Prashant/Daniel,

Besides what Alex has already sent you may like to have a look at the
discussions in the archives. This thing has been discussed a lot of times in
the past.

Check
http://discuss.microsoft.com/SCRIPTS/WA-MSD.EXE?A2=ind0008&L=OSPF&P=R2394
and
http://discuss.microsoft.com/SCRIPTS/WA-MSD.EXE?A2=ind0008&L=OSPF&P=R2279 .

Besides the MTU check is (from RFC2328)

        If the Interface MTU field in the Database Description packet
        indicates an IP datagram size that is larger than the router can
        accept on the receiving interface without fragmentation, the
        Database Description packet is rejected.

So if we did not want an MTU mismatch in any case, we could always send a
value 0 for the Interface MTU, as is done on virtual links.

Thanks,
Vishwas

-----Original Message-----
From: Daniel Gryniewicz [mailto:dangNEXTHOP.COM]
Sent: Friday, June 21, 2002 12:07 AM
To: OSPFDISCUSS.MICROSOFT.COM
Subject: Re: mtu in DD exchange.

True, and it'll be resent forever, taking 9000 bytes of bandwidth every
time.
:( And I'll still use it in my SPF, and so the black hole will still
occure.

Daniel

On Thu, 20 Jun 2002 11:35:20 -0700
prashant desai <prashant_s_desaiHOTMAIL.COM> wrote:

> that makes sense :-)
> but the drop lsa will not be acked....never reachd ospf
>
> thanks
> Prashant
>
>
> >From: Daniel Gryniewicz <dangNEXTHOP.COM>
> >Reply-To: Mailing List <OSPFDISCUSS.MICROSOFT.COM>
> >To: OSPFDISCUSS.MICROSOFT.COM
> >Subject: Re: mtu in DD exchange.
> >Date: Thu, 20 Jun 2002 14:30:58 -0400
> >
> >So, I send you the 9000 byte packet. My IP layer thinks it's fine. Your
> >IP
> >layer drops it. An adjacency never forms. Or, worse yet, the adjacency
> >formed (the HELLOs and DDs were all small enough) and then I get and
flood
> >a
> >huge LSA from someone else. You drop it, but I assume you have it. My
SPF
> >puts a route from this LSA through you, but you don't have it, so
> >blackhole.
> >I *have* to know that every packet I send to you can make it. The MTU
> >agreement assures that packets won't get dropped because of MTU mismatch.
> >
> >Daniel
> >
> >On Thu, 20 Jun 2002 11:27:52 -0700
> >prashant desai <prashant_s_desaiHOTMAIL.COM> wrote:
> >
> > > exactly..the ip layer should drop the packet..why should OSPF care
about
> > > it..dont you think its the ip layers job to check mtu conformance of
> > > incomming packets ?
> > >
> > >
> > > >From: Daniel Gryniewicz <dangNEXTHOP.COM>
> > > >Reply-To: Mailing List <OSPFDISCUSS.MICROSOFT.COM>
> > > >To: OSPFDISCUSS.MICROSOFT.COM
> > > >Subject: Re: mtu in DD exchange.
> > > >Date: Thu, 20 Jun 2002 13:59:07 -0400
> > > >
> > > >What fragments? All OSPF (baring virtual links) traffic is only on
one
> > > >link.
> > > >What the MTU agreement assures is that everyone on the link agrees
what
> >the
> > > >MTU *for that link* should be. If I think the MTU is 9000, and you
> >think
> > > >the
> > > >MTU is 1500, you'll silently drop all my packets. My IP layer won't
> > > >fragment,
> > > >because it thinks the MTU is large enough.
> > > >
> > > >Daniel
> > > >
> > > >On Thu, 20 Jun 2002 10:45:44 -0700
> > > >prashant desai <prashant_s_desaiHOTMAIL.COM> wrote:
> > > >
> > > > > Hi,
> > > > > Can anyone give me a good reason why OSPF routers have to agree on
> > > > > the MTU size during DD exchange process..
> > > > > As OSPF runs over IP..fragmentation or reassembly should not be
its
> > > > > concern..and they will be transparently handled by the IP layer..
> > > > >
> > > > > Prashant
> > > > >

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]