February 2012 - Bird-users

Neighbor lost for bird6
by Simone Morandini 27 Feb '12

27 Feb '12

Hi all, I'm facing a weird problem: we run BIRD with identical configuration on 2 different servers, both connected on the same peering lan. We had "Error: Neighbor lost" error message for all the IPv6 bgp speakers, on both machines, while the v4 deamon had no problems. Issuing a "kill -s HUP" didn't change anything, while shutting the deamon down and then re-starting it actually solved the problem. From the logfile I see that all the speakers went to "flush" state, then "stop", and then "start", but none of the sessions came up. Any clue on what could be the cause? TIA, Simone. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.

3 4

Strange error message
by fredrik danerklint 27 Feb '12

27 Feb '12

Hi! I'm trying to turn the session up against a peer to no success. What I've got was this: 2012-02-26 18:22:57 <RMT> b6: Received: No supported AFI/SAFI: 000201 2012-02-26 18:25:29 <RMT> b4: Received: No supported AFI/SAFI: 000101 What does that mean? -- //fredan

3 3

How redistribute routes from kernel table to OSPF protocol?
by Владислав Гришин 22 Feb '12

22 Feb '12

1. Routes into kernel routing table proxmox-host1:/# ip ro list 10.18.1.25 dev venet0 scope link 10.18.1.26 dev venet0 scope link 10.16.1.0/24 dev vmbr0 proto kernel scope link src 10.16.1.27 10.17.1.0/24 dev eth1 proto kernel scope link src 10.17.1.27 default via 10.16.1.1 dev vmbr0 proxmox-host1:/# 2. Bird Config file filter export_OSPF { # redistribute connected if ( source = RTS_DEVICE ) then { print "net accepted:", net; ospf_metric2 = 20; accept; } # redistribute static if ( source = RTS_STATIC_DEVICE ) then { print "net accepted:", net; ospf_metric2 = 20; accept; } if ( source = RTS_STATIC ) then { print "net accepted:", net; ospf_metric2 = 20; accept; } reject; } filter import_OSPF { if ( source = RTS_OSPF_EXT2 && net != 0.0.0.0/0 ) then { print "net accepted:", net; accept; } reject; } protocol direct { interface "eth1"; interface "venet0"; } protocol kernel { # table main; learn; # Learn all alien routes from the kernel device routes; # persist off; # Don't remove routes on bird shutdown # scan time 20; # Scan kernel routing table every 20 seconds scan time 2; # Scan kernel routing table every 60 seconds import all; # Default is import all export all; # Default is export none # kernel table 5; # Kernel table to synchronize with (default: main) } protocol ospf myospf { router id 10.17.1.27; debug all; export filter export_OSPF; import filter import_OSPF; area 0.0.0.0 { interface "eth1" { hello 10; retransmit 5; cost 10; transmit delay 1; dead count 4; wait 40; type broadcast; priority 0; authentication cryptographic; password "ХХХХХХХХХХ"; }; }; } 3. Output bird console bird> show route all 0.0.0.0/0 via 10.16.1.1 on vmbr0 [kernel1 2012-02-22 12:47:54] * (10) Type: inherit unicast univ 10.16.1.0/24 via 10.17.1.28 on eth1 [myospf 2012-02-22 12:48:04] * E2 (150/10/20) [10.17.1.28] Type: OSPF-boundary unicast univ OSPF.metric1: 10 OSPF.metric2: 20 OSPF.tag: 0x00000000 OSPF.router_id: 10.17.1.28 10.17.1.0/24 dev eth1 [direct1 2012-02-22 12:47:54] * (240) Type: device unicast univ bird> But I don't see these routes in a bird 10.18.1.25 dev venet0 scope link 10.18.1.26 dev venet0 scope link (Only default route present from kernel.) It is necessary to declare these routes to OSPF neighbors. Vladislav Grishin

