The relay transport is shared: one connection per relay server carries the
streams of every peer using it. When it drops, each of those peers gets a
Disconnected verdict from evalConnStatus even when ICE is still carrying its
traffic, because peerUsesRelay comes from HasRelayAddress(), which only reports
that management offered relay servers, not that we are connected to one. The
guard answers Disconnected with the aggressive retry, so every peer starts
sending offers over signal for a transport that no offer can restore: the relay
client's own guard is what reconnects it.
Feed relayManager.Ready() into the status inputs and return PartiallyConnected
when ICE is up and the missing side is the shared transport. That is the
existing "one path works, the other does not" branch, which retries three times
and then hourly instead of walking the exponential ladder forever.
Peers are not left waiting for the hourly tick: when the transport comes back,
Manager.onServerConnected notifies srWatcher, the guard resets the ticker to
800ms and iceState.reset() clears the hourly mode.
The verdict is unchanged when the transport is up but this peer is unreachable
over relay - it may have moved to another server, and only an offer carries its
new relay address - and in force-relay mode, where relay is the only transport.
Avoid invalid disconnection notifications in case the closed race dials.
In this PR resolve multiple race condition questions. Easier to understand the fix based on commit by commit.
- Remove store dependency from notifier
- Enforce the notification orders
- Fix invalid disconnection notification
- Ensure the order of the events on the consumer side
- Clients now subscribe to peer status changes.
- The server manages and maintains these subscriptions.
- Replaced raw string peer IDs with a custom peer ID type for better type safety and clarity.
Fix WireGuard watcher related issues
- Fix race handling between TURN and Relayed reconnection
- Move the WgWatcher logic to separate struct
- Handle timeouts in a more defensive way
- Fix initial Relay client reconnection to the home server
This update adds new relay integration for NetBird clients. The new relay is based on web sockets and listens on a single port.
- Adds new relay implementation with websocket with single port relaying mechanism
- refactor peer connection logic, allowing upgrade and downgrade from/to P2P connection
- peer connections are faster since it connects first to relay and then upgrades to P2P
- maintains compatibility with old clients by not using the new relay
- updates infrastructure scripts with new relay service
riccardom