[Outages-discussion] [outages] 3556 problems

[randal k]

I've been pondering the same thing -- but if the Edge-RR sessions went
down, shouldn't the RR then stop announcing after $timeout? Then the
RR-Carrier/Interconnect should have timed out as well, and /all/
routes should have eventually disappeared entirely.

I am concerned that the RFO does not address this situation *at all*
-- yea flowspec broke your BGP, but why the heck were you ghosting
routes for hours? Perhaps they have insane timers on their RR?

On Sun, Aug 30, 2020 at 1:24 PM Frank Bulk <frnkblk at iname.com> wrote:
>
> Is there a scenario where an internal flowspec announcement to filter out
> certain traffic could have prevent BGP updates from AS3356's edge routers to
> get delivered to AS3356's route reflectors, such that even when external BGP
> peers dropped their peering sessions with AS3356 that the routing updates
> didn't get back to AS3356's route reflectors?
>
> Frank
>
> -----Original Message-----
> From: Outages <outages-bounces at outages.org> On Behalf Of Randy Bush via
> Outages
> Sent: Sunday, August 30, 2020 1:14 PM
> To: outages at outages.org
> Subject: Re: [outages] 3556 problems
>
> > initial research identified that an offending flowspec announcement
> > prevented Border Gateway Protocol (BGP) from establishing
>
> opsec paranoid here wonders about the content of the flowspec
> announcement and whence it originated.
>
> randy
> _______________________________________________
> Outages mailing list
> Outages at outages.org
> https://puck.nether.net/mailman/listinfo/outages
>
>
> _______________________________________________
> Outages-discussion mailing list
> Outages-discussion at outages.org
> https://puck.nether.net/mailman/listinfo/outages-discussion