- Bird-users - Mailing Lists

Improper exports with BGP6 over IPv4
by Ave 31 Jul '20

31 Jul '20

Hello, I believe I stumbled upon a bug in BIRD 2.x while doing something that can only be summed up as "testing things on someone else's network with limited access provided to me". When peering over IPv4 yet exporting a full table, I am noticing that all exports, withdraw or update, are exported as withdrawals. Export side: Routes: 0 imported, 85839 exported, 0 preferred Route change stats: received rejected filtered ignored accepted Import updates: 0 0 0 0 0 Import withdraws: 0 0 --- 0 0 Export updates: 85840 0 0 --- 85840 Export withdraws: 0 --- --- --- 0 Import side: Routes: 0 imported, 0 exported, 0 preferred Route change stats: received rejected filtered ignored accepted Import updates: 0 0 0 0 0 Import withdraws: 85823 0 --- 85823 0 Export updates: 0 0 0 --- 0 Export withdraws: 0 --- --- --- 0 I've also tested with an instance where there were both updates and withdrawals exported, and as a result a sum of both were displayed as imported withdrawals. The configurations I've tested with are quite simple: Full import config (seems to import "correctly", but as exports seem to be faulty it all gets recorded as withdrawals): > log syslog all; > router id 192.168.4.22; > protocol device { > scan time 5; > } > > protocol bgp bird1 { > local 192.168.4.22 as 65529; > neighbor 192.168.4.1 as 65530; > multihop; > > ipv6 { > import all; > export none; > }; > } Trimmed export config that doesn't export correctly: > protocol bgp bird2 { > local 192.168.4.1 as 65530; > neighbor 192.168.4.22 as 65529; > multihop; > > ipv6 { > import none; > export all; > }; > } Things I've so far: - Doing BGP4 over IPv4 (renaming "ipv6" blocks to "ipv4" in configs). This had no issues, exports and imports were correct. - Doing BGP4 over IPv6 (doing same change as last point, plus changing IPv4 addresses to IPv6 ones in configs). This had no issues, exports and imports were correct. - Switching from IPv4 to IPv6 (changing IPv4 addresses to IPv6 ones in configs). This resulted in proper imports, hinting that problem only happens on BGP6 over IPv4. - Exporting to a non-bird2 BGP daemon. Specifically, I exported to a mikrotik router running ROS v6.46.6. It also classified these as withdrawals ("Withdrawn Received 85 979", with empty "Prefix Count" and "Updates Received"). Thanks and I hope I'm not just doing something wrong and wasting your time as a result, Ave

2 1

Binary Bird2 for EdgeRouter (EdgeOS, Debian 9, mips)
by Skyler Mäntysaari 31 Jul '20

31 Jul '20

Dear list, Is there somewhere a repository where I can find a mips version of Bird2? SID debian repository does not support mips architecture. Best regards, Skyler Mäntysaari

3 10

Re: Bird 1.6 support for CentOS 8
by Serhii Slipenko 31 Jul '20

31 Jul '20

Hello, Sorry for reply the mail. Does anybody know plans for Bird 1.6 and Centos 8 ? My bests, Serhii Slipenko From: Serhii Slipenko <serhii.slipenko(a)open-xchange.com> Date: Monday, 27 July 2020 at 09:39 To: <bird-users(a)network.cz> Subject: Bird 1.6 support for CentOS 8 Hello, I would like to ask you about BIRD 1.6 and support of that version on Centos 8 / Debian 10. Is there some notice in the “news” section? Does it going to be supported or there is some plans to drop the support for BIRD 1.6 in Centos 8? my bests, Serhii Slipenko

1 0

suse repository for bird 2.0.7
by Farhad 29 Jul '20

29 Jul '20

Hello, Is there bird version 2 in any public repository for suse linux? Regards, Farhad

2 4

syntax error
by Farhad 27 Jul '20

27 Jul '20

Dears, I am new to bird. I am trying to follow the examples here: https://bird.network.cz/?get_doc&v=16&f=bird-6.html#ss6.2 I get syntax errors for these lines: route 10.17.3.0/24 via 10.17.3.215 bfd; table testable; in the 'protocol static' section. protocol bfd { interface "eth1" { min rx interval 20 ms; min tx interval 50 ms; idle tx interval 300 ms; }; } protocol static { table testable; export all; import all; check link; route 10.17.3.0/24 via 10.17.3.215 bfd; } Regards, Farhad

2 2

avoid transit because of default route
by Julien Sansonnens 27 Jul '20

27 Jul '20

Dear list, I only export three prefixes I own to my peers and upstreams. In theory, nothing prevents one of my peers to choose my router as default route (without my consent), and to pass his packets through my network (only in output, therefore). How to prevent this abuse? It doesn't depend on the BGP layer, but on the IP kernel isn't it ? Is it common to filter incoming IP packets according to their destination? Cheers, julien

