Re: [Idr] TCP & BGP: Some don't send terminate BGP when holdtimer expired, because TCP recv window is 0

[Enke Chen <[e-mail]>]

Hi, Jakob:

It's about recovery from corner cases that occur rarely than other
"holdtime expiration" events.

Keep it simple and robust.

Thanks.   -- Enke

On Fri, Dec 11, 2020 at 11:15 PM Jakob Heitz (jheitz) <[e-mail]>
wrote:

> Kill and restart all from scratch with BGP is a scary prospect,
>
> especially if it's just going to repeat.
>
>
>
> How about we graceful restart it.
>
> Require GR to be in effect on the session.
>
> Then either RST or silently drop the session and start a new one.
>
> The receiver should accept the new session and do GR.
>
> And, yeah, use a long timer. Don't restart again until at least the EOR
>
> is both sent and received and only then start the boom timer again.
>
>
>
> Regards,
>
> Jakob.
>
>
>
> *From:* Idr <[e-mail]> *On Behalf Of * Enke Chen
> *Sent:* Friday, December 11, 2020 10:15 PM
> *To:* [e-mail]
> *Cc:* Enke Chen <[e-mail]>
> *Subject:* Re: [Idr] TCP & BGP: Some don't send terminate BGP when
> holdtimer expired, because TCP recv window is 0
>
>
> Hi, Folks:
>
>
>
> There is an interesting article titled "When TCP Sockets Refuse to Die":
>
>
>       https://blog.cloudflare.com/when-tcp-sockets-refuse-to-die/
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__blog.cloudflare.com_when-2Dtcp-2Dsockets-2Drefuse-2Dto-2Ddie_&d=DwMGaQ&c=V9IgWpI5PvzTw83UyHGVSoW3Uc1MFWe5J8PTfkrzVSo&r=OPLTTSu-451-QhDoSINhI2xYdwiMmfF5A2l8luvN11E&m=lifWUvp3-DcEhCcrTN7S0OaBssKd2YgRWJK9wKS6kME&s=iaULKaznGYFPvvgj5LqIC6FRCjyl4DU6NzlDLBRNoKA&e=>
>
>
>
> which recommends using the TCP keepalive option and the TCP_USER_TIMEOUT
> option together to deal with several TCP "stuck" scenarios.
>
>
>
> To strike a balance between maintaining routing stability and working
> around these corner cases, how about we recommend using these two TCP
> options with a timeout value larger than the BGP per-session holdtimer
> (e.g.,  2 * bgp_holdtimer) ?
>
>
>
> Thanks.   -- Enke
>
> ------
> Re: [Idr] TCP & BGP: Some don't send terminate BGP when holdtimer expired,
> because TCP recv window is 0
>
> "Jakob Heitz (jheitz)" <[e-mail]> Sat, 12 December 2020 03:29 UTCShow
> header
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__mailarchive.ietf.org_arch_browse_idr_&d=DwMGaQ&c=V9IgWpI5PvzTw83UyHGVSoW3Uc1MFWe5J8PTfkrzVSo&r=OPLTTSu-451-QhDoSINhI2xYdwiMmfF5A2l8luvN11E&m=lifWUvp3-DcEhCcrTN7S0OaBssKd2YgRWJK9wKS6kME&s=VXdBMh8C3fKDigTD71jkq9slGVhCFKdLa5EqbvlCUbg&e=>
>
> Good point Keyur.
>
> A receiver may be overwhelmed for a long time and not open its TCP window to avoid
>
> silly window syndrome or some other reason. The receiver may still be functional
>
> and able to clear its backlog, albeit in a long time. Resetting such a session
>
> will only make the situation worse. Telling the difference between this case
>
> and a receiver stuck in a bug is difficult.
>
>
>
> Regards,
>
> Jakob.
>
>