| 26 Jun 2025 |
matthewcroughan @ 39c3 (DECT 94667 or 97340 or 67192) | Nope, a reboot did not change this state | 13:40:26 |
matthewcroughan @ 39c3 (DECT 94667 or 97340 or 67192) | enp196s0 is still not attached to the bridge in the above config | 13:40:34 |
matthewcroughan @ 39c3 (DECT 94667 or 97340 or 67192) | bridge link still only shows tap0 | 13:40:52 |
antifuchs | Why are you dhcp’ing on the bridge participant interface instead of the bridge itself? That can’t result in a working config (this is unrelated, but also a problem) | 13:55:07 |
antifuchs | Anyway I recall something like this happened on one of my machine, I’ll go take a look | 13:56:07 |
adamcstephens | i don't think it's a violation to put the IP on the individual interface, but it is preferable to put it on the bridge for sure | 14:03:42 |
matthewcroughan @ 39c3 (DECT 94667 or 97340 or 67192) | Is there a way to set promiscuous mode via networkd? | 14:06:29 |
matthewcroughan @ 39c3 (DECT 94667 or 97340 or 67192) | Or do I have to make a systemd service to run ip link set dev foo promisc on | 14:06:57 |
| @zeromute:matrix.org left the room. | 14:20:52 |
antifuchs | hm, my issue was rooted in the interface not being identifiable by the matchConfig I gave (have switched to mac address matching now). That shouldn't affect you if the interface name you are matching corresponds correctly | 14:24:32 |
adamcstephens | you can use networkctl status <if> to see which networkd files it's matching | 14:25:20 |
| 27 Jun 2025 |
| R̴̨͕͇͍̞̮̐̅͆̌̀̉̐͋̈́̃̀͒́̎̅̚̚̚͠͝Ĕ̵̡̛͖͖̟̙̫̱͈̘̞̭͍͍͑̌̄͑̓̋̓̀̈̏̈́͊̇͊͆̉͂̏̀̃̚͘͝͝ͅͅD̶̡̢͔̱̖̮͙͉̘̺͓͍̩̮͈͍͗̃̀̏͌͘͜ͅŚ̸̬̭̯̬͙͇͓̬̩̳̤͚͓̤̩̺͉͖̉͛̓̿̎͊̿̆́̐͂̇͌̄̇̓͘ͅͅT̴̞̫̘̝͇͔̟̪̪̦͂̔̎̀̎ͅŎ̷̡̬̹̪͈̭̣͈̭̭͉̦̖̝̘̪͖͔̥̦̘̻̳Ṋ̶̛̫͈̳̘͚̜̔̋͆̅̈́͊̑͊̉̌̈́̾͑̈́̚ͅË̸̡̨̨̛͇̜̖͔͖̻̟̗̠̙͓̘̗̥͉͇̜͑͆͊͑͑̀̓͒͜͝͝ changed their display name from Redstone to R̴̨͕͇͍̞̮̐̅͆̌̀̉̐͋̈́̃̀͒́̎̅̚̚̚͠͝Ĕ̵̡̛͖͖̟̙̫̱͈̘̞̭͍͍͑̌̄͑̓̋̓̀̈̏̈́͊̇͊͆̉͂̏̀̃̚͘͝͝ͅͅD̶̡̢͔̱̖̮͙͉̘̺͓͍̩̮͈͍͗̃̀̏͌͘͜ͅŚ̸̬̭̯̬͙͇͓̬̩̳̤͚͓̤̩̺͉͖̉͛̓̿̎͊̿̆́̐͂̇͌̄̇̓͘ͅͅT̴̞̫̘̝͇͔̟̪̪̦͂̔̎̀̎ͅŎ̷̡̬̹̪͈̭̣͈̭̭͉̦̖̝̘̪͖͔̥̦̘̻̳Ṋ̶̛̫͈̳̘͚̜̔̋͆̅̈́͊̑͊̉̌̈́̾͑̈́̚ͅË̸̡̨̨̛͇̜̖͔͖̻̟̗̠̙͓̘̗̥͉͇̜͑͆͊͑͑̀̓͒͜͝͝. | 00:55:56 |
Zhaofeng Li | Kea 3.0 😮
https://gitlab.isc.org/isc-projects/kea/-/wikis/Release-Notes/release-notes-3.0.0 | 06:41:33 |
Zhaofeng Li | tl;dr:
Build system: Meson Native HTTP/TLS support: The DHCPv4, DHCPv6, and DHCP-DDNS daemons now have native support for API sockets over HTTP and TLS (HTTP), without the need to use the Control Agent (CA). Classification: Class tags: Option class-tagging allows an option value to be conditionally applied to the response based on the client's class membership. The effect is similar to using an if-block in ISC DHCP to conditionally include options at a given scope. [...] We have added the ability to choose from multiple occurrences of the same option based on class-tagging. IPv6-only preferred: Kea now supports v6-only-preferred mode, as described in RFC8925, more precisely. [...] Until recently, however, Kea sent back the option and assigned the IPv4 address at the same time; this is now fixed. Usability: Kea now prints options with binary data in a human-readable form whenever possible Build system improvements: Kea now compiles properly with Boost 1.87
| 06:41:38 |
Zhaofeng Li | * tl;dr:
- Build system: Meson
- Native HTTP/TLS support: The DHCPv4, DHCPv6, and DHCP-DDNS daemons now have native support for API sockets over HTTP and TLS (HTTP), without the need to use the Control Agent (CA).
- Classification: Class tags: Option class-tagging allows an option value to be conditionally applied to the response based on the client's class membership. The effect is similar to using an if-block in ISC DHCP to conditionally include options at a given scope. [...] We have added the ability to choose from multiple occurrences of the same option based on class-tagging.
- IPv6-only preferred: Kea now supports
v6-only-preferred mode, as described in RFC8925, more precisely. [...] Until recently, however, Kea sent back the option and assigned the IPv4 address at the same time; this is now fixed.
- Usability: Kea now prints options with binary data in a human-readable form whenever possible
- Build system improvements: Kea now compiles properly with Boost 1.87
| 06:42:08 |
Tom | Insbesondere das sie v6-only-prefered gefixt haben ist ja mal eine gute Nachricht :) | 06:51:53 |
Tom | * Especially that they've fixed v6-only-preferred is good. :) | 08:49:01 |
hexa | this is on my todo list, but the 2.6.3 update is still broken https://github.com/NixOS/nixpkgs/pull/411875 | 09:36:27 |
hexa | if someone wants to take a look | 09:36:32 |
| @psibi:matrix.org left the room. | 09:38:21 |
| nbp changed their display name from nbp to nbp — PTO. | 17:26:45 |
| 29 Jun 2025 |
antifuchs | I have a distinct memory that if you passed an NS server to the kernel via the command line, the scripted initrd had a working resolver config... but somehow now, it no longer does. I added a hack to create an /etc/resolv.conf in my system config, but that seems suboptimal. | 01:49:55 |
hexa | systemd-initrd ultras ❤️🔥 | 01:50:47 |
hexa | * ❤️🔥 systemd-initrd ultras ❤️🔥 | 01:50:52 |
hexa | * ❤️🔥 systemd-initrd ultras ❤️🔥 | 01:51:00 |
emily | please don't spend itme trying to get anything working with scripted initrd :p | 01:54:25 |
hexa | I wish we had proper deprecation tooling | 01:54:58 |
hexa | networking.interfaces needs to go 😄 | 01:55:25 |
hexa | nobody wants to maintain our own network setup stack | 01:55:40 |
hexa | someone should package up https://codeberg.org/liske/ifstate as an alternative | 01:55:47 |