November 2012 - Bird-users

Neighbor lost on interface flap
by Simone Morandini 30 Nov '12

30 Nov '12

Hi guys, we use two machines running BIRD v. 1.3.8 as a bgp route-server on our IXP peering lan, and I noticed an apparently strange behavior: we had to reload the switch module on which our route server is connected, but after that, all v4 neighbors were showing "Error: neighbor lost", while all the v6 neighbors were ok. After quitting the daemon and re-starting it, all the v4 bgp peers were up & running. This happened all the times we had to reload the switch module, that is, all the times we "disconnected" the network interface from the peering lan. Any clue on what could be the cause of this? Thanks, Simone. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.

1 0

ospf virtual link not coming up
by Martin Kraus 30 Nov '12

30 Nov '12

Hi. I'm playing with ospf and got stuck on a virtual link with a cisco router. The cisco virtual link is up and I can see hello packets coming to the linux router but nothing going out. It's connected through area 0.0.0.1. I can see both routers in lsdb. debug in bird produces following: OSPF: Received packet for unknown vlink (ID 192.168.70.1, IP 10.5.1.1) I assume I'm doing something wrong but I can't see what as the config seems very simple. bird RID 10.4.7.1: area 0.0.0.1 { interface "r2" { cost 10; type pointopoint; neighbors { 10.4.8.2; }; }; virtual link 192.168.70.1 { }; }; bird> show ospf interfaces Virtual link to 192.168.70.1: Peer IP: 0.0.0.0 Transit area: 0.0.0.1 (1) Interface: "(none)" State: down Priority: 0 Cost: 0 Hello timer: 10 Wait timer: 40 Dead timer: 40 Retransmit timer: 5 bird> show ospf lsadb area 1 type 1 Area 0.0.0.1 Type LS ID Router Age Sequence Checksum 0001 192.168.60.1 192.168.60.1 353 80000008 5485 0001 192.168.70.1 192.168.70.1 232 80000006 b37b 0001 10.4.7.1 10.4.7.1 146 80000019 1312 cisco RID 192.168.70.1: router ospf 1 area 0.0.0.1 virtual-link 10.4.7.1 R3#show ip ospf virtual-links Virtual Link OSPF_VL2 to router 10.4.7.1 is up Run as demand circuit DoNotAge LSA allowed. Transit area 0.0.0.1, via interface FastEthernet0/1, Cost of using 20 Transmit Delay is 1 sec, State POINT_TO_POINT, Timer intervals configured, Hello 10, Dead 40, Wait 40, Retransmit 5 Hello due in 00:00:04 thanks for help Martin