1 2

Graceful Restart functionality
by Aris Lambrianidis 22 Feb '12

22 Feb '12

Hello Ondrej(s) and all, I wanted to ask if you guys consider implementing full RFC 4724 (Graceful Restart Mechanism for BGP) support for BIRD. --Aris Lambrianidis

1 0

RIPng (Sorry for beating a dead horse!)
by David Splittberger 20 Feb '12

20 Feb '12

So I've just started using bird and just wanted to verify with you all what I'm seeing in RIPng. I've scoured the mailing list archives so I'm pretty sure this is the same bug that has been reported before but I figured I would just see if maybe I've actually got a misconfiguration instead. So I'm running BIRD 1.3.2 on a couple of ubuntu boxes: Linux newport 2.6.28-11-server #42-Ubuntu SMP Fri Apr 17 02:45:36 UTC 2009 x86_64 GNU/Linux bird> show status BIRD 1.3.2 setup is pretty simple. I've got two boxes but logically i've made 3 by using two tables on one to simulate two routers. configs exactly are: eth2 - table2 RA - eth0 ---------------------- eth0 RB eth1 ----------------------- eth1 table3 RB -eth3 RouterA: router id 2.2.3.5; table t4; table t5; protocol kernel kern1 { import all; export all; table t4; kernel table 2; scan time 20; } protocol kernel kern2 { import all; export all; table t5; kernel table 3; scan time 20; } protocol device { scan time 10; } protocol direct dir1 { table t4; import all; interface "eth0","eth2"; } protocol direct dir2 { table t5; import all; interface "eth1","eth3"; } protocol rip my_rip1 { table t4; import filter { print "importing"; accept; }; export filter { print "exporting"; accept; }; port 521; honor neighbor; interface "eth0" { metric 2; mode broadcast; }; authentication none; period 30; garbage time 60; } protocol rip my_rip2 { table t5; import filter { print "importing"; accept; }; export filter { print "exporting"; accept; }; port 521; honor neighbor; interface "eth1" { metric 2; mode broadcast; }; authentication none; period 30; garbage time 60; } RouterB: router id 2.2.3.4; protocol kernel kern1 { import all; export all; scan time 20; } protocol device { scan time 10; } protocol direct { import all; interface "eth0","eth1"; } protocol rip my_rip1 { import filter { print "importing"; accept; }; export filter { print "exporting"; accept; }; port 521; honor neighbor; interface "eth0" { metric 2; mode broadcast; }; interface "eth1" { metric 2; mode broadcast; }; authentication none; period 30; garbage time 60; } Now, with mode set to broadcast when using RIPng, I see no updates being sent out on my links. Packet captures turn up nothing and debug <rip instance> all shows nothing. If I change mode broadcast to multicast, I see updates being sent but I then get these messages: bird> debug >>> my_rip2: fe80::215:17ff:fe0d:86f5 send me routing info but he is not my neighbor >>> my_rip1: fe80::215:17ff:fe0d:86f4 send me routing info but he is not my neighbor >>> my_rip2: fe80::215:17ff:fe0d:86f5 send me routing info but he is not my neighbor >>> my_rip1: fe80::215:17ff:fe0d:86f4 send me routing info but he is not my neighbor >>> my_rip2: fe80::215:17ff:fe0d:86f5 send me routing info but he is not my neighbor >>> my_rip1: fe80::215:17ff:fe0d:86f4 send me routing info but he is not my neighbor >>> my_rip2: fe80::215:17ff:fe0d:86f5 send me routing info but he is not my neighbor >>> my_rip1: fe80::215:17ff:fe0d:86f4 send me routing info but he is not my neighbor >>> my_rip2: fe80::215:17ff:fe0d:86f5 send me routing info but he is not my neighbor >>> my_rip1: fe80::215:17ff:fe0d:86f4 send me routing info but he is not my neighbor I hadn't seen those messages turn up in anyone elses posts to the mailing list about RIPng so it's making me think I have a misconfiguration somewhere. Thanks David

