June 2025 - Bird-users - Mailing Lists

BIRD 3.1.0, 2.17, 3.0.2, 2.16.2 released
by Ondrej Filip 24 Aug '25

24 Aug '25

Dear BIRD Users, we have released four versions of BIRD at once; 2.17 and 3.1.0 with new features and 2.16.2 and 3.0.2 fixing bugs in 2.16.1 and 3.0.1, respectively. It may become common in the future to keep certain versions stable with only fixes for some longer time. Specifically, 3.1.0 and 2.17 have already successfully migrated to Debian testing, so these versions may get quite a long chain of updates. In 2.17 and 3.1.0, we added TCP-AO support for BGP on Linux. We haven't found any resource for TCP-AO in any flavor of BSD; if you know of any, please tell us. We also implemented graceful restart timer minima and maxima in BGP, format specification for link-local-only next hops in BGP, to improve compatibility with other routing stacks, and also we merged a new next hop option for IPv4 routes in Babel. In 3.1.0, we also added a compatibility layer for show route outputs, to allow your BIRD 2 automation to work with BIRD 3. Please tell us if you find any significant difference there. In all versions, we fixed several rare crashes in BFD and table removal, and other minor bugs. In 3.1.0 and 3.0.2, we also fixed quite some crashes in route propagation and state machines. Also setting route gateway in 3.x by filters is now fixed. It also looks like that some bugs in 3.1.0 slipped through our release process, most notably failure to build on FreeBSD 13 with Netlink and some strange crash on BGP neighbor flap. We are currently investigating that and we may release 3.1.1 soon. We apologize for for that. Anyway, the packages are as always in our repository at https://pkg.labs.nic.cz/ and on our website at https://bird.network.cz/ Have a pleasant spring or autumn, and happy routing! Ondrej and Maria

6 5

Relaxed handling of OTC attribute
by André Grüneberg 26 Jun '25

26 Jun '25

Hi Bird team, We are currently evaluating the implementation of RFC9234 for our IXP route servers. Looking at it naively, one just needs to set local role rs_server in the protocol. And indeed, routes from peers will be rejected and this is logged. Instead of just logging, we would really like to apply our "blame and shame" policy, i.e. make the invalid routes (in our case, anything with an OTC set) visible in our looking glass (similar to RPKI invalids). To do so, we'd need the "ineligible" routes to be imported into the main table, tagged in a sensible way. I understand that RFC9234 section 5 mandates that the behaviour wrt OTC attribute handling shall not be configurable by the operator. But ineligible does not require the route to be invisible (see section 3). Would it be possible to implement a more relaxed behaviour by allowing the import of ineligible routes (but never export)? Our current alternative is to avoid using BGP roles capability, but only implement OTC handling in filters. Thanks and best wishes, André -- André Grüneberg, Managing Director andre.grueneberg(a)bcix.de +49 30 2332195 42 BCIX Management GmbH Albrechtstr. 110 12103 Berlin Germany Geschäftsführer/Managing Directors: Jens Lietzmann, André Grüneberg Handelsregister: Amtsgericht Charlottenburg, HRB 143581 B

7 16

Set src address of the route
by Max Tulyev 19 Jun '25

19 Jun '25

Hello, is it possible for bird (v 2.x) to manually set in the routing table in Linux the source address of the route came via BGP? i.e. I got the route 192.168.0.0/16 with nexthop 10.1.1.1, and need to export to the kernel something like this: "ip ro add 192.168.0.0/16 via 10.1.1.1 dev eth0.5 proto bird SRC 172.16.0.1" The real life example: the bird router connected to Internet Exchange point, learned the route from the peer or the route server. When something from this router wants to connect to something within this route, the source address of this connection will be the address of the Internet Exchange interface, which is not globally routed, and connection will fail. Any ideas? Thank you! -- WBR, Max Tulyev MT6561-RIPE - Jabber maxtul(a)netassi.st - Phone +447441953543 - Telegram @mt6561

3 2

Bird from Debian/Ubuntu repo or from sources?
by Mike Neo 16 Jun '25

16 Jun '25

Hi, for simple BGP routing with two peers (Global + IX). What is better to choose? Bird from Debian/Ubuntu repositories or from sources? Kind regards, Mike

2 2

[3.1.2 and older] birdc doesn't output the possible command expansions list
by Артём Варакса 16 Jun '25

16 Jun '25

Hello, I've observed an issue with `bird` 3.1.2 that is also reproducible in 2.17.1, but not with 2.15.1 (i. e. a regression somewhere between 2.15.1 and 2.17.1). After starting the `bird` service, on 2.17.1 and newer: > # birdc > bird> show stat > Ambiguous command, possible expansions are: On 2.15.1, the output is correct: > # birdc > bird> show stat > Ambiguous command, possible expansions are: > show status Show router status > show static [<name>] Show details of static protocol > bird> This was tested on ALT Linux (downstream issue: https://bugzilla.altlinux.org/54787) Per the maintainer, it's likely an issue in `bind` itself rather than the package, hence reporting here. Thanks!

