[j-nsp] BGP over GRE

[Scott Morris]

    gr-0/0/0 {
        unit 0 {
            clear-dont-fragment-bit;
            tunnel {
                source xxx.xxx.115.178;
                destination yyy.yyy.112.2;
                no-path-mtu-discovery;
            }
            family inet {
                mtu 1500;
                address 10.255.255.1/30;
            }
        }           
    } 
Protocols {
  bgp  {
        group STDIO {
            type external;
            traceoptions {
                file STDIO;
                flag packets send receive detail;
            }
            local-address 10.255.255.1;
            import No-Routes-In;
            authentication-key "$9$Opu01IcreWLxdhSs4aZq.5QF/A0IEy"; ##
SECRET-DATA
            export BadMyLinks;
            peer-as 7963;
            neighbor 10.255.255.2;
        }
    }


Yes, my autonomous-system is defined...  My other eBGP peers work perfectly
fine (one directly connected, two multihop) without issue.

I'd love to give more log details, but there aren't any!  :)

Notice the traceoptions in the config there.  Here is the ENTIRE file (yes,
the date/time on the router is accurate):
smorris at Emanon-Edge> show log STDIO 
Apr 13 15:48:53 trace_on: Tracing to "/var/log/STDIO" started
Apr 13 20:42:27.099946 bgp_peer_mgmt_clear: NOTIFICATION sent to
10.255.255.2 (External AS 7963): code 6 (Cease) subcode 4 (Administratively
Reset), Reason: Management session cleared BGP neighbor
Apr 13 20:42:27.101386 bgp_send: sending 21 bytes to 10.255.255.2 (External
AS 7963)
Apr 13 20:42:27.101410
Apr 13 20:42:27.101410 BGP SEND (null) -> 10.255.255.2
Apr 13 20:42:27.103228 BGP SEND message type 3 (Notification) length 21
Apr 13 20:42:27.103288 BGP SEND Notification code 6 (Cease) subcode 4
(Administratively Reset)
Apr 13 20:42:27.103566 bgp_send: sending 21 bytes to 10.255.255.2 (External
AS 7963) failed: Bad file descriptor
Apr 18 23:12:37.048543 bgp_peer_mgmt_clear: NOTIFICATION sent to
10.255.255.2 (External AS 7963): code 6 (Cease) subcode 4 (Administratively
Reset), Reason: Management session cleared BGP neighbor
Apr 18 23:12:37.048719 bgp_send: sending 21 bytes to 10.255.255.2 (External
AS 7963)
Apr 18 23:12:37.048742
Apr 18 23:12:37.048742 BGP SEND (null) -> 10.255.255.2
Apr 18 23:12:37.048788 BGP SEND message type 3 (Notification) length 21
Apr 18 23:12:37.048807 BGP SEND Notification code 6 (Cease) subcode 4
(Administratively Reset)
Apr 18 23:12:37.049050 bgp_send: sending 21 bytes to 10.255.255.2 (External
AS 7963) failed: Bad file descriptor

smorris at Emanon-Edge>

So something freaky is going on, just not sure what it is!

Scott

-----Original Message-----
From: Kristian Larsson [mailto:kristian at spritelink.se] 
Sent: Friday, May 04, 2007 2:18 AM
To: Scott Morris
Cc: juniper-nsp at puck.nether.net
Subject: Re: [j-nsp] BGP over GRE

Could we please have som configuration excerpts?
Right now the only similar scenario I can come up with was when I was doing
iBGP over GRE, but then I made the mistake of terminating both the GRE and
the iBGP session on the other routers loopback and so it failed. Since this
is an eBGP session it's even less complicated. But please, provide some
configurations and some more log messages wouldn't hurt.

  Kristian.

On Fri, May 04, 2007 at 12:34:37AM -0400, Scott Morris wrote:
> This is a very strange question, and very strange scenario...  but I'm 
> also getting some very strange errors, so I'm hoping that someone here 
> may have seen this before and can give me some hint of whatever I'm 
> apparantly not thinking of!
>  
> I have a GRE tunnel from a J2300 to a Cisco router.  The GRE is very 
> simple, and it works just fine (at least as far as pinging the other 
> end of the tunnel or telnetting to it goes!).  However, BGP over this 
> tunnel does not work.  Again, it SHOULD be a very simple setup, 
> directly connected ebgp peers, no multihop, nothing strange.
>  
> It's not working.  On the Cisco side, it shows as an authentication 
> error (at first we weren't doing any, but then turned it on just to 
> see if it made a difference which it did not).  But on the Juniper 
> side, the error shows up ONLY in the /var/log/messages file.
>  
> May  4 00:31:01  Emanon-Edge rpd[2797]: bgp_pp_recv: NOTIFICATION sent 
> to
> 10.255.255.2+34211 (proto): code 2 (Open Message Error) subcode 5 
> (authentication failure), Reason: no group for 10.255.255.2+34211 
> (proto) from AS 7963 found (peer idled), dropping him
> 
> This was the same message with or without authentication enabled on 
> the BGP portion.  When I do a "monitor interface" to try to watch the 
> traffic, I don't even SEE any outbound bgp traffic.  Same thing with 
> traceoptions detail.  This "no group" message is quite vexing.
>  
> Anyone seen anything like this before?  I'm hoping it's something 
> simple and driven by lack of sleep, but nothing is leaping at me.
>  
> TIA,
>  
> Scott
>  
> _______________________________________________
> juniper-nsp mailing list juniper-nsp at puck.nether.net 
> https://puck.nether.net/mailman/listinfo/juniper-nsp

-- 
Kristian Larsson                                   KLL-RIPE
Network Engineer                       SpriteLink [AS39525]
+46 704 910401			     kristian at spritelink.se