2 1

Bird 1.6 support for CentOS 8
by Serhii Slipenko 27 Jul '20

27 Jul '20

Hello, I would like to ask you about BIRD 1.6 and support of that version on Centos 8 / Debian 10. Is there some notice in the “news” section? Does it going to be supported or there is some plans to drop the support for BIRD 1.6 in Centos 8? my bests, Serhii Slipenko Solution Engineer

1 0

export filter matches, but not announced
by micah anderson 27 Jul '20

27 Jul '20

Hi, I have a peer configured with an export filter called 'bgp_nullroute', and if I do 'birdc show route filter bgp_nullroute' i see a netblock there. I thought that would mean that this netblock would be exported/announced to the peer, but its not being exported, but there is something that I do not understand, I would appreciate a nudge in the right direction! protocol bgp nullroute { description "nullroute server"; local as ASN; source address 199.254.238.1; neighbor 192.168.1.1 as 64512; multihop 255; import filter bgp_in_nullroute; export filter bgp_nullroute; } protocol static static_nullroute { import filter static_nullroutes; include "static_nullroutes.conf"; } file static_nullroutes.conf has: route 204.13.164.220/32 blackhole; define ASN = 16652; filter bgp_nullroute { if (ASN,666) ~ bgp_community && net.len = 32 && our_network() then { accept; } else reject; } bird> show route all filter bgp_nullroute 204.13.164.220/32 blackhole [static_nullroute 09:11:08] (200) Type: static unicast univ BGP.community: (16652,666) bird> show protocols all nullroute name proto table state since info nullroute BGP master up 2020-07-23 10:01:28 Established Description: nullroute server Preference: 100 Input filter: bgp_in_nullroute Output filter: bgp_nullroute Routes: 0 imported, 0 exported, 0 preferred Route change stats: received rejected filtered ignored accepted Import updates: 0 0 0 0 0 Import withdraws: 0 0 --- 0 0 Export updates: 802145 134 802011 --- 0 Export withdraws: 71 --- --- --- 0 BGP state: Established Neighbor address: 192.168.1.1 Neighbor AS: 64512 Neighbor ID: 192.168.1.220 Neighbor caps: refresh AS4 Session: external multihop AS4 Source address: 199.254.238.1 Hold timer: 159/180 Keepalive timer: 3/60 -- micah

3 5

Static routes, preferences and ordering
by William 24 Jul '20

24 Jul '20

Hi All, Discovered something interesting in bird2 2.0.7 (via bird2 package on Centos7, 2.0.7-1.el7) when retrying to use statics to override a default route coming via BGP. The following block would not see the higher preference route added (not even listed in 'show route protocol WANSTATICS'), but when commented out and config reloaded the gw via 172.16.4.250 would vanish, then come back when I restarted WANSTATICS. protocol static WANSTATICS { ipv4 { table WAN; }; route 0.0.0.0/0 via 172.16.1.254 { preference = 110; # override upstream }; route 0.0.0.0/0 via 172.16.4.250 { preference = 90; # prefer one to come from upstream via BGP }; check link on; } It was working on another sibling router node, but the two route blocks were reversed: protocol static WANSTATICS { ipv4 { table WAN; }; route 0.0.0.0/0 via 172.16.4.250 { preference = 90; # prefer one to come from upstream via BGP }; route 0.0.0.0/0 via 172.16.1.254 { preference = 110; # override upstream }; check link on; } Lo and behold, switch them around on the problematic node and suddenly everything behaves, 'show route protocol WANSTATICS' lists the new route and it propagates into the routing table too. There's nothing that I could see in the documentation to explain this, wondering if anyone else has seen something similar? Regards, William

2 1

Re: [BFD] BFD is Not working after system reboot.
by seunghwan hwang 23 Jul '20

23 Jul '20

Hello K. What does "disabled" setting mean? This is my configuration ================================== protocol bfd { multihop { min rx interval 300 ms; min tx interval 300 ms; multiplier 3; }; } protocol bgp as1 { import all; export all; local ~ as ~; neighbor ~ as ~; enable route refresh off; hold time 90; bfd on; } protocol bgp as2 { import all; export all; local ~ as ~; neighbor ~ as ~; enable route refresh off; hold time 90; bfd on; } ================================== Please check it. Regards. 2020년 7월 20일 (월) 오후 7:48, Kees Meijs | Nefos <kees(a)nefos.nl>님이 작성: > Hello Seunghwan, > > Haven't you got a "disabled" setting in de BIRD configuration file in > the protocol portion? > > Regards, > Kees > > On 20-07-2020 03:54, seunghwan hwang wrote: > > > > I've already tried capturing tcpdump. > > BIRD's BFD neighbor was captured sending packets, but the packets sent > > by BIRD were not captured. > > In this state, BIRD's BFD packets were captured after restarting BFD > > protocol. > > The BFD packet was captured Only after restarting the BFD protocol . > > Is there an issue with BFD Protocol? > > Please Check it. > > > >

1 0