BGP Peering shuts down with "add paths tx;"

Tim Weippert weiti+bird-users at weiti.org
Sun Jan 14 15:48:15 CET 2024


Hi Ondrej, 

I tested it with my IPv6 Peering, with no routes announced the Peering established correctly (from my point of view):
[..]
    Local capabilities
      Multiprotocol
        AF announced: ipv6 flow6
      Route refresh
      Graceful restart
      4-octet AS numbers
      ADD-PATH
        RX:
        TX: ipv6
      Enhanced refresh
      Long-lived graceful restart
    Neighbor capabilities
      Multiprotocol
        AF announced: ipv6
      Route refresh
      Graceful restart
        Restart time: 120
        AF supported: ipv6
        AF preserved:
      4-octet AS numbers
      ADD-PATH
        RX: ipv6
        TX: ipv6
[..]

But as soon as routes get announced the Peering went down:

B_TEST_v6 BGP        ---        start  14:19:48.911  Active        Received: Invalid network field
  BGP state:          Active
    Neighbor address: 2a05:XXX:XXX:X
    Neighbor AS:      65000
    Local AS:         65000
    Connect delay:    0.214/5
    Last error:       Received: Invalid network field
  Channel ipv6
    State:          DOWN
    Table:          T_TEST_v6
    Preference:     100
    Input filter:   (unnamed)
    Output filter:  (unnamed)
    IGP IPv6 table: master6
  Channel flow6
    State:          DOWN
    Table:          TF_TEST_v6
    Preference:     100
    Input filter:   ACCEPT
    Output filter:  ACCEPT

And the Cisco side went down with:


%ROUTING-BGP-3-MALFORM_UPDATE : Malformed UPDATE message received from neighbor <IPV6 Neighbor IP> (VRF: TEST) - message length 215 bytes, error flags 0x00000008, action
+taken "Reset". Error details: "Error 0x00000008, Field "MP_REACH-NLRI", Attribute 14 (Flags 0x00, Length 0), Data [53]". NLRIs: [IPv4 Unicast]

I send you the complete pcap per PM, and also the complete show command from above.

Regards, 
tim

On Thu, Jan 11, 2024 at 05:42:02PM +0100, Ondrej Zajicek wrote:
> On Tue, Jan 09, 2024 at 05:10:49PM +0100, Tim Weippert via Bird-users wrote:
> > 
> > Hi All,
> > 
> > i try to activate "add paths tx;" from my Bird2 (2.0.12) Routeserver to a Cisco IOS-XR Systems (7.9.2), after i add
> > the config on Bird side, the IOS-XR shuts down the peering:
> > 
> > %ROUTING-BGP-3-MALFORM_UPDATE : Malformed UPDATE message received from neighbor <IPV6-RR> (VRF: INTERNET) - message length 160 bytes, error flags 0x00000008, action taken "Reset". Error details: "Error 0x00000008, Field "MP_REACH-NLRI", Attribute 14 (Flags 0x00, Length 0), Data [6d]". NLRIs: [IPv4 Unicast]
> > %ROUTING-BGP-5-ADJCHANGE : neighbor <IPV6-RR> Down - BGP Notification sent, illegal network (VRF: INTERNET) (AS: 65000)
> > ...
> >
> > Does this means, that bird doesn't add the path-id as expected?
> > Is there some known caveouts around?
> 
> Hi
> 
> Well, it should work as expected.
> 
> Can you establish a session without sending anything (to not trigger the bug) and send the output of 'show protocols all'?
> 
> Or perhaps make a tcpdump/pcap of such session with updates?
> 
> -- 
> Elen sila lumenn' omentielvo
> 
> Ondrej 'Santiago' Zajicek (email: santiago at crfreenet.org)
> "To err is human -- to blame it on a computer is even more so."

-- 
Tim Weippert
http://weiti.org - weiti at weiti.org
GPG Fingerprint - E704 7303 6FF0 8393 ADB1  398E 67F2 94AE 5995 7DD8


More information about the Bird-users mailing list