2 3

OSPFv3 (IPv6) over OpenVPN IPv4 PTP connection.
by Paul Libert 24 Nov '12

24 Nov '12

Hi List ! I'm trying to get bird6 OSPF working over a OpenVPN link on Ubuntu Linux. bird (IPv4) works fine. This is a Ubuntu Precise box but I upgraded bird to last version (1.3.8) It's talking with other Ubuntu boxes running quagga over OpenVPN links using tun interfaces. The relevant part of the bird6.conf is protocol ospf ospf1 { tick 2; rfc1583compat yes; area 0.0.0.0 { stub no; interface "tun*" { type ptp; }; }; } This is exactly the same as the one in the bird.conf file. On IPv4, the machines can establish OSPF communication as show by the output from birdc: bird> show ospf interface ospf1: Interface tun0 (peer 10.200.0.2) Type: ptp Area: 0.0.0.0 (0) State: ptp Priority: 1 Cost: 10 Hello timer: 10 Wait timer: 40 Dead timer: 40 Retransmit timer: 5 Interface tun1 (peer 10.200.1.1) Type: ptp Area: 0.0.0.0 (0) State: ptp Priority: 1 Cost: 10 Hello timer: 10 Wait timer: 40 Dead timer: 40 Retransmit timer: 5 On IPv6, the bird6 never try to establish connection, interface stay in the down state because they are marked as stub: bird> show ospf interface ospf1: Interface tun0 (IID 0) Type: ptp Area: 0.0.0.0 (0) State: down (stub) Priority: 1 Cost: 10 Hello timer: 10 Wait timer: 40 Dead timer: 40 Retransmit timer: 5 Interface tun1 (IID 0) Type: ptp Area: 0.0.0.0 (0) State: down (stub) Priority: 1 Cost: 10 Hello timer: 10 Wait timer: 40 Dead timer: 40 Retransmit timer: 5 I tracked it down to ospf_iface_stubby in proto/ospf/iface.c returning 1 because IA_HOST bit is set. IA_HOST bit is set because in sysdep/linux/netlink.c, in nl_parse_addr, ipa_equal(ifa.ip, addr) is true. I cannot add a peer IPv6 address on a OpenVPN tun interface (and I think that's what's make it work on IPv4 so the ip address output looks like: 9: tun0: <POINTOPOINT,MULTICAST,NOARP,UP,LOWER_UP> mtu 1450 qdisc pfifo_fast state UNKNOWN qlen 100 link/none inet 10.200.0.1 peer 10.200.0.2/32 scope global tun0 inet6 2001:470:e553:1::1/64 scope global valid_lft forever preferred_lft forever inet6 fe80::216:3eff:fe79:ed82/128 scope link valid_lft forever preferred_lft forever 10: tun1: <POINTOPOINT,MULTICAST,NOARP,UP,LOWER_UP> mtu 1450 qdisc pfifo_fast state UNKNOWN qlen 100 link/none inet 10.200.1.2 peer 10.200.1.1/32 scope global tun1 inet6 fe80::216:3eff:fe79:ed82/128 scope link valid_lft forever preferred_lft forever I tried to use 'neighbors' stanza specifying the ip address of the peer, with types in ptp, ptmp, broadcast, nonbroadcast without success. I'm probably missing something obvious and would be glad if someone could point it out to me. If I use quagga on both side of the link, everything works. For reference, here is a tshark output of the link traffic: 0.000000 10.200.1.1 -> 224.0.0.5 OSPF 68 Hello Packet 0.000225 10.200.1.2 -> 224.0.0.5 OSPF 68 Hello Packet 0.852653 fe80::216:3eff:feb4:31b -> ff02::5 OSPF 76 Hello Packet 10.000924 10.200.1.1 -> 224.0.0.5 OSPF 68 Hello Packet 10.001042 10.200.1.2 -> 224.0.0.5 OSPF 68 Hello Packet 10.031023 10.200.1.1 -> 224.0.0.5 OSPF 84 LS Update 10.862718 fe80::216:3eff:feb4:31b -> ff02::5 OSPF 76 Hello Packet 11.044247 10.200.1.2 -> 10.200.1.1 OSPF 64 LS Acknowledge 20.002061 10.200.1.1 -> 224.0.0.5 OSPF 68 Hello Packet 20.002272 10.200.1.2 -> 224.0.0.5 OSPF 68 Hello Packet 20.872998 fe80::216:3eff:feb4:31b -> ff02::5 OSPF 76 Hello Packet It shows the IPV4 OSPF broadcast of both node and the v6 broadcast of the quagga node. Cheers ! Paul

