[j-nsp] Fwd: Quick way to Shift MPLS traffic away from an interface

[Wojciech Janiszewski]

Hi Adam,

RFC 5817 could be a solution for that problem but apparently it's not
supported yet. Pitty.

Regards,
Wojciech
22 maj 2015 15:13 "Adam Vitkovsky" <Adam.Vitkovsky at gamma.co.uk> napisał(a):

> Hello Tim,
>
> > tim tiriche
> > Sent: 21 May 2015 04:02
> >
> > Hello,
> >
> > What is the quick way to shift LSP traffic from an interface after
> > increasing the igp metric?
> >
> If you have configured loose EROs (hops) in your LSPs and you have also
> used "optimize-timer" (as unfortunately Juniper does not have a reoptimize
> timer enabled by default) the LSPs should reoptimize automatically after a
> period of time configured.
>
>
> > question:
> >
> > - What command can I use to find all lsp traversing the iface and a good
> > way to clear them? I am assuming I would need to run clear mpls
> > optimize-aggressive on the lsp's on that particular router only? Is my
> > understanding correct?
> >
> I think you can use "show mpls lsp" to list the active RSVP LSPs
> traversing the box.
>
> But how to trigger LSP reoptimization from a mid-point router that is a
> good question indeed.
> Since the LSPs are signalled from the head-end router they can be rerouted
> only by the head-end router.
> Saying that you have to force the mid-point router to send a RSVP PathErr
> msg towards the head-end router,
> the head-end router should then perform the LSP reoptimization
> As Ivan proposed breaking the RSVP session is the obvious choice
> -but preferably you'd want to do the reoptimization in a make before break
> fashion (though if you don't have the LSPs configured as adaptive it
> doesn't matter)
> -anyways I really don't know if it's possible to trigger reoptimization
> from a mid-point router when you are not using BW reservations.
>
>
>
> > - Is it a good idea to turn on optimize-aggressive?
> If you are using RSVP LSPs just for pure MPLS forwarding and not for BW
> reservations or you're not using SRLGs then you can reduce the CSPF to just
> comparing IGP metric.
>
> >
> > Any best practices or pointers would be appreciated!
> I think that good practice would be "optimize-timer" so that paths are
> optimized automatically (e.g. once a day) and to use adaptive LSPs to
> enable make before break.
>
>
> addam
>
> ---------------------------------------------------------------------------------------
>  This email has been scanned for email related threats and delivered
> safely by Mimecast.
>  For more information please visit http://www.mimecast.com
>
> ---------------------------------------------------------------------------------------
> _______________________________________________
> juniper-nsp mailing list juniper-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp
>