Cloud-flare WARP tunnels dies

Trying to get cloudflare-warp running, creating a tunnel between my home network and my WARP client on ios/Windows

Cloudflare-warp is installed on Proxmox LXC with Ubuntu 22.04 - Which runs my wireguard VPN as well.
Initially everything looks fine, tunnel goes up when I restart the service - at least when looking in Cloudflare, but shortly after goes down with no trace in logs, except this once in a while:
2024-03-08T08:36:25.239Z DEBUG warp::warp_service::network_change: Routes changed: NewNeighbour; Destination: 192.168.1.254;

ip route shows no route towards cloudflare, which I think is a bit strange if cloudflare thinks things are up.

I am trying to makwe my whole 192.168.0.0/24 network reachable from client devices.

Any recommendations on how to move on ?

Updated everything, now it looks stable.
I do have a tun interface, so don’t suspect that.

But doing a traceroute from my PC I get timeout, like traffic not reaching my Linux - Which still does not have a route towards Cloudflare.

Tracing route to 192.168.1.211 over a maximum of 30 hops

  1     9 ms     8 ms     8 ms  162.158.133.63
  2    11 ms    10 ms     8 ms  162.158.133.62
  3     *        *        *     Request timed out.

I have this in my server logs:

2024-03-08T09:05:16.357Z DEBUG main_loop: warp_settings::raw_settings: gateway_id=e..4 Using gateway ID
2024-03-08T09:05:16.358Z DEBUG warp_api::api_client: Method: Post, Path: v0/accounts/e..4/reg/2f..d/devicestate, CombinedDeviceSettings { account_id: "e..4", status: "Connected", mode: "tunnel_only", always_on: true, reg_id: "2..d", doh_subdomain: "e..4.cloudflare-gateway.com", switch_locked: false, client_version: "2024.2.62", client_platform: "linux", warp_metal: "none", warp_colo: "none", handshake_latency_ms: Some(9), estimated_loss: Some(0.0) }

How did you even get it running in an lxc? In an LXC on proxmox it always tells me disconnected, manual I had to spin up a VM.

Any insight on how you solved it? I cannot make it work either on Ubuntu/Proxmox

Will look it up when I have time.
Things are rock solid. Both LXC and VM.

Solved it. Added these to my container conf:

lxc.cgroup2.devices.allow: c 10:200 rwm
lxc.mount.entry: /dev/net dev/net none bind,create=dir

Thanks anyway!