From: Yakout Esmat (yesmat@xxxxxxxxxxxxxx)
Date: Tue Jul 16 2002 - 21:47:36 GMT-3
Gents,
Thanks very much for your participation and feed back
Yakout
-----Original Message-----
From: Chris Hugo [mailto:chrishugo@yahoo.com]
Sent: Wednesday, July 17, 2002 6:34 AM
To: Carlos G Mendioroz; Yakout Esmat
Cc: kris.keen@aon.com.au; Groupstudy; nobody@groupstudy.com
Subject: Re: Default route over ISDN
Hi,
I realize most folks might not have the Alex Zinin book but since you guys
are discussing a topic I just read through I will post a couple of snippets
from his book that I feel might clear the air a little bit.
Page 242-243 and section 5.4.2.3
----start of excerpt----
If you revisit the logic of the forwarding engine in Cisco routers. you
can see that when a packet is routed along a route referencing only an
interface, the packets destination address is used as the next-hop address.
This is done just the same way as when a packet is delivered to a host on a
directly connected subnet: The forwarding engine finds the interface route,
takes the destination address as the next hop and passes this information to
the packet-delivery procedure. The packet delivery code uses ARP or map
tables to find data-link details corresponding to the next hop address. The
same process is used when a static route is configured with INTERFACE ONLY.
Another example of incorrect use of static routes in a Frame Relay
environment was in a point-to-multipoint scenario. The forwarding engine
used the destination IP addresses in the packets to look through the Frame
Relay map table. As no matching entry was found--the map statements were
configured for the next-hop addresses corresponding to remote routers, not
hosts to remote subnets-- the majority of the packets dropped.
The same problem can show up if static routes referencing only interfaces
are used for DDR: The dialer maps contain the address of remote routers,
whereas actual next-hop address used by the forwarding engine are the
destination addresses from the packets.
So what is rule of thumb?
Use static routes over interfaces only if the corresponding interface is
point-to-point, that is if encapsulation is HDLC, PPP, or is a
point-to-point sub int. Otherwise, always configure static routes with both
the output interface and the next-hop address.
----end of excerpt----
hth,
chris hugo
Carlos G Mendioroz <tron@huapi.ba.ar> wrote:
Yes, it also works. But WHY would you want to do that ?!!!
The static to the interface is bringing in trouble. Use a floating to
the next hop!
Again, I'm missing the reason of your wanting it not to show in the
routing table...
Yakout Esmat wrote:
>
> I don't think it is possible too, I haven't seen any documentation
where
> local interface is used on floating static route.
>
> I have no routers to test it at the moment, but will post the results
once I
> did that.
>
> -----Original Message-----
> From: kris.keen@aon.com.au [mailto:kris.keen@aon.com.au]
> Sent: Tuesday, July 16, 2002 4:42 PM
> To: Yakout Esmat
> Cc: Groupstudy; nobody@groupstudy.com; Carlos G Mendioroz
> Subject: RE: Default route over ISDN
>
> I would doubt that, if you point a static to a connected interface,
does
> the distance default to 0? Can you override admin distance? like
this..
> I will mock this up as I'd be intrested to see
>
> -----------------------------------------------------------------
> Kris Keen - CCNP, CCDP, CNE
> Network Support Specialist - Network Systems
> Aon Risk Services Australia Limited
> (612) 9253 7272
> 0404862970
> E: Kris.Keen@aon.com.au
>
> "Yakout Esmat"
> >
> s.com.au> cc: "Groupstudy"
>
> Sent by: bcc:
> nobody@groupst Subject: RE: Default route over
> ISDN
> udy.com
>
> 16/07/2002
> 04:05 PM
> Please respond
> to "Yakout
> Esmat"
>
> I guess this also might work, only if you have another default route
> pointing to different direction, and if this disappears use the
floating
> one.
> But not sure that this is possible "ip route 0.0.0.0 0.0.0.0 bri0 200"
>
> -----Original Message-----
> From: Carlos G Mendioroz [mailto:tron@huapi.ba.ar]
> Sent: Tuesday, July 16, 2002 10:53 AM
> To: Yakout Esmat
> Cc: Groupstudy
> Subject: Re: Default route over ISDN
>
> Well,
> as Mas suggested, putting a dialer string works, just tested it.
>
> What I do not understand is your reason for making this an interface
> route.
> If what you want is that the route is hidden, I guess a floating
static
> is in order.
>
> Yakout Esmat wrote:
> >
> > Carlos,
> >
> > It makes sense, but if I need to hide the default route from routing
> table,
> > then I have to point to my local bri interface, otherwsie if I point
to
> next
> > hop it will be installed in my routing table.
> >
> > What you are saying is that if I point to my interface it doesn't
work??
> >
> > -----Original Message-----
> > From: Carlos G Mendioroz [mailto:tron@huapi.ba.ar]
> > Sent: Tuesday, July 16, 2002 3:12 AM
> > To: Yakout esmat
> > Cc: Groupstudy
> > Subject: Re: Default route over ISDN
> >
> > Yakout,
> > when you send a packet to an interface (via ip route 0.0.0.0/0 int
x)
> > somehow the interface has to encapsulate it and send it.
> >
> > On DDR interfaces that also accounts for knowing where do you have
to
> > dial to. This is "solved" via maps iif you have a destination IP,
but
> > in your first case, you don't. Therefore, "encapsulation error".
> >
> > Does this make sense to you ?
> >
> > Yakout esmat wrote:
> > >
> > > Any body aware of this problem..
> > >
> > > Scenario:
> > >
> > > 10.1.1.1/24--[R1]------(192.168.1.0/24)-----[R2]----
> > > ISDN
> > >
> > > 1) R2 is configured with a static default route
> > >
> > > Problem:
> > >
> > > 1) If the default static route on R2 is pointing to BRI0 interface
"ip
> > route
> > > 0.0.0.0 0.0.0.0 bri0", R2 can't ping 10.1.1.1 on R1. Getting
> > "encapsulation
> > > failed" error when debuging ip packets, and R2 can't initiate a
dial.
> > >
> > > 2) If I cahnge the default to point to next hop instead "ip route
> 0.0.0.0
> > > 0.0.0.0 192.168.1.1", every thing is sweet
> > >
> > > IOS 12.1, legacy ISDN with single dialer map statement is used, IP
> > classless
> > > is on...
> > >
> > > I must be missing some thing.
> > >
> > > Thanks
> > >
> > > Yakout
This archive was generated by hypermail 2.1.4 : Sat Sep 07 2002 - 19:36:33 GMT-3