[Bug]: Dual stack peer's IPv6 kicks out their IPv4 every 30 seconds #3706
Description
Contact Details
What happened?
Hello there,
I was experiencing a lot of connection drops on my netmaker virtual network, and after a couple of hours of diagnosing, I think I've narrowed down the problem to peers that have a dual-stack IP (i.e. simultaneous IPv4 and IPv6)
I'm on the Community Edition, and my network topology is pretty simple; basically I have one publicly accessible gateway G, and several peers P0, P1, etc..., each of which is behind a regular NAT (no open ports). And my goal is to simply to group these NAT restricted peers under one virtual network, so that they can communicate with one another.
I configured my netmaker virtual network to only provide virtual IPv4s, and I disable the IPv6 toggle.
However, peers with a dual-stack IP have their connection drop nearly every 30 seconds for a span of 5-10 seconds.
Running wg show while the connection is down (checked via continuous pinging) reveals that the netmaker interface disappears during this downtime. And 5-10 seconds later, when a fresh new handshake is established, the wireguard netmaker interface returns (wg show).
Moreover, right before my peer dual-stack device disconnects, I notice on my dashboard that my peer device's IPv6 has been discovered, only for it to disappear again once downtime is over.
And surely enough, disabling the IPv6 option on my windows wifi network adapter fixes the issue.
So it seems like my IPv6 is kicking out my IPv4, because they both come from the same device (i.e. same public-key/device-id), and in response, the netclient on my device performs a fresh pull, only for the cycle to be repeated.
I can't expect my peers to globally disable their IPv6 network adapter, so I would really appreciate if you can look into this potential bug, or perhaps guide me if I'm setting up things incorrectly.
Thank you
Version
v1.1.0
What OS are you using?
Windows
Relevant log output
Contributing guidelines
- Yes, I did.