- Bird-users - Mailing Lists

Bird Daemon Performance and usage information
by Eloi Carbó Solé 22 May '17

22 May '17

Hello, I am working on my MSc. dissertation which is about improving existing Bird's integration with UCI configuration scheme for OpenWRT/LEDE firmwares (if someone is interested [0]). As part of my documentation I am trying to highlight Bird's performance against other well-known software as well as adoption rates in IXPs or other organisations. There is some information I already used in the past for the project that gave birth to this integration in OpenWRT (~2014) but I cannot find presentations/papers/graphics using the latest versions or news/announcements from IXPs or organisations using Bird. Right now I am referencing some presentations [1], news [2] and old data coming from years 2012-2015 but, is there any extra documents I could use as evidence of Bird's performance/adoption rate or latest presentations to the public? Best regards, -- Eloi Carbó [0] https://github.com/eloicaso/bird-openwrt [1] https://people.netfilter.org/pablo/netdev0.1/slides/ondrej- zajicek-slides-netdev01.pdf [2] https://joinup.ec.europa.eu/community/osor/case/bird- manages-routing-worlds-largest-internet-exchanges

3 3

Re: [PATCH 2/4] Static protocol supports SADR
by Dean 22 May '17

22 May '17

On 05/22/2017 01:57 PM, Ondrej Zajicek wrote: > On Sun, May 21, 2017 at 11:01:35PM +0200, Dean Luga wrote: >> From: dean <dluga93(a)gmail.com> >> >> A new channel, sadr_ip6, is used for SADR both in the kernel >> and static protocols. In the static protocol, routes can be >> inserted with the following syntax: >> >> route <dst_prefix> from <src_prefix> via "interface" >> route <dst_prefix> from <src_prefix> recursive <ip> > Hi > > You should add parsing of SADR networks to conf/confbase.Y together with > other network types in net_ nonterminal symbol. That would add support > of SADR to static procotol automatically. > > Also, SADR-based recursive routes seems like a whole new can of worms. > And we are currently rethinking recursive routes internally. Perhaps > you could postpone changes to hostentry and stick with SADR recursive > routes dependent on non-SADR IGP tables, which fits to the current > framework. Note that this was updated in 2.0-pre1 so that for other > network types than IP4/IP6, igp tables are of type IP4/IP6. > Ok. How do I go about this? Do I make the changes and submit all 4 patches again? Or another patch with the config fix and reverting the recursive SADR routes?

2 1

Source Address Dependent Routing patch
by Dean 22 May '17

22 May '17

Hello, This is another try at extending the static and OSPFv3 protocols to support Source Address Dependent Routing (SADR), also called Source Specific Routing. This basically means that routing will take into account not only the destination address, but the source address as well. IPv6 subtrees are needed to support SADR in the kernel. They should be available for Linux kernel versions newer than 3.11. If they are not enabled, the kernel needs to be compiled with the option to enable them. The patches are based on bird-2.0.0-pre0.

3 6

Problem with BIRD importing routes from neighbor to master table
by Lyubomir Millanov 22 May '17

22 May '17

Hello, I'm having this problem with BIRD refusing to import a specific route into the routing table. This is strange because this is the session with the neighbor is not new and all of his other prefixes work fine. This is a newly added prefix. What I see: bird> show route all 185.200.156.0/22 185.200.156.0/22 via 185.1.40.2 on eth1 [as41059 2017-05-11] * (100) [AS61437i Type: BGP unicast univ BGP.origin: IGP BGP.as_path: 41313 16154 61437 BGP.next_hop: 185.1.40.2 BGP.med: 100 BGP.local_pref: 70 BGP.community: (0,42) (0,3856) (0,5483) (0,6939) (0,8400) (0,13335) (0,15169) (0,15958) (0,34772) (15169,13200) (16154,10000) (41059,2) (41059,65007) (41313,2) (49401,8877) (49401,49401) BGP.ext_community: (ro, 0, 197216) bird> sh route protocol as16154 all 185.200.156.0/22 via 185.1.40.12 on eth0 [as16154 2017-05-12] * (100) [AS61437i Type: BGP unicast univ BGP.origin: IGP BGP.as_path: 16154 61437 BGP.next_hop: 185.1.40.12 BGP.local_pref: 100 BGP.community: (15169,13300) (16154,10000) (41313,6699) (65000,208) For some reason the prefix is not installed in the master routing table even though it's accepted from the neighbor. We use IN-filtering based on queries with irr tool and those queries are correct - the prefix is correctly announced. >From the above I can see that the neighbor prefix is learned after the first one (same prefix, learned from another path). I've cleared the bgp session with the neighbor without any change of the behavior. Is there a way to clear the route from the master routing table? I'd appreciate if someone can give me advice on the matter. Best regards, Lyubo

