cleared. However, my
interest in what Chris is observing is very interesting.
Nigel
>From: "Khalid Siddiq" >To: "Nigel Taylor" , >Subject: RE: NSSA and type
7 LSA's -same problem. Configs incl. >Date: Wed, 4 Sep 2002 15:34:29
+0500 > >but i am sure that in first case when KEITH is redistributing
the rip in NSSA area then their is multiple area connected to the NSSA
ASBR, and because of this the p bit is set to 0 by R1 router. > > > > >>
> > > > > > >> > > R1 > > > > >> > > ! > > > > >> > > router ospf 100 > >
> > >> > > redistribute rip subnets > > > > >> > > passive-interface
TokenRing0 > > > > >> > > network 146.10.11.0 0.0.0.255 area 11 > > > >
>> > > network 146.10.21.0 0.0.0.3 area 21 > > > > >> > > default-metric
5 > > > > >> > > area 11 range 146.10.11.0 255.255.255.0 > > > > >> > >
area 21 nssa > > > > >> > > ! > > > > >> > > router rip > > > > >> > >
version 1 > > > > >> > > redistribute ospf 100 metric 10 > > > > >> > >
passive-interface Loopback100 > > > > >> > > passive-interface Serial1 >
> > > >> > > network 146.10.0.0 > > > > >> > > neighbor 146.10.18.8 >
>but in second case when nick redistribute connected route in nssa i dont
know ??? nick what version u are using on that router ? >regards, >khalid
> >-----Original Message----- >From: Nigel Taylor
[mailto:nigel_taylor@hotmail.com] >Sent: Wednesday, September 04, 2002
2:24 PM >To: Khalid Siddiq; ccielab@groupstudy.com >Subject: Re: NSSA and
type 7 LSA's -same problem. Configs incl. > > >Khalid, > The ASBR in
question is connected to a single area as noted by >the config paste by
the original poster. As you pointed out the problem is >trying to
determine why R1 is clearing the P-bit which is the original >question,
which was asked by Peter. > >Just after making my post, Nick Shah, posted
the result of some quick >testing in which case the redistribution of the
connected interface into the >OSPF process did not provide/meet the
necessary for the P-bit being set and >until he made the connected
interface part of the NSSA the type 7's was not >translated into type 5's
> >AS noted in my post Alex Zinin's rationale does seem to provide some
insight >into the problem we're currently discussing and if anything
gives us a >reason for doing further reading on the topic. > >I simply
love this list... :-) > >Nigel > >----- Original Message ----- >From:
"Khalid Siddiq" >To: "Nigel Taylor" ; >Sent: Wednesday, September 04,
2002 5:46 AM >Subject: RE: NSSA and type 7 LSA's -same problem. Configs
incl. > > > >There are two problems discussed; the previous problem is
because of the >fact that the ASBR is connected to two areas, i.e. 11 &
21. that's why it is >clearing the P bit. > >Second, problem where R1 is
generation NSSA with p bit cleared is most >likely be the case of a bug.
>regards, >khalid >-----Original Message----- >From: Nigel Taylor
[mailto:nigel_taylor@hotmail.com] >Sent: Wednesday, September 04, 2002
9:19 AM >To: ccielab@groupstudy.com >Subject: Re: NSSA and type 7 LSA's
-same problem. Configs incl. > > >Chris, > I think you're missing what
Peter is trying to say. As you've >noted(and Peter as well) in normal
conditions the only time the P-bit is >cleared is to eliminate the
translation of type 7's from being translated by >other NSSA ABR's.
However, as Peter point's out the problem here would be >why is the ASBR
clearing the P-bit. I looked around and came back to (as >Howard pointed
out) one of the best books on IP routing (next to Doyle's) >Alex Zinin's
- Cisco IP Routing. > >This is just a thought but if anyone cares to
comment of the validity of my >understanding and or the possibility if
these mechanisms applying across the >board with reference to NSSA LSA
processing, please do so. Page 497, >9.2.5.3, Alex notes the use of the
redistribute command on Cisco OSPF router >and the use of a special
process called the OSPF scanner, which is started. >Reason for the
scanner is to ensure synchronization of the routing table - >specifically
the redistributed routes - and self-originated >AS-external-LSAs. I noted
the redistribute command being used on R1 which >caught my attention. >
>He further notes that it's not enough that the ASBR to originate LSAs.
The >remote router(s) must have corresponding entry in their RIBs to
install >external routes derived from AS-external-LSAs announced by a
particular >router. The most important thing her noted was as follows,
"To ensure that >routers residing in the same area as the ASBR install
such entries and that >ABR's properly announce the ASBR location in the
ASBR-summary-LSAs, the ASBR >sets the E-bit in it's router LSA. > >Can
you confirm the setting of the E-bit in the router LSAs? Also, the >final
paragraph provide some insight as to a number of thing we could look >at
in trying to identify the reasoning behind what you're observing. >
>Thoughts... >Nigel > >P.S. Folks add this books to your Intra-domain
Routing Library. It >provides an level of insight into so many other
aspects of Cisco's protocol >implementation. >(Disclaimer: I was not paid
to make that comment... It's all Howard's >fault. He mentioned the book
in the list - I just was just crazy enough to >buy it :-) ) > > > >
>----- Original Message ----- >From: "Chris Larson" >To: "Peter van Oene"
; "Khalid Siddiq" >; >Sent: Tuesday, September 03, 2002 9:42 PM >Subject:
Re: NSSA and type 7 LSA's -same problem. Configs incl. > > > > The only
reason that I know that this would happen is if the router were > > both
ASBR and ABR and R2 is not. Also, an earlier post had mentioned > >
reacability. I have taken the config off for now, but I remeber that
>subnet > > being an O route.
>__________________________________________________________________ >To
unsubscribe from the CCIELAB list, send a message to
>majordomo@groupstudy.com with the body containing: >unsubscribe ccielab
>
------------------------------------------------------------------------
Join the worlds largest e-mail service with MSN Hotmail. Click Here
This archive was generated by hypermail 2.1.4 : Mon Oct 07 2002 - 07:43:43 GMT-3