You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When testing in relay-only mode we end up with very spammy DISCO messages blowing up the datagram sends queues in the magicsocket:
On each datagram being sent the NodeState::get_send_addrs discovers it has no UDP information so it requests a "full ping" which consists of pinging all the addresses, including the relay one as well as sending a call-me-maybe. So that is two extra DISCO messages being queued to send over the relay on top of the datagram being sent.
I'm not sure if this is only an artifact of the relay-only debugging mode or would also occur if hole punching fails. The relay-only debugging mode is far from perfect and does a lot of things an actual relay-only mode should not do.
The text was updated successfully, but these errors were encountered:
When testing in relay-only mode we end up with very spammy DISCO messages blowing up the datagram sends queues in the magicsocket:
On each datagram being sent the
NodeState::get_send_addrs
discovers it has no UDP information so it requests a "full ping" which consists of pinging all the addresses, including the relay one as well as sending a call-me-maybe. So that is two extra DISCO messages being queued to send over the relay on top of the datagram being sent.I'm not sure if this is only an artifact of the relay-only debugging mode or would also occur if hole punching fails. The relay-only debugging mode is far from perfect and does a lot of things an actual relay-only mode should not do.
The text was updated successfully, but these errors were encountered: