October 2021 - Bird-users

Broadcast RIP to each broadcast address on interface with multiple aliases
by bird21＠sensation.net.au 23 Oct '21

23 Oct '21

Hi all, I want to broadcast a last-resort old-school default route via RIPv1. The following config works: filter default_route { if net ~ [ 0.0.0.0/0 ] then { accept; } reject; } # ...snip... protocol rip { export filter default_route; interface "em0" { version 1; mode broadcast; update time 5; }; } ...but it only sends to the broadcast address of the main interface IP. How can I get it to send to the broadcast addresses of the aliases as well? These are the networks on the em0 adapter: x.x.x.192/28 (main) 192.168.x.0/24 (alias) 192.168.y.0/24 (alias) After consulting https://bird.network.cz/doc/bird-3.html#proto-iface , I tried the following to force additional/different addresses: interface x.x.x.193/28 { # actual interface IP/prefix interface x.x.x.192/28 { # network IP/prefix interface x.x.x.x { # interface IP only interface "em0" x.x.x.x { # interface name and IP interface "em0" x.x.x.x/28 { # interface name and IP/prefix All of these are rejected with the error "Interface name/mask expected, not IP prefix", regardless of whether I list the main interface IP, or main plus aliases (comma separated). Am I missing something simple in the syntax, or is the RIPv1 implementation not capable of doing this? Thanks in advance.

2 2

OSPFv3 between BIRD2 and RouterOS (Mikrotik)
by Kees Meijs | Nefos 21 Oct '21

21 Oct '21

Hi list, First of all: I'm very sorry if this has been discussed before. Anyway, we're experiencing possible (configuration) incompatibiliy between BIRD2 and Mikrotik RouterOS. To be precise, we're on BIRD 2.0.7 and RouterOS 6.48.5 (Level 6). Configuration snippet from BIRD: > protocol ospf v3 ospf_v6 { > # Do not generate ECMP (equal-cost multipath) routes. > ecmp no; > > area 0.0.0.0 { > interface "eth0" { > authentication cryptographic; > password "SOMETHING"; > bfd yes; > }; > > interface "eth1" { > authentication cryptographic; > password "SOMETHINGELSE"; > bfd yes; > }; > > interface "dummy0" { > stub; > }; > > interface "eth2" { > stub; > }; > > interface "eth3" { > type pointopoint; > bfd yes; > }; > }; > } The dummy0 interface has bound a /32 IPv4 and a /128 IPv6 address for BGP lookback. OSPFv3 should provide us with IGP in order to facilitate iBGP. Between BIRD nodes this works flawless. When adding the Mikrotik router to the mix, the /128 addresses don't get propagated. Other routes do and OSPFv3 seems to work fine, but the stub routes seem to be ignored. What am I missing here? Thanks! Cheers, Kees -- https://nefos.nl/contact <https://nefos.nl/contact> Nefos IT bv Ambachtsweg 25 (industrienummer 4217) 5627 BZ Eindhoven Nederland KvK 66494931 /Bereikbaar op maandag, dinsdag, donderdag en vrijdag tussen 09:00u en 17:00u./

2 3

bird adversites a direct route from an interface without carrier
by Alarig Le Lay 21 Oct '21

21 Oct '21

Hi, On one router I have an interface that used to be connected to another machine, so the state is NO-CARRIER: core01-arendal ~ # ip addr show enp1s0 2: enp1s0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000 link/ether 00:0d:b9:48:c1:c0 brd ff:ff:ff:ff:ff:ff inet 10.0.4.1/31 scope global enp1s0 valid_lft forever preferred_lft forever inet6 2001:4640:a14f:fffd::2/127 scope global valid_lft forever preferred_lft forever inet6 fe80::20d:b9ff:fe48:c1c0/64 scope link valid_lft forever preferred_lft forever On another router, I still have this route from OSPF, advertised from core01-arendal: core02-arendal ~ # birdc BIRD 2.0.8 ready. bird> show route all for 10.0.4.0 Table master4: 10.0.4.0/31 unicast [ospf_ipv4 17:51:54.613] * E2 (150/1/10000) [45.91.126.254] via 10.0.4.5 on enp2s0 Type: OSPF-E2 univ igp_metric: 1 OSPF.metric1: 1 OSPF.metric2: 10000 OSPF.tag: 0x00000000 OSPF.router_id: 45.91.126.254 bird> show route all for 2001:4640:a14f:fffd::2 Table master6: 2001:4640:a14f:fffd::2/128 unicast [ospf_ipv6 17:51:58.613] * I (150/1) [45.91.126.254] via fe80::20d:b9ff:fe48:c1c2 on enp2s0 Type: OSPF univ igp_metric: 1 OSPF.metric1: 1 OSPF.router_id: 45.91.126.254 bird> And here is the export filter for OSPF: export filter { if (proto = "kernel_grt_ipv4" || proto = "direct1" || source = RTS_DEVICE) then { accept; } else reject; }; Given that the interface is unplugged, I expected bird to withdrawn the route rather than keeping it into OSPF. Regards, -- Alarig Le Lay

2 5

Bird just doesn't want to find OSPF neighbors although they are there and can communicate
by Lukas Haase 20 Oct '21

20 Oct '21

Hi, I am having big trouble getting Bird/OSPF working. I broke it down to a simple setup: Point-to-point wireguard link between Node1 (192.168.56.224/31) and Node2 (192.168.56.225/31) using nbma. There were nothing but troubles with multicast and to avoid all these hassles for now, I directly switched to nbma. Config Station 1: protocol ospf test { area 0.0.0.0 { interface "wg-tun" { cost 10; type nbma; authentication cryptographic; password "VWj3QH8LAtWIzQca"; neighbors { 192.168.56.225; }; }; }; } Config Station 2: protocol ospf test { area 0.0.0.0 { interface "wg-tun" { type nbma; cost 10; authentication cryptographic; password "VWj3QH8LAtWIzQca"; neighbors { 192.168.56.224; }; }; }; } - Ping from both sides to the other works. Yet the neighbor table just stay empty: node1:~# birdc show ospf neighbors BIRD 1.6.8 ready. test: Router ID Pri State DTime Interface Router IP node1:~# node2~:# birdc show ospf neighbors BIRD 1.6.8 ready. test: Router ID Pri State DTime Interface Router IP node2:~# Both OSPF Hello are received via tcpdump (see below). What else can be missed? Thanks, Lukas node1:~# tcpdump -n -vvv -i wg-tun 'proto ospf' tcpdump: listening on wg-tun, link-type RAW (Raw IP), capture size 262144 bytes 23:41:09.515892 IP (tos 0xc0, ttl 1, id 65434, offset 0, flags [none], proto OSPF (89), length 80) 192.168.56.224 > 192.168.56.225: OSPFv2, Hello, length 44 Router-ID 192.168.56.193, Backbone Area, Authentication Type: MD5 (2) Key-ID: 1, Auth-Length: 16, Crypto Sequence Number: 0x00012747 Options [External] Hello Timer 10s, Dead Timer 40s, Mask 255.255.255.254, Priority 1 Designated Router 192.168.56.224 23:41:19.589835 IP (tos 0xc0, ttl 1, id 10125, offset 0, flags [none], proto OSPF (89), length 80) 192.168.56.225 > 192.168.56.224: OSPFv2, Hello, length 44 Router-ID 192.168.56.194, Backbone Area, Authentication Type: MD5 (2) Key-ID: 1, Auth-Length: 16, Crypto Sequence Number: 0x00209357 Options [External] Hello Timer 10s, Dead Timer 40s, Mask 255.255.255.254, Priority 1 Designated Router 192.168.56.225 node2:~# tcpdump -n -vvv -i wg-tun 'proto ospf' tcpdump: listening on wg-tun, link-type RAW (Raw IP), snapshot length 262144 bytes 23:41:29.723176 IP (tos 0xc0, ttl 1, id 2705, offset 0, flags [none], proto OSPF (89), length 80) 192.168.56.224 > 192.168.56.225: OSPFv2, Hello, length 44 Router-ID 192.168.56.193, Backbone Area, Authentication Type: MD5 (2) Key-ID: 1, Auth-Length: 16, Crypto Sequence Number: 0x00012748 Options [External] Hello Timer 10s, Dead Timer 40s, Mask 255.255.255.254, Priority 1 Designated Router 192.168.56.224 23:41:39.732903 IP (tos 0xc0, ttl 1, id 12839, offset 0, flags [none], proto OSPF (89), length 80) 192.168.56.225 > 192.168.56.224: OSPFv2, Hello, length 44 Router-ID 192.168.56.194, Backbone Area, Authentication Type: MD5 (2) Key-ID: 1, Auth-Length: 16, Crypto Sequence Number: 0x00209358 Options [External] Hello Timer 10s, Dead Timer 40s, Mask 255.255.255.254, Priority 1 Designated Router 192.168.56.225

4 8

bgp_large_community wildcard in conditional
by Brooks Swinnerton 18 Oct '21

18 Oct '21

Hello, I was wondering if it's possible to create a conditional that matches on a wildcard for a part of a BGP large community. For example: if ([*, 600, 6939]) ~ bgp_large_community then reject; This currently results in a syntax error at the position of the comma after the *.