1 0

Re: Please help, local interfaces not being recognized as local
by Jerry Scharf 18 May '17

18 May '17

Ondrej, Thanks for the response. That fixed it. jerry On 5/18/17 3:10 AM, Ondrej Zajicek wrote: > On Wed, May 17, 2017 at 06:28:59PM -0700, Jerry Scharf wrote: >> Hi, >> >> So I have narrowed things down but still don't understand what is going on. >> The symptoms are as follows: >> >> I have local interfaces with either simple addresses on net or on vlans. The >> peer is on the local subnet of the interface. Some are ibgp and some are >> ebgp. If I set direct, they all fail and I don't see any listener on port >> 179. As soon as I change them to multihop, everything works fine. > Hi > > You are missing 'protocol device' which is necessary for BIRD to find out > network interfaces. > -- Soundhound Devops "What could possibly go wrong?"

1 0

Please help, local interfaces not being recognized as local
by Jerry Scharf 18 May '17

18 May '17

Hi, So I have narrowed things down but still don't understand what is going on. The symptoms are as follows: I have local interfaces with either simple addresses on net or on vlans. The peer is on the local subnet of the interface. Some are ibgp and some are ebgp. If I set direct, they all fail and I don't see any listener on port 179. As soon as I change them to multihop, everything works fine. Also, I have defined a bunch of local aliases on lo. I then create a static block that lists all the addresses as via "lo". When I show static, they all show as dormant and they are not being redistributed by the bgp speaker. This is on a centos 7 system currently stock configured runing 1.4.5 from epel. I had been doing things with systemd-networkd and alternate kernels, but I did a clean install now. I have also tried 1.6.0 bird/centos7 build with the same issues. If someone can help me debug this, I would really appreciate it. If not, I may have to abandon using bird in favor of another routing daemon. That is certainly not my first choice. thanks, jerry here's my current config (with multihop specified and bfd disabled. There are currently no other files being included.) # this is automatically generated from borderbird4.j2, version 0.3 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 none; # Default is import all export all; # Default is export none # kernel table 5; # Kernel table to synchronize with (default: mai n) } protocol direct { interface "lo", "*"; preference 32000; # make sure bgp routes don't overwrite direct } log "/var/log/birdlog" all; router id 10.100.16.10; protocol static { preference 16000; route 198.49.100.45/32 via "lo"; route 8.25.217.18/32 via "lo"; route 198.49.100.18/32 via "lo"; route 8.25.217.45/32 via "lo"; route 8.6.161.18/32 via "lo"; route 8.5.161.45/32 via "lo"; } function issecondary() { return net ~ [198.49.100.45,8.25.217.18,198.49.100.18,8.25.217.45,8.6.161.18 ,8.5.161.45]; } function isprimary() { return net ~ [0.0.0.4/30]; } function net_outside() { return net ~ [172.18.10.0/24+,8.5.161.0/24+,198.49.100.0/24+]; } function is_internal_local() { return net ~ [10.100.32.10/32,10.100.32.0/20,10.100.16.10/32,10.100.16.0/20] ; } filter inside_export { if ((net = 0.0.0.0/0) || (net ~ 10.0.0.0/8)) then accept; else reject; } filter inside_import { if is_internal_local() then reject; if net ~ [10.0.0.0/8+, 172.16.0.0/12+] then accept; else reject; } filter load_default { if net = 0.0.0.0/0 then accept; else reject; } filter outside_only { if isprimary() then bgp_local_pref=100; else if issecondary() then bgp_local_pref=80; else if net_outside() then bgp_local_pref=25; if net_outside() then accept; else reject; } template bgp tocore { description "bgp to data center as"; local as 65302; multihop; export filter inside_export; import filter inside_import; # bfd on; } protocol bgp inside_netA from tocore { neighbor 10.100.16.3 as 65302; source address 10.100.16.10; } protocol bgp inside_netB from tocore { neighbor 10.100.32.3 as 65302; source address 10.100.32.10; } template bgp toedge { description "bgp to edge/ISP as"; local as 64102; multihop; export filter outside_only; import filter load_default; allow local as; # bfd on; } protocol bgp outside_edge_A from toedge { neighbor 172.18.10.1 as 18741; source address 172.18.10.4; } protocol bgp outside_edge_B from toedge { neighbor 172.18.10.2 as 18741; source address 172.18.10.5; } protocol bfd { interface "eth*" { interval 300 ms; multiplier 3; }; } include "/soundhound/etc/bird/*4.part.cfg"; -- Jerry Scharf, Soundhound DevOps "What could possibly go wrong?"

