Re: [Idr] SendHoldTimer default value (was: BGP Message level ACKs)

> Snijders <job=40fastly.com@dmarc.ietf.org> wrote:

>> Raszuk suggested 60 minutes.

Indeed I mentioned 60 min based on the below quote from Nick:

*this draft is about unsticking connections which have been jammed fortime
periods longer than hours [bgpzombies], and it's normal for BGPsession
uptime to be measured in months rather than hours. For example,here's the
output of an entirely unremarkable dfz leaf router:*

So credit here goes to him not me,

Thx,
Robert

Snijders <job=40fastly.com@dmarc.ietf.org> wrote:

> Dear Jeff, all,
>
> Everyone - thank you so much for all the feedback so far, we'll try to
> incorporate all suggestions we received in the next version of the
> draft.
>
> On Fri, Mar 10, 2023 at 08:46:07AM -0500, Jeffrey Haas wrote:
> > Please keep in mind the mix of peer scale and holdtimers that may get
> > used in a deployment.  I'm supportive of something like a
> > sendholdtimer, when the value is long enough.
>
> The current version of the draft (-09) suggests 8 minutes for the
> SendHoldTimer value. We the authors picked this number by simply
> doubling the suggested HoldTimer value.
>
> Nick suggested a procedure, and a default of 20 minutes here:
> https://mailarchive.ietf.org/arch/msg/idr/vaN72cmrtHXS4aBVPrwCFJ-1TMg/
> Raszuk suggested 60 minutes.
>
> Jeff, you say "the value needs to be long enough"
>
> But what is long enough? We are open to suggestions.
>
> Kind regards,
>
> Job
>
> _______________________________________________
> Idr mailing list
> Idr@ietf.org
> https://www.ietf.org/mailman/listinfo/idr
>