4 4

prefer ipv6 examples in the documentation
by Alexander Zubkov 06 Jun '25

06 Jun '25

Hi everybody! In spite or recent Maria's activity, I decided to checked the BIRD documentation and found many examples where legacy :) IPv4 addresses are used. My proposal is to preferably use IPv6 examples in the documentation. I tried to spot the places where IPv4 examples can be replaced or complemented by IPv6 examples, and prepared a patch with possible changes. Regards, Alexander Zubkov

4 4

[3.1.2] Queuing an already queued event to another queue is not supported.
by Matt Corallo 03 Jun '25

03 Jun '25

Tried to test 3.1.2 and it mostly seems to work okay, but after some runtime I got the above error and crash. Sadly don't have a lot more to help debug this, though it may have been during a reload. Thanks, Matt

4 8

Correct replacement for 'birdc reload in all' in BIRD 3.1?
by Robert Scheck 02 Jun '25

02 Jun '25

Hello, I'm currently running the following commands in a systemd.service as part of a systemd.timer: - rpki-client -B -c -j -m -o - bird -p -c /var/lib/rpki-client/bird - birdc configure soft - birdc reload in all Since upgrading from BIRD 2.16 to 3.1, the last command fails. However, https://gitlab.nic.cz/labs/bird/-/blob/stable-v3.1/doc/migration-bird3.md mentions just: > "Reload of filters is now done by reload filters command, contrary to > just reload in BIRD 2. Am I overlooking something? What's the correct replacement command for BIRD 3.1? Regards, Robert

2 1

[3.1.1 on EL8] CPU 100% with NetworkManager
by zx＠zxinc.org 02 Jun '25

02 Jun '25

Hi there, When I upgrade bird from 2.16.2 to 3.1.1 today, it became unstable that it often causes CPU usage to 100% for bird, and later also for NetworkManager, and there are some notable syslogs: 6月 02 16:04:28 zx-router bird[2197971]: I/O loop cycle took 14046.255 ms for 43 events 6月 02 16:04:48 zx-router bird[2197971]: I/O loop cycle took 15307.277 ms for 42 events 6月 02 16:04:57 zx-router bird[2197971]: I/O loop cycle took 8562.921 ms for 43 events 6月 02 16:05:16 zx-router bird[2197971]: I/O loop cycle took 9799.271 ms for 85 events 6月 02 16:05:25 zx-router bird[2197971]: I/O loop cycle took 9330.265 ms for 80 events 6月 02 16:05:30 zx-router bird[2197971]: I/O loop cycle took 5004.867 ms for 44 events 6月 02 16:06:09 zx-router bird[2197971]: I/O loop cycle took 38759.438 ms for 84 events 6月 02 16:06:17 zx-router bird[2197971]: I/O loop cycle took 5465.396 ms for 6 events 6月 02 16:06:23 zx-router bird[2197971]: I/O loop cycle took 5906.858 ms for 44 events 6月 02 16:06:37 zx-router bird[2197971]: I/O loop cycle took 12896.032 ms for 9 events 6月 02 16:06:56 zx-router bird[2197971]: I/O loop cycle took 8217.969 ms for 81 events 6月 02 16:07:08 zx-router bird[2197971]: I/O loop cycle took 9012.775 ms for 25 events 6月 02 16:07:26 zx-router bird[2197971]: I/O loop cycle took 11506.570 ms for 45 events 6月 02 16:07:56 zx-router bird[2197971]: I/O loop cycle took 30014.580 ms for 85 events 6月 02 16:08:05 zx-router bird[2197971]: I/O loop cycle took 6333.289 ms for 6 events 6月 02 16:08:14 zx-router bird[2197971]: I/O loop cycle took 8751.816 ms for 42 events 6月 02 16:08:29 zx-router bird[2197971]: I/O loop cycle took 13017.064 ms for 46 events 6月 02 16:08:51 zx-router bird[2197971]: I/O loop cycle took 13223.357 ms for 80 events 6月 02 16:09:44 zx-router bird[2197971]: I/O loop cycle took 17730.328 ms for 5 events 6月 02 16:10:37 zx-router bird[2197971]: I/O loop cycle took 8061.950 ms for 8 events 6月 02 16:10:46 zx-router bird[2197971]: I/O loop cycle took 9008.065 ms for 65 events 6月 02 16:10:54 zx-router bird[2197971]: I/O loop cycle took 6035.965 ms for 4 events 6月 02 16:11:03 zx-router bird[2197971]: I/O loop cycle took 9959.243 ms for 67 events I'm exporting full table to kernel with NetworkManager running on RHEL8 related systems (CentOS/Alma/Rocky 8). There is no problem before upgrading.