2 1

Redistribute loopback in OSPF
by Quentin Ritoul 18 May '17

18 May '17

Hello, I'm starting to use Bird, and I try to redistribute loopback in ospf. (So it seems easy...) rt01 : interface ens18 > 10.0.4.5/24 interface loopback > 10.99.99.1/32 (and 127.0.0.1/8) rt02 : interface ens18 > 10.0.4.6/24 interface loopback > 10.99.99.100/32 (and 127.0.0.1/8) Below the configuration of rt01 and rt02 : rt01 : filter loopbacks { if (net = 10.99.99.1/32 ) then accept; else reject; } protocol ospf { import all; export filter loopbacks; area 0 { interface "ens18" { cost 5; type broadcast; hello 5; retransmit 2; wait 10; dead 20; authentication cryptographic; password "toto"; }; }; } rt02 : filter loopbacks { if (net = 10.99.99.100/32 ) then accept; else reject; } protocol ospf { import all; export filter loopbacks; area 0 { interface "ens18" { cost 5; type broadcast; hello 5; retransmit 2; wait 10; dead 20; authentication cryptographic; password "toto"; }; }; } But it doesn't work. Each router don't receive the route 10.99.99.X bird> sh route 0.0.0.0/0 via 10.0.4.1 on ens18 [kernel1 22:01:40] * (10) 10.0.4.0/24 dev ens18 [direct1 22:01:40] * (240) dev ens18 [ospf1 22:01:48] I (150/5) [10.0.4.5] Could you help me please ? Regards, Quentin

3 2

OSPFv3 flooding unknown LSAs
by Dean 17 May '17

17 May '17

Hello, I am using bird 2.0.0 (from the int-new branch) and I noticed that OSPF is flooding unknown LSAs. According to RFC 5340, unknown LSAs shouldn't be flooded if the U-bit isn't set. But the lsa_flooding_allowed function doesn't check the U-bit or whether the LSA has an unknown type. Is that a mistake or did I miss something? Regards, Dean Luga.

1 1

are my posts working?
by Jerry Scharf 17 May '17

17 May '17

Hi, Sorry to spam the list and people, but I have tried to send a couple messages to the list with nothing showing up. jerry -- Jerry Scharf, Soundhound DevOps "What could possibly go wrong?"

