DNS query fails on large DNS answers

[nazarewk]

Describe the problem

Querying

To Reproduce

Steps to reproduce the behavior:

1. Set up the Network Resource pointing to a domain resource (along with Routing Peer)
2. try to resolve a large DNS record using Netbird's resolver (100.XXX.255.254 on MacOS/Windows, peer's Netbird IP on Linux)

The DNS query fails as soon as the answer exceeds wireguard interface's MTU value (1280).

Expected behavior

The DNS request succeeds.

It should be achieved with either of/and:

• the internal resolver should inform the Client about the truncated reply and the Client should retry the query over TCP
• (a wild guess, not an expert here) the internal resolver should support chunked queries with EDNS0

Are you using NetBird Cloud?

Yes

NetBird version

0.35+ (Network implemented)

Is any other VPN software installed?

no

Debug output

Already extensively reviewed tcpdump and trace level Netbird client logs on both sides (the client and the routing peer) on a live session to determine the root cause.

Additional context

To resolve a 1806 bytes sized A record setting up MTU to 2000 on the wt0 interface on the Routing Peer made the issue go away.

[lixmal]

Relevant changes

#3632
#3639