4 4

Crash width Assertion '!rejected || !accepted' failed at nest/rt-table.c:1192
by Alexey Boldyrev 02 Jun '25

02 Jun '25

Backtrace: Apr 26 18:05:26 router bird[751]: Netlink: Invalid argument Apr 26 18:05:26 router bird[751]: Netlink: Invalid argument Apr 26 18:05:26 router bird[751]: ... Apr 27 02:09:23 router bird[751]: AS4200220008: Error: Hold timer expired Apr 27 02:10:59 router bird[751]: AS4200220008: Error: Hold timer expired Apr 27 13:14:03 router bird[751]: Assertion '!rejected || !accepted' failed at nest/rt-table.c:1192 Apr 27 13:14:03 router systemd-coredump[1345547]: [🡕] Process 751 (bird) of user 967 dumped core. Stack trace of thread 824: #0 0x0000777b9f380624 n/a (libc.so.6 + 0x97624) #1 0x0000777b9f326ba0 raise (libc.so.6 + 0x3dba0) #2 0x0000777b9f30e582 abort (libc.so.6 + 0x25582) #3 0x000065083a2cb01f n/a (/usr/bin/bird + 0xe001f) #4 0x000065083a248712 n/a (/usr/bin/bird + 0x5d712) #5 0x000065083a249117 n/a (/usr/bin/bird + 0x5e117) #6 0x000065083a21d0cf n/a (/usr/bin/bird + 0x320cf) #7 0x000065083a2cf838 n/a (/usr/bin/bird + 0xe4838) #8 0x000065083a21d0cf n/a (/usr/bin/bird + 0x320cf) #9 0x000065083a2d1ee7 n/a (/usr/bin/bird + 0xe6ee7) #10 0x0000777b9f37e708 n/a (libc.so.6 + 0x95708) #11 0x0000777b9f402aac n/a (libc.so.6 + 0x119aac) Stack trace of thread 751: #0 0x0000777b9f386be2 n/a (libc.so.6 + 0x9dbe2) #1 0x0000777b9f37ae33 n/a (libc.so.6 + 0x91e33) #2 0x0000777b9f37ae74 n/a (libc.so.6 + 0x91e74) #3 0x0000777b9f3f553e __poll (libc.so.6 + 0x10c53e) #4 0x000065083a1f1a14 n/a (/usr/bin/bird + 0x6a14) #5 0x0000777b9f310488 n/a (libc.so.6 + 0x27488) #6 0x0000777b9f31054c __libc_start_main (libc.so.6 + 0x2754c) #7 0x000065083a1f2565 n/a (/usr/bin/bird + 0x7565) Stack trace of thread 826: #0 0x0000777b9f386be2 n/a (libc.so.6 + 0x9dbe2) #1 0x0000777b9f37ae33 n/a (libc.so.6 + 0x91e33) #2 0x0000777b9f37ae74 n/a (libc.so.6 + 0x91e74) #3 0x0000777b9f3f553e __poll (libc.so.6 + 0x10c53e) #4 0x000065083a2d2020 n/a (/usr/bin/bird + 0xe7020) #5 0x0000777b9f37e708 n/a (libc.so.6 + 0x95708) #6 0x0000777b9f402aac n/a (libc.so.6 + 0x119aac) ELF object binary architecture: AMD x86-64 Apr 27 13:14:03 router systemd[1]: bird.service: Main process exited, code=dumped, status=6/ABRT Apr 27 13:14:03 router systemd[1]: bird.service: Failed with result 'core-dump'. Apr 27 13:14:03 router systemd[1]: bird.service: Consumed 3min 7.732s CPU time, 6.1M memory peak. Version: 3.1.0

4 3