GroupStudy.com GroupStudy.com - A virtual community of network engineers
 Home  BookStore  StudyNotes  Links  Archives  StudyRooms  HelpWanted  Discounts  Login
RE: OSPF area 0 question [7:131756] posted 07/02/2008
[Chronological Index] [Thread Index] [Top] [Date Prev][Date Next] [Thread Prev][Thread Next]


Thanks. I've always been opposed to doing things "just for a test lab". The
military often says "train like you fight, fight like you train", or "sweat
during training saves blood during battle."

Things you do for a practice lab only might be what flashes into your mind
in the real thing -- just as incredibly bad BGP practices that seem to be
needed for passing the lab could get you fired by an ISP.

-----Original Message-----
From: nobody@xxxxxxxxxxxxxx [mailto:nobody@xxxxxxxxxxxxxx] On Behalf Of TJ
Sent: Wednesday, July 02, 2008 3:09 PM
To: cisco@xxxxxxxxxxxxxx
Subject: RE: OSPF area 0 question [7:131756]

"Anyway, while it's unlikely in production configurations, I've learned, the
hard way, to look very hard at lab configs, to make absolutely sure that
some neglected direct connection, leftover static route, etc., isn't leaking
connectivity. Is that something that might get snuck into a test simulation
or CCIE exam?"

Indeed; that's why we use "config replace nvram:" as the first step of every
lab :).
	... "config replace" is a beautiful command; especially for training

	... and especially WRT IPv6 (where a merge of configs would
too-frequently result in multiple addresses on interfaces!)



/TJ

-----Original Message-----
From: nobody@xxxxxxxxxxxxxx [mailto:nobody@xxxxxxxxxxxxxx] On Behalf Of
Howard C. Berkowitz
Sent: Tuesday, July 01, 2008 5:22 PM
To: cisco@xxxxxxxxxxxxxx
Subject: RE: OSPF area 0 question [7:131756]

I may not be picturing your configuration correctly, but, back in the days
of steam-powered routers in the basic Cisco lab, the OSPF lab had one router
in each workgroup (i.e., one nonzero area) connected to the backbone. The
routers in the group connected to one another.

More often than I want to remember, I'd tell students to SHUT their area
0.0.0.0 interface, and I'd confidently explain how they could not get to the
other areas.  

Alas, too often, they could. I then reread all the instructions, and
discovered that in the previous lab instruction, the R0 of each workgroup
connected to R3 of the next workgroup. Now, there were no network statements
matching those interfaces, so the lab designer was content that OSPF
wouldn't see the next workgroup.

Unfortunately, said lab designer had not instructed the students to SHUT the
workgroup to workgroup end links. Since each workgroup was in a different
classful network for the RIP and IGRP labs, the each end link happily got
one other classful network into it, VIA DIRECT CONNECT.

So, when I'd tell students to ping, they'd get to the next workgroup with
OSPF disabled on the backbone. Only when I asked them to ping two workgroups
away would the ping fail, although in about every other class, there was
some helpful individual that thought it was OK to telnet to a router in the
adjacent workgroup, and ping from there (that sound is the echo of my head
beating against the desk).

Anyway, while it's unlikely in production configurations, I've learned, the
hard way, to look very hard at lab configs, to make absolutely sure that
some neglected direct connection, leftover static route, etc., isn't leaking
connectivity. Is that something that might get snuck into a test simulation
or CCIE exam?

Another clue came when the students would have debug icmp enabled, and see
echo requests coming in. With no routing protocol, their router wouldn't
know the return path.

Paranoia can be useful.

-----Original Message-----
From: nobody@xxxxxxxxxxxxxx [mailto:nobody@xxxxxxxxxxxxxx] On Behalf Of
Georgecooldude
Sent: Tuesday, July 01, 2008 4:46 PM
To: cisco@xxxxxxxxxxxxxx
Subject: Re: OSPF area 0 question [7:131756]

Earlier I said this:
"Without area 0 nothing will work. Area 1 - area 1 will not talk. area 1 -
area 2 also won't work. Try it in the lab and see."

Gotta eat my words now. Just tried it again in my home lab with two 1721s
connected via serial interfaces using OSPF area 100. Adding loopback
interfaces into the ospf process using the network command under area 100
works just fine. Sorry for any confusion and I guess I also learned
something today. Sure I tested that previously on some 2500 series (about 2
yrs ago) but maybe I overlooked something that time or was using a bugged
IOS.

To recap.
Area 1 to area 1 will work just fine without an area 0 existing. As long as
your using single area OSPF.

Regards




On Tue, Jul 1, 2008 at 6:45 PM, Dino Costantini
wrote:

> Hi Daniel,
>
> an ABR won't put in its routing table an inter-area route learned from 
> another ABR though a non-backbone area. I think it may flood the LSA 
> in the area anyway, I am not sure.
>
> Anyway this will break all inter-area communication because the ABR 
> won't be able to route packets destined to other areas.
>
> Intra-area communication won't be affected as the inter-area flooding 
> will work fine.
>
> Bye,
>
> Dino
>
> Daniel Tsay (dtsay) wrote:
> > Hi,
> >
> > I would like to ask a question and hopefully someone can help.
> >
> > Regarding the area 0 in the OSPF, I know that the area 0 is needed it.
> > What about if I don't configure area 0, but just only area 1 or 2. 
> > Can it still communicate between those two areas?
> >
> > Thanks in advance.




Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=131790&t=131756
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html