Hi Jeffrey, > On May 9, 2023, at 17:15, Jeffrey (Zhaohui) Zhang <[email protected]> wrote: > > Hi Acee, > > For argument sake (this is non-blocking), I don't see difference between > static routes and protocol-learned routes. Why would protocol-learned routes > need to have nexthop-specific backup while static routes don't?
The static route definition is read-write so this would need to be supported as a feature since I don’t know anyone who supports this. If someone really wants this, they can do it with a separate augmentation model. There are other static options that are more widely implemented that we haven’t included in the base model. I don’t see a requirement given that backup next-hops can be defined via higher preference. I know that you are arguing that this backup isn’t “next-hop specific”. However, if you have ECMP, normally an IGP would use the ECMP as a back up as well (though, I’m aware that options have been implemented to override this). Thanks, Acee > > Thanks. > Jeffrey > > -----Original Message----- > From: Acee Lindem <[email protected]> > Sent: Thursday, May 4, 2023 10:51 AM > To: Jeffrey (Zhaohui) Zhang <[email protected]> > Cc: Yingzhen Qu <[email protected]>; Routing Directorate > <[email protected]>; [email protected]; > [email protected]; [email protected] > Subject: Re: Rtgdir last call review of draft-ietf-rtgwg-yang-rib-extend-16 > > [External Email. Be cautious of content] > > > Hi Jeffrey, > > >> On May 1, 2023, at 5:05 PM, Jeffrey (Zhaohui) Zhang <[email protected]> >> wrote: >> >> Hi Yingzhen, >> I can go with that if that’s what the authors/WG’s preference, but my >> preference would be to cover it in the base specification itself. What’s the >> harm? >> Anyway, this is not a blocking comment. > > The way the backup static route use case is handled is by configuring > multiple next-hops with different preferences. This is one of the most useful > extensions provided by this augmentation. > > Thanks, > Acee > > >> Thanks. >> Jeffrey >> From: Yingzhen Qu <[email protected]> >> Sent: Monday, May 1, 2023 5:01 PM >> To: Jeffrey (Zhaohui) Zhang <[email protected]> >> Cc: [email protected]; [email protected]; >> [email protected]; [email protected] >> Subject: Re: Rtgdir last call review of >> draft-ietf-rtgwg-yang-rib-extend-16 >> [External Email. Be cautious of content] Hi Jeffrey, Considering >> this is not commonly used, I'd suggest if someone really needs it they can >> do an easy augmentation using the grouping defined in this draft. >> Thanks, >> Yingzhen >> On Mon, May 1, 2023 at 1:52 PM Jeffrey (Zhaohui) Zhang <[email protected]> >> wrote: >> Hi Yingzhen, >> From: Yingzhen Qu <[email protected]> >> Sent: Monday, May 1, 2023 4:46 PM >> To: Jeffrey (Zhaohui) Zhang <[email protected]> >> Cc: [email protected]; [email protected]; >> [email protected]; [email protected] >> Subject: Re: Rtgdir last call review of >> draft-ietf-rtgwg-yang-rib-extend-16 >> [External Email. Be cautious of content] Hi Jeffrey, Thanks for the >> review, please see my answers below. >> Thanks, >> Yingzhen >> On Mon, May 1, 2023 at 11:43 AM Zhaohui Zhang via Datatracker >> <[email protected]> wrote: >> Reviewer: Zhaohui Zhang >> Review result: Has Issues >> >> I have the following one nit comment and one question: >> >> augment "/rt:routing/rt:ribs/rt:rib/" >> + "rt:routes/rt:route/rt:next-hop/rt:next-hop-options/" >> + "rt:next-hop-list/rt:next-hop-list/rt:next-hop" >> { >> description >> "Augment the multiple next hops with repair path."; >> uses repair-path; >> } >> >> The description is slightly misleading. It is to agument a single >> next-hop in the next-hop-list, not "multiple next hops". >> [Yingzhen] how about: "Augment the next-hop with a repair path." >> Zzh> Good. >> Shouldn't the repair path be applicable to static routes as well? >> [Yingzhen]: Theoretically you can have a repair-path for a static route, but >> have you seen this in deployment? >> Zzh> Whether anyone implemented/deployed it that way, I think it’s quite >> reasonable and desired to have it covered in the spec. For example, a static >> route could be using if1 by default but if2 as backup (in case if1 is down). >> Zzh> Jeffrey >> Juniper Business Use Only >> >> Juniper Business Use Only > > > > Juniper Business Use Only _______________________________________________ rtgwg mailing list [email protected] https://www.ietf.org/mailman/listinfo/rtgwg