2 1

not seeing statics being advertised to bgp neighbors
by Jerry Scharf 16 May '17

16 May '17

Hi, I have the following setup and am not seeing what I expect. I had this working before, so I am sure it is something I have changed in the configurations when automating them. I just can't see what it is. Situation is that that have a bunch of host aliases (/32) defined on the lo interface. For each of these, I create a static in bird so there is a route. Then I filter and forward those via BGP to the edge routers that talk to ISPs. A single machine is presenting a number of addresses in different routing blocks and addresses. here is the list of loopback aliases [root@pnpborder1 bird]# ip addr show lo 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet 198.49.100.18/32 brd 198.49.100.18 scope global lo valid_lft forever preferred_lft forever inet 8.25.217.45/32 brd 8.25.217.45 scope global lo valid_lft forever preferred_lft forever inet 8.6.161.18/32 brd 8.6.161.18 scope global lo valid_lft forever preferred_lft forever inet 8.5.161.45/32 brd 8.5.161.45 scope global lo valid_lft forever preferred_lft forever inet 198.49.100.45/32 brd 198.49.100.45 scope global lo valid_lft forever preferred_lft forever inet 8.25.217.18/32 brd 8.25.217.18 scope global lo valid_lft forever preferred_lft forever inet 4.7.20.4/32 brd 4.7.20.4 scope global lo valid_lft forever preferred_lft forever inet 4.7.20.5/32 brd 4.7.20.5 scope global lo valid_lft forever preferred_lft forever inet 173.1.2.36/32 brd 173.1.2.36 scope global lo valid_lft forever preferred_lft forever inet 173.1.2.37/32 brd 173.1.2.37 scope global lo valid_lft forever preferred_lft forever inet6 dead:beef::1/32 scope global valid_lft forever preferred_lft forever inet6 ::1/128 scope host valid_lft forever preferred_lft forever the routing table doesn't show the routes and the neighbors don't get them (did a tcpdump to make sure it wasn't filtering on the far side.) [root@pnpborder1 bird]# netstat -nr Kernel IP routing table Destination Gateway Genmask Flags MSS Window irtt Iface 0.0.0.0 - 0.0.0.0 ! - - - - 4.7.20.0 0.0.0.0 255.255.255.240 U 0 0 0 edge10Bvl801 4.7.20.0 0.0.0.0 255.255.255.240 U 0 0 0 edge10Avl801 10.100.16.0 0.0.0.0 255.255.240.0 U 0 0 0 eth4 10.100.32.0 0.0.0.0 255.255.240.0 U 0 0 0 eth5 10.201.0.0 0.0.0.0 255.255.240.0 U 0 0 0 eth0 169.254.0.0 0.0.0.0 255.255.0.0 U 0 0 0 eth0 172.18.10.0 0.0.0.0 255.255.255.0 U 0 0 0 edge10Bvl800 172.18.10.0 0.0.0.0 255.255.255.0 U 0 0 0 edge10Avl800 173.1.2.32 0.0.0.0 255.255.255.240 U 0 0 0 edge10Bvl803 173.1.2.32 0.0.0.0 255.255.255.240 U 0 0 0 edge10Avl803 can anyone spot what I did wrong? thanks, jerry config: # this is automatically generated from borderbird4.j2, version 0.3 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 none; # Default is import all export all; # Default is export none # kernel table 5; # Kernel table to synchronize with (default: main) } protocol direct { interface "lo", "*"; preference 32000; # make sure bgp routes don't overwrite direct } log "/var/log/birdlog" all; router id 10.100.16.10; protocol static { preference 16000; route 198.49.100.45/32 via "lo"; route 8.25.217.18/32 via "lo"; route 198.49.100.18/32 via "lo"; route 8.25.217.45/32 via "lo"; route 8.6.161.18/32 via "lo"; route 8.5.161.45/32 via "lo"; } function issecondary() { return net ~ [198.49.100.45,8.25.217.18,198.49.100.18,8.25.217.45,8.6.161.18,8.5.161.45]; } function isprimary() { return net ~ [0.0.0.4/30]; } function net_outside() { return net ~ [172.18.10.0/24+,8.5.161.0/24+,198.49.100.0/24+]; } function is_internal_local() { return net ~ [10.100.32.10/32,10.100.32.0/20,10.100.16.10/32,10.100.16.0/20]; } filter inside_export { if ((net = 0.0.0.0/0) || (net ~ 10.0.0.0/8)) then accept; else reject; } filter inside_import { if is_internal_local() then reject; if net ~ [10.0.0.0/8+, 172.16.0.0/12+] then accept; else reject; } filter load_default { if net = 0.0.0.0/0 then accept; else reject; } filter outside_only { if isprimary() then bgp_local_pref=100; else if issecondary() then bgp_local_pref=80; else if net_outside() then bgp_local_pref=25; if net_outside() then accept; else reject; } template bgp tocore { description "bgp to data center as"; local as 65302; direct; export filter inside_export; import filter inside_import; bfd on; } protocol bgp inside_netA from tocore { neighbor 10.100.16.3 as 65302; source address 10.100.16.10; } protocol bgp inside_netB from tocore { neighbor 10.100.32.3 as 65302; source address 10.100.32.10; } template bgp toedge { description "bgp to edge/ISP as"; local as 18741; # direct; export filter outside_only; import filter load_default; allow local as; bfd on; } protocol bgp outside_edge_A from toedge { neighbor 172.18.10.1 as 18741; source address 172.18.10.4; } protocol bgp outside_edge_B from toedge { neighbor 172.18.10.2 as 18741; source address 172.18.10.5; } protocol bfd { interface "eth*" { interval 300 ms; multiplier 3; }; } include "/soundhound/etc/bird/*4.part.cfg"; -- Jerry Scharf, Soundhound DevOps "What could possibly go wrong?"

1 0