2 2

Multiple Table interface Problem
by Stefano Bonino 23 Nov '12

23 Nov '12

Dear all, we have some problem trying to attach an interface to a particular routing table. We have created the routing table 60 in the kernel here * odorizzi@vpn1:~$ cat /etc/iproute2/rt_tables # # reserved values # 255 local 254 main 253 default 0 unspec # # local # #1 inr.ruhep 60 auslre * we have sync the bird configuration as reported # Turn on global debugging of all protocols #debug protocols all; # The direct protocol automatically generates device routes to # all network interfaces. Can exist in as many instances as you wish # if you want to populate multiple routing tables with device routes. protocol direct { interface "-vlan*", "-eth1", "-eth2","eth0"; } protocol direct { table auslre; # interface "vlan60"; interface "eth1", "eth2"; } # This pseudo-protocol performs synchronization between BIRD's routing # tables and the kernel. If your kernel supports multiple routing tables # (as Linux 2.2.x does), you can run multiple instances of the kernel # protocol and synchronize different kernel tables with different BIRD tables. protocol kernel { learn; # Learn all alien routes from the kernel persist; # Don't remove routes on bird shutdown scan time 20; # Scan kernel routing table every 20 seconds import all; # Default is import all export all; # Default is export none # kernel table 32367; # Kernel table to synchronize with (default: main) } protocol kernel { learn; # Learn all alien routes from the kernel persist; # Don't remove routes on bird shutdown scan time 20; # Scan kernel routing table every 20 seconds import all; # Default is import all export all; # Default is export none kernel table 60; # Kernel table to synchronize with (default: main) table auslre; } # This pseudo-protocol watches all interface up/down events. protocol device { scan time 10; # Scan interfaces every 10 seconds table auslre; primary "eth1" 192.168.0.1; primary 192.168.0.0/30; primary "eth2" 172.16.0.1; primary 172.16.0.0/30; } # Static routes (again, there can be multiple instances, so that you # can disable/enable various groups of static routes on the fly). protocol static { # disabled; # Disable by default table auslre; # Connect to a non-default table preference 1000; # Default preference of routes debug { states, routes, filters, interfaces, events, packets }; debug all; route 0.0.0.0/0 via 192.168.0.2; # route 198.51.100.0/25 reject; # route 10.0.0.0/8 reject; # route 10.1.1.0:255.255.255.0 via 198.51.100.3; # route 10.1.2.0:255.255.255.0 via 198.51.100.3; # route 10.1.3.0:255.255.255.0 via 198.51.100.4; # route 10.2.0.0/24 via "vlan60"; } #protocol static { # debug { states, routes, filters, interfaces, events, packets }; # debug all; # route 0.0.0.0/0 via 195.62.186.254; #} # Pipe protocol connects two routing tables... Beware of loops. #protocol pipe { # peer table auslre; # Define what routes do we export to this protocol / import from it. # import all; # default is all # export all; # default is none # import none; # If you wish to disable imports # import filter test_filter; # Use named filter # import where source = RTS_DEVICE; # Use explicit filter #} # RIP aka Rest In Pieces... #protocol rip MyRIP { # You can also use an explicit name # preference xyzzy; # debug all; # port 1520; # period 7; # infinity 16; # garbage time 60; # interface "*" { mode broadcast; }; # honor neighbor; # To whom do we agree to send the routing table # honor always; # honor never; # passwords { # password "nazdar"; # }; # authentication none; # import filter { print "importing"; accept; }; # export filter { print "exporting"; accept; }; #} protocol ospf ospf_test { tick 2; rfc1583compat yes; area 0.0.0.0 { stub no; networks {195.62.186.128/25; }; interface "eth0" { type pointopoint; authentication none; neighbors { 195.62.186.184 eligible; }; }; }; } protocol ospf ospf_auslre { tick 2; table auslre; rfc1583compat yes; area 0.0.0.0 { stub no; networks {192.168.0.0/30; }; interface "eth1" { # hello 9; # retransmit 6; # cost 10; # transmit delay 5; # dead count 5; # wait 50; # type broadcast; type pointopoint; authentication none; # password "pass"; neighbors { 192.168.0.2 eligible; # 10.1.1.4; }; }; interface "eth2" { type pointopoint; authentication none; }; # interface "arc0" { # rx buffer large; # type nonbroadcast; # poll 14; # dead 75; # neighbors { # 10.1.1.2 eligible; # 10.1.1.4; # }; # strict nonbroadcast yes; # }; # interface "xxx0" { # passwords { # password "abc" { # id 1; # generate to "22-04-2003 11:00:06"; # accept to "17-01-2004 12:01:05"; # }; # password "def" { # id 2; # generate from "22-04-2003 11:00:07"; # accept from "17-01-2003 12:01:05"; # }; # }; # authentication cryptographic; # }; }; # area 20 { # stub 1; # interface "ppp1" { # hello 8; # authentication none; # }; # interface "fr*"; # virtual link 192.168.0.1 { # password "sdsdffsdfg"; # authentication cryptographic; # }; # }; } #protocol bgp { # disabled; # description "My BGP uplink"; # local as 65000; # neighbor 198.51.100.130 as 64496; # multihop; # hold time 240; # startup hold time 240; # connect retry time 120; # keepalive time 80; # defaults to hold time / 3 # start delay time 5; # How long do we wait before initial connect # error wait time 60, 300;# Minimum and maximum time we wait after an error (when consecutive # # errors occur, we increase the delay exponentially ... # error forget time 300; # ... until this timeout expires) # disable after error; # Disable the protocol automatically when an error occurs # next hop self; # Disable next hop processing and always advertise our local address as nexthop # path metric 1; # Prefer routes with shorter paths (like Cisco does) # default bgp_med 0; # MED value we use for comparison when none is defined # default bgp_local_pref 0; # The same for local preference # source address 198.51.100.14; # What local address we use for the TCP connection # password "secret"; # Password used for MD5 authentication # rr client; # I am a route reflector and the neighor is my client # rr cluster id 1.0.0.1; # Use this value for cluster id instead of my router id # export where source=RTS_STATIC; # export filter { # if source = RTS_STATIC then { # bgp_community = -empty-; bgp_community = add(bgp_community,(65000,5678)); # bgp_origin = 0; # bgp_community = -empty-; bgp_community.add((65000,5678)); # if (65000,64501) ~ bgp_community then # bgp_community.add((0, 1)); # if bgp_path ~ [= 65000 =] then # bgp_path.prepend(65000); # accept; # } # reject; # }; #} but we cannot see that interface inside that routing table. we tried to force the traffic from/to that interface going to the right table as reported but without effect. * * *root@vpn1:~# ip rule show* *0: from all lookup local * *32764: from all oif eth1 lookup auslre * *32765: from all iif eth1 lookup auslre * *32766: from all lookup main * *32767: from all lookup default* * * Do you have any idea how to solve that? Probably we are struggling with the wrong configuration file bird.conf and we are not able to attached correctly an interface to a specific routing table. Thank you very much guys Bye Stefano

2 1

Does Bird accept configuration changes using the command line?
by 3.listas＠adminlinux.com.br 14 Nov '12