3 2

Unable to fetch route from tunnel device
by Dave Lloyd 17 Feb '12

17 Feb '12

I'm trying to set up an OpenVPN server that routes for multiple sites and I'm having a bit of trouble getting Bird to add the routes that go through the tun0 device. I've configured Bird to import routes from both the OS and tun0, but filter out any routes that don't go through the tun0 IP (I only want to advertise routes that go through my OpenVPN tunnel). I'm running Bird 1.3.6. On startup, I receive messages like these: >>> KRT: Received route 172.24.1.16/28 with strange next-hop 172.16.254.1 >>> Netlink: File exists >>> Netlink: File exists >>> OSPF: Socket error: Operation not permitted This appears to be coming from this section of code: sysdep/linux/netlink/netlink.c: 846 if (!ng || (ng->scope == SCOPE_HOST)) { log(L_ERR "KRT: Received route %I/%d with strange next-hop %I", net->n.prefix, net->n.pxlen, ra.gw); return; } My network setup is as follows: Local IP of tun0 on OpenVPN server: 172.16.254.1/24 IP of tun0 on remote server: 172.24.1.28/28 Routes (from ip route) on OpenVPN server: 172.16.254.1 dev tun0 proto bird 172.24.1.0/28 via 172.16.254.1 dev tun0 172.24.1.16/28 via 172.16.254.1 dev tun0 Routes (from ip route) on OpenVPN client: 172.16.0.0/24 via 172.24.1.28 dev tun0 Bird config: log syslog { debug, trace, info, remote, warning, error, auth, fatal, bug }; filter no_local prefix set local_nets; ip tun0; { local_nets = [ 172.16.0.0/16+ , 192.168.0.0/16+ , 10.0.0.0/8+ ]; tun0 = 172.16.254.1; printn "gw: ", gw; if gw = tun0 then accept "accept: gw is ",gw; reject "reject: gw is ", gw; } protocol kernel { learn; # Learn all alien routes from the kernel persist; # Don't remove routes on bird shutdown scan time 10; # Scan kernel routing table every 20 seconds device routes on; # this is on so we can accept routes with local nexthop import filter no_local; # don't import "local" networks import none; # Default is import all export all; # Default is export none } protocol device { scan time 10; # Scan interfaces every 10 seconds } protocol static { } protocol direct { interface "tun0"; } protocol ospf MFOSPF { export all; area 172.16.0.0 { interface "eth4" { type broadcast; authentication none; neighbors { 10.19.7.11; }; }; interface "tun0" { type ptmp; authentication none; }; }; } Any ideas about what I'm doing wrong here? The network configuration (aside from Bird) works as I can ping from another IP, 172.16.1.11, through the OpenVPN server and out to 172.24.1.28. Is there a way to get Bird to advertise a route via OSPF that goes through a local IP (for a tun device) that is not on the same subnet as the remote network? Thanks much! --dlloyd

2 1

Again on communities...
by Simone Morandini 01 Feb '12

01 Feb '12

Hi guys, one question (again) on communities: consider the following filtering function, applied on pipe protocols (no filtering on bgp protocol): function bgp_out(int peeras) { if !(source = RTS_BGP) then return false; if (0, peeras) ~ bgp_community then return false; if (65004, peeras) ~ bgp_community then return true; if (0, 65004) ~ bgp_community then return false; return true; } If a given peer wants to peer only with two specific ASNs, he will then send these communities: (65004,ASN1) (65004,ASN2) to the route server. Is it correct to say that such a filter simply doesn't work for this purpose? That is, routes are however announced to all the participants? Should it be modify to be somehow like this: function bgp_out(int peeras) { if !(source = RTS_BGP) then return false; if (community attribute is non-empty) { apply some filter logic; return false; } else return true; } Thanks, Simone. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.

2 1