mediasoup: [Rust] Worker panics when running with multiple workers

Bug Report

mediasoup version: 0.15

I’m trying to run mediasoup with 2 (eventually more) workers per server (on a 2 core server, so one worker per core). After switching to 2 workers, I’ve been getting pretty common worker panics:

thread 'mediasoup-worker-cd9eec89-8672-4293-b4eb-3a857fdd0341' panicked at /usr/local/cargo/registry/src/index.crates.io-6f17d22bba15001f/mediasoup-0.15.0/src/router/data_consumer.rs:277:21:
Wrong message from worker: NotificationRef { handler_id: Ok("\u{c}\0\0\0\u{8}\0\u{e}\0\u{7}\0\u{8}\0\u{8}\0\0\0\0\0\0\u{4}\u{c}\0\0\0\0\0\u{6}\0\u{8}\0\u{4}\0\u{6}\0\0\0\u{4}\0\0\0q\0\0\0WRTC::WebRtcServer::OnStunDataReceived() | ignoring received STUN packet with unknown remote ICE usernameFragment\0\0\0"), event: Ok(DataconsumerBufferedAmountLow), body: Err(Error { source_location: ErrorLocation { type_: "Notification", method: "body", byte_offset: 36 }, error_kind: InvalidOffset }) }

thread 'mediasoup-worker-87b809a2-a4f9-4231-89a2-b4f5fe2cd40c' panicked at /usr/local/cargo/registry/src/index.crates.io-6f17d22bba15001f/mediasoup-0.15.0/src/router/data_consumer.rs:277:21:
Wrong message from worker: NotificationRef { handler_id: Ok("\u{c}\0\0\0\u{8}\0\u{e}\0\u{7}\0\u{8}\0\u{8}\0\0\0\0\0\0\u{4}\u{c}\0\0\0\0\0\u{6}\0\u{8}\0\u{4}\0\u{6}\0\0\0\u{4}\0\0\0q\0\0\0WRTC::WebRtcServer::OnStunDataReceived() | ignoring received STUN packet with unknown remote ICE usernameFragment\0\0\0"), event: Ok(DataconsumerBufferedAmountLow), body: Err(Error { source_location: ErrorLocation { type_: "Notification", method: "body", byte_offset: 36 }, error_kind: InvalidOffset }) }

thread 'mediasoup-worker-bb92335a-9b5b-4b48-b4bc-127543ac7041' panicked at /usr/local/cargo/registry/src/index.crates.io-6f17d22bba15001f/mediasoup-0.15.0/src/router/webrtc_transport.rs:479:21:
Wrong message from worker: NotificationRef { handler_id: Err(Error { source_location: ErrorLocation { type_: "Notification", method: "handler_id", byte_offset: 36 }, error_kind: InvalidUtf8 { source: Utf8Error { valid_up_to: 0, error_len: Some(1) } } }), event: Ok(TransportSctpStateChange), body: Err(Error { source_location: ErrorLocation { type_: "Notification", method: "body", byte_offset: 36 }, error_kind: InvalidOffset }) }

thread 'mediasoup-worker-6e6b7487-e501-42bf-9681-7450e79a0dfb' panicked at /usr/local/cargo/registry/src/index.crates.io-6f17d22bba15001f/mediasoup-0.15.0/src/router/webrtc_transport.rs:465:21:
Wrong message from worker: NotificationRef { handler_id: Err(Error { source_location: ErrorLocation { type_: "Notification", method: "handler_id", byte_offset: 36 }, error_kind: InvalidOffset }), event: Ok(WebrtctransportDtlsStateChange) }

and

thread 'mediasoup-worker-28a9b972-22da-4324-8e10-4284fbab7089' panicked at /usr/local/cargo/registry/src/index.crates.io-6f17d22bba15001f/mediasoup-0.15.0/src/worker/channel.rs:256:91:
called `Result::unwrap()` on an `Err` value: Error(Char { character: 'h', index: 29 })

The last one is weird because it seems unrelated to the others.

I’ve never seen any of these panics before switching to multiple workers, so I’m assuming it’s related to that. I also only see this in prod so it might be difficult to try and reproduce