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

[Nick Hilliard <[e-mail]>]

Jeffrey Haas wrote on 03/08/2022 19:53:
> The authors will want to consider what the operational considerations are
> for the other side of that connection

in theory, the side which operates the sendhold timer will tear down the 
tcp session and any further tcp packets will be replied to with rst, 
which should in theory reset the session at the remote side, which 
should cause a non-graceful peer shutdown, and I guess eventually things 
will clear themselves up. Of course, when you're talking about 
overloaded cpus / buggy implementations whose digestive systems were 
poor enough to have caused this problem in the first place, who knows 
what will happen in practice. The principle of walking away from bgp 
sessions which are detected as catastrophically bad is probably not a 
bad one to aspire to. At least, it will light up the NMS with angry red 
lights and nee-naw sounds.

Nick