Starting netbird client wipes out IPv6 default route

[sandroshu]

Before posting

• I searched existing discussions and issues, including closed ones, and checked the relevant docs.
• I believe this is a product bug rather than a configuration or setup question.
• I can reproduce this issue, or for intermittent issues I've included trigger, frequency, and timing details below.
• I removed or anonymized sensitive data from logs, screenshots, and configuration.

Affected area

Client / Agent, Routes / Exit nodes

Deployment type

Self-hosted - quickstart script

Operating system or environment

Linux

NetBird version and upgrade status

0.71.2

Did this work before?

Yes, this worked before

Regression details

-0.71.0 OK
-0.71.1 ??
-0.71.2 Broken

Summary

Starting netbird client/agent on debian removes the default ipv6 route. Never comes back.

Current behavior

Removes ipv6 default route.
After starting netbird service when I add the default route back manually it will work.

Expected behavior

Do not remove default route

Steps to reproduce

1. check ipv6 default route exists
2. start netbird
3. check routes again, does not exists

Environment and topology

No IPv6 default route configured in netbird, IPv6 group is NOT added to the client, client/agent does NOT have IPv6 address from netbird.
Same issue when IPv6 is enabled for the group, so this option does not change this issue.
I have both enabled in netbird because we use both.

Self-hosted details, if available

No response

Logs, status output, or debug evidence

root@access:~# ip -6 r l
2a00:xxx:xxx:xxx::/64 dev eth0 proto kernel metric 256 expires 5279sec pref medium
fde9:92ce:e042::/64 dev eth0 proto kernel metric 256 expires 5279sec pref medium
fe80::/64 dev eth0 proto kernel metric 256 pref medium
default via fe80::5e02:14ff:feb0:556e dev eth0 proto ra metric 1024 expires 1679sec hoplimit 64 pref medium

root@access:~# systemctl start netbird

root@access:~# ip -6 r l
2a00:xxx:xxx:xxx::/64 dev eth0 proto kernel metric 256 expires 5272sec pref medium
fde9:92ce:e042::/64 dev eth0 proto kernel metric 256 expires 5272sec pref medium
fe80::/64 dev eth0 proto kernel metric 256 pref medium

Related issues or discussions

No response

Impact

No response

Additional context

No response

[lixmal]

These kind of issues should be accompanied by a debug bundle.

Looking at the routes it's likely that the ra route was removed by the kernel because netbird enabled ipv6 forwarding: is this a routing peer?

If that's the case you should be able to fix it with sysctl -w net.ipv6.conf.eth0.accept_ra=2 for the time being

[sandroshu]

Thanks for your reply!

It is a routing peer, but only a single IPv4 network set.

I have added accept_ra=2 to sysctl. I'll reboot it later to see if the default route stays.

Debug bundle
Upload file key:
a770f3854575ec30551a636122833a695daeb5b000d0880c6b43e9a2e4ee6c01/797eb7c0-9f5b-4007-b118-8705b7efdead

[crntnvdl]

Same issue here, it also (and more importantly I think) breaks systemd-networkd-wait-online.service in my case which delays massively other services starting.
networkctl will show the that the ipv6 default gateway is wiped and the interface is stuck in configuring state making systemd-networkd-wait-online.service block until timeout.
Looking further I believe it is caused by the systemd-networkd config file created to avoid it wiping the routes on reload. By manually removing it the issue disappears.
This has been happening for some versions already, sorry I don't remember when it started, but I never had the time to actually diagnose the issue.