14 Nov '12

Hi, I have a network of ~ 200 servers with Quagga routers in the network core. I'm evaluating the impact of using the Bird in my routers and I have a question. In Quagga the administrator can change the configuration using a command line. For example, I have a shell script that changes the metric of a route in specific cases. Does Bird allow something similar? Does Bird accept configuration changes using the command line? Thanks -- Thiago Henrique www.adminlinux.com.br

3 4

Segfault on interface addition when not using the device protocol
by Apollon Oikonomopoulos 13 Nov '12

13 Nov '12

Greetings, I came across this issue while trying to debug (unsuccessfully so far) a probably unrelated segfault on one of our routers. When using the kernel protocol in a configuration without a device protocol, then adding a new interface with an address causes bird to segfault. The relevant backtrace for bird 1.3.8 (or 1.3.7 it's the same) is the following: #0 kif_choose_primary (i=0x5555557c8080) at krt.c:147 #1 0x0000555555564059 in ifa_recalc_primary (i=0x5555557c8080) at ../../nest/iface.c:469 #2 0x0000555555564d89 in ifa_update (a=0x7fffffffdfd0) at ../../nest/iface.c:546 #3 0x0000555555596534 in nl_parse_addr (h=<optimized out>) at netlink.c:545 #4 0x0000555555596a38 in nl_async_msg (h=0x5555557c4f70) at netlink.c:980 #5 nl_async_hook (sk=<optimized out>, size=<optimized out>) at netlink.c:1027 #6 0x0000555555590bed in sk_read (s=0x5555557c4e80) at io.c:1442 #7 0x0000555555591a09 in io_loop () at io.c:1647 #8 0x000055555555bf89 in main (argc=<optimized out>, argv=<optimized out>) at main.c:675 Steps to reproduce: 1. Use the following config: router id 192.0.2.1; protocol kernel { import none; export none; persist; scan time 20; } 2. Start bird bird -d -c bird.conf 3. Add a dummy interface ip link add crash type dummy 4. Add an IP address to the interface ip addr add crash 192.0.2.2/24 <bird segfaults> The issue is not present when the device protocol is defined in the configuration. I realize that the device protocol is virtually mandatory, but I still regard a segfault as a bug :). Regards, Apollon

1 0

RIP routes
by Gustavo Ponza 04 Nov '12

04 Nov '12

Hi Ondrej, here is a problem I am *not able* to solve: my main BIRD operation concerns with the OSPF routing I have with our CisarNet and Hamnet *hamradio* backbones and, mainly, with the 44.x.x.x/8 and 10.x.x.x/8 networks. Now, since I must distribute, by using the RIP2 protocol, *only* the 44.x.x.x/8 prefixes to another 'ham' gateway which is able *only* to understand and manage the RIP2 protocol, how to rule out the 10.0.0.0/8 prefixes from being distributed to it? And since other *spurious* public IP addresses are collected by the OSPF network, how to rule out those from being distributed, too? many thanks gus -- ----==-- _ 73, gustavo i0ojj ---==---(_)__ __ ____ __ --==---/ / _ \/ // /\ \/ / -=====/_/_//_/\_,_/ /_/\_\ A proud member of linux team An advanced CisarNet backbone TCP/IP server in Colleferro/Rome-JN61lr @ir0rm-7.ampr.org linked to Hamnet site at Nuernberg (Germany)

1 0

BGP not starting
by Thomas St-Pierre 01 Nov '12

01 Nov '12

Hi, I'm a new BIRD user so I may be missing something, if so please forgive me. I created a simple config for an ibgp neighbor: template bgp rr_client { local as XXXXX; rr client; import all; export all; } protocol bgp R1 from rr_client { neighbor 172.19.242.51 as XXXXX; } The other end is a cisco router with a similarly simple config. However when I start up bird I get "Waiting for 172.19.242.51 to become my neighbor", and nothing ever happens. bird doesn't start listening on 179 either. For information, both bgp endpoints are in the same /24 subnet (ie directly connected). However if I add multihop then it comes up and everything's great. Is there a reason why I have to use multi hop when the devices are on the same lan? Thanks! Thomas

2 2

Using arp as a routing protocol (yuck)
by Alex Bligh 01 Nov '12

01 Nov '12

I sort of hesitate to ask this as it veers well into the realms of 'disgusting'. I'd like to use bird to do routing via arp. This is in a scenario where economic constraints prohibit purchase of sensible equipment. What I'd like to do is: a) if a route appears (it will always be long subnet route0, 'advertise' it, by repeatedly promiscuously arp'ing for it. b) proxy arp for it c) (horrors) (a) and (b) for proxy ndp I'm taking it that the good taste of bird authors has prevented this misfeature to date. Would it be easy to hack in? -- Alex Bligh

1 0