5 7

Deleting a single route in BGP
by Johannes Ludwig 05 Oct '21

05 Oct '21

Hello everyone! Currently I'm writing my bachelor thesis that also contains an implementation of an extension in BIRD and BGP. One building block is deleting a single route in BGP without triggering an Update message (with route withdraws) and the route must be deleted programmatically, so without receiving an Update message. The problem is, that I see no support for that. I can't call rte_update2 with new == NULL, because this would delete all routes that lead to the specified network. But I want to delete only one specific route! The problem description and what I have done so far: Delete a specific route without triggering a BGP Update Message: Since BGP has no use case where only a specific route should be deleted, this problem is difficult (at least for me). Normally we receive a Route Withdraw Message and can call rte_update2() with the not reachable network and when rte * new == NULL all routes that lead to the target network are deleted. But in my use case I only want to remove one specific route. At the moment I add specific pflags to the rte I want to delete and in rte_update2 I check if new->pflags has the specific value. If so, I jump to drop in rte_update2, where the route is deleted. The route is deleted but no recalculation is done, therefore I set new to the best route of the reachable network (net->routes) and jump to recalc in rte_update2(). I add my specific pflags and when bgp_rt_notify in attrs.c is called I return. Unbelievable but this works -- almost! My problem is that the route is deleted in the BIRD BGP Routingtable, so when I call birdc show route all it shows me all routes without the deleted one (so far so good), BUT the kernel routing table is not changed. And this is the point where I have to change my approach. To summarize my problem in one sentence: How can I delete a single route in BIRD without declaring a network as unreachable and calling rte_update2() with new == NULL, because this approach would lead to the deletion of all routes that lead to a network. I hope someone can help me and maybe explain where I can look, what functions could be useful and what things I have to consider. Many thanks in advance! Johannes Ludwig

2 1

Lower number of IPv6 routes received than expected
by Adrian Brooks 04 Oct '21

04 Oct '21

Hello, I'm hoping someone can help me out with some troubleshooting please. I'm using two IPv4 sessions to send the Internet routing table into Bird. Taking the first session, the advertising router tells me it's sending 871273 IPv4 unicast routes and 134440 IPv6 unicast routes towards Bird. All the IPv4 routes arrive, but I only get ~16k IPv6 routes. I don't have any filters configured to reject routes so I'm puzzled as to what's going on. How could I troubleshoot further? I didn't see any obviously related messages in the log. BIRD 2.0.8 ready. bird> show route import table uplink1.ipv4 count 871273 of 871273 routes for 871273 networks in table import bird> bird> show route import table uplink1.ipv6 count 16091 of 16091 routes for 16091 networks in table import The second session (sourced from a different router) is better, but there's still not enough: bird> show route import table uplink2.ipv6 count 45357 of 45357 routes for 45357 networks in table import Thanks in advance for any pointers, Regards, Adrian Adrian BROOKS Telecom Networks - Fixed Services Platforms Tel. : +352 2424 2301 adrian.brooks(a)post.lu<mailto:adrian.brooks@post.lu> POST Technologies 20, Rue de Reims L-2417 Luxembourg www.posttechnologies.lu<http://www.posttechnologies.lu/> [logo_technologies(a)2x.png] Ce message et toutes les pièces jointes (ci-après le "message") sont établis à l'intention exclusive de ses destinataires et sont confidentiels. Si vous recevez ce message par erreur, merci de le détruire et d'en avertir immédiatement l'expéditeur. Toute utilisation de ce message non conforme à sa destination, toute diffusion ou toute publication, totale ou partielle, est interdite, sauf autorisation expresse. Internet ne permettant pas d'assurer l'intégrité de ce message, POST Luxembourg décline toute responsabilité au titre de ce message, dans l'hypothèse où il aurait été modifié.

2 2

bird-2.0.8 in Debian testing
by Jakub Ružička 03 Oct '21

03 Oct '21

Hello Community, latest bird2-2.0.8 is now available from Debian testing (and unstable) \o/ HOWTO get on Debian bullseye: echo "deb http://deb.debian.org/debian testing main" >> /etc/apt/sources.list apt update apt -t testing install bird2 # TEST ALL THE THINGS This can be backported into bullseye-backports. Cheers, Jakub Ružička CZ.NIC packager 📦

3 2