Re: [Idr] Fwd: New Version Notification for draft-spaghetti-idr-bgp-sendholdtimer-05.txt

[heasley <[e-mail]>]

Thu, Aug 04, 2022 at 08:56:43PM +0200, Robert Raszuk:
> Thx for the example. Of course UPDATES work as KEEPALIVES, but this case
> seems to be about not reading/draining either.

yes, as has been mentioned several times on this thread, there are cases
where the remote never drains the socket for days or more.  A bug, clearly.

> The current draft for adoption already recommends mangling with BGP State
> Machine. If we go and solve it at TCP no change to BGP State Machine is
> needed. So this is more than implementation choice.

you mean alter; mangling is your opinion.

RFC7606 is an example of altering BGP to accomodate broken implementations
and, while I personally would prefer a way to punish the broken
implementation, the RFC explains how that is not possible and this is a
reasonable way to address the problem.  This draft seems similar in this
respect.

Tell us how to fix it in TCP.  Jeff has already explained how it would
not necessarily address the problem, depending upon how an implementation
separates the TCP and BGP.  TCP KAs may (probably will) continue to work,
so what else is there?

> So assume we bring the session back. Do we require manual UP ?

IMO, implementation/operator decision.  Could be like existing
max-prefix exceeded implementations.  This also seems like a post-adoption
discussion item; whether the draft makes any recommendation at all.