Re: Graceful restart comment

[Pedro Roque Marques <[e-mail]>]

Jeffrey Haas writes:

> On Mon, Apr 29, 2002 at 01:12:04PM -0700, Gargi Nalawade wrote:
>> I agree. The semantics of the NOTIFICATION message shouldnt be
>> changed.  We seem to need a new message type, say an INFORMATIONAL
>> message that sends out information (notifies actually) to the peer,
>> but doesnt necessarily cause a session reset.

> Other things this would be useful for is when packets are discarded
> by an implementation where it is unnecessary to actually tear down
> the peering session.  A good example is when a peer sends you an
> MP_REACH_NLRI that you don't support, or you're going to start
> ignoring a given AFI/SAFI because of errors in the PDU, but are
> going to leave the peering session up.

Jeff,
If you have errors in a PDU the best thing to do is to tear down the
session not to selectivly ignore bits... if you can't trust your peer
to format updates correctly then you can't trust any of the nlri that
it has sent you.

Accepting bad routing information is usually much harder to debug than
not accepting information...

I can't think of any scenario where it is reasonable to discard
packets from the neighbor while maintaining the session up... if your
neighbor is sending you packets for an NLRI that it hasn't negociated,
chances are that there is something else wrong with it.

The "be liberal in what you accept" principle is a very good idea when
we are talking about encoding formats but personally i don't think it
applies when your peer demonstrates that it's internal state is fubar.

  Pedro.