[j-nsp] BFD between Cisco - Juniper when FRR is enabled does not torn down primary tunnel

[Robert Kern]

Hi David,

we were using both version 9.2 and 9.4 with mpls oam feature, but as fas as
I know Cisco does not support this functionality. For testing BFD we teared
down logical interface, so the meaning of BFD actually come in place
(physical port stayed up).

On Cisco side I saw that BFD actually trigered FRR, but on Juniper side
primary Tunnel stayed up - it actually didn't come down. According to
documentation BFD should triger also FRR to torn primary tunnel down.

Am I missing some configuration? How to configure BFD for RSVP - I found
only BFD for MPLS (oam)?

Regards,

Robert

On Fri, Apr 3, 2009 at 1:53 PM, David Ball <davidtball at gmail.com> wrote:

>  How long does it take for your FRR to kick in?  Is your logical or
> physical link actually down (you mention both below) ?  I imagine if
> you want BFD-speed failover, you may want to look at 9.4, which
> apparently supports BFD for LDP and RSVP sessions.
>
> David
>
>
>
> On 03/04/2009, Robert Kern <cj11st at gmail.com> wrote:
> > Hi all,
> >
> > we have run into a problem with BFD between Cisco and Juniper box when
> > MPLS-FRR is configured.
> >
> > ISIS is used as IGP protocol and one hop MPLS-TE tunnels are configured
> with
> > link protection. BFD is configured on both sides (under IS-IS protocol).
> > After primary logical link is dropped between Cisco-Juniper (simulating
> DWDM
> > system), BFD session and isis adj are torn down on both sides. The
> problem
> > is that Juniper does not signal to MPLS-TE FRR that physical interface is
> > logicaly down and FRR should be used. Instead it keeps  primary tunnel up
> > showing faulty interface as outgoing. The result is that traffic is
> > blackholed. On the other side Cisco re-routes traffic corectly.
> >
> > Am I missing some configuration or this is a known issue?
> >
> > Config on both sides (FRR without BFD works fine):
> >
> > ge-1/3/0
> >         description Link_to_Cisco;
> >         mtu 9110;
> >         hold-time up 0 down 0;
> >         unit 0
> >             family inet
> >                 address 10.100.111.50/30;
> >
> >             family iso;
> >             family inet6
> >                 address 2A00:EE00:0:12:10:100:111:50/64;
> >
> >             family mpls;
> >
> > protocols
> >     rsvp
> >         interface ge-1/3/0.0
> >             authentication-key
> > "$9$9cAaCORSrvxNd9AIclKx7jHqmfzAtO1IcApclMXbwHqm"; ## SECRET-DATA
> >             bandwidth 850m;
> >             link-protection
> >                 path
> >                     10.100.111.66 strict;
> >                     10.100.111.53 strict;
> > mpls
> >         traffic-engineering mpls-forwarding;
> >         explicit-null;
> >         ipv6-tunneling;
> >         standby;
> >         label-switched-path Protect-ge020
> >             to 10.100.100.5;
> >             ldp-tunneling;
> >             link-protection;
> >             primary path1;
> >
> >         path path1
> >             10.100.111.49 strict;
> >         interface ge-1/3/0.0;
> >
> >
> > isis
> >
> >         lsp-lifetime 65000;
> >         spf-options
> >             delay 50;
> >             holddown 2000;
> >
> >         topologies ipv6-unicast;
> >         overload timeout 600;
> >         traffic-engineering
> >             family inet
> >                 shortcuts;
> >
> >             family inet6
> >                 shortcuts;
> >
> >
> >         level 1 disable;
> >         level 2
> >             authentication-key "$9$L04XVYoJD.P5bsfz3/0OxNd"; ##
> SECRET-DATA
> >             authentication-type md5;
> >             wide-metrics-only;
> >
> >
> >
> >         interface ge-1/3/0.0
> >             point-to-point;
> >             bfd-liveness-detection
> >                 version automatic;
> >                 minimum-interval 300;
> >                 minimum-receive-interval 300;
> >                 multiplier 3;
> >
> > On Cisco side I have under interface:
> >
> > bfd interval 300 min_rx 300 multiplier 3
> >
> > and under router isis:
> >
> > bfd all-interfaces
> >
> >
> > Regards,
> >
> > Robert
> > _______________________________________________
> > juniper-nsp mailing list juniper-nsp at puck.nether.net
> > https://puck.nether.net/mailman/listinfo/juniper-nsp
> >
>