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
Connections should keep track of the last time they received a KEEPALIVE message from the other guy (handled on line 193 of UdpRemoteConnection) and call disconnect() if a keepalive hasn't been received in a certain amount of time.
Keepalives are sent every 7 seconds currently (TrConstants.UDP_KEEP_ALIVE_DURATION), so 30 seconds might be a good timeout.
The text was updated successfully, but these errors were encountered:
A disconnecting node which knows it's disconnecting because of TopologyMaintenance should also call a disconnect() session. This isn't happening as can be seen in LongTests.main().
If a node reaches its max peers, and is "changing" some of its peers because of a maintenance session, it should notify the peers that it's removing those peers so that they can in turn remove the "changer". Currently the node which is disconnected from doesn't know that it has been removed by the "changer" and may still try start sessions with the "changer".
Connections should keep track of the last time they received a KEEPALIVE message from the other guy (handled on line 193 of UdpRemoteConnection) and call disconnect() if a keepalive hasn't been received in a certain amount of time.
Keepalives are sent every 7 seconds currently (TrConstants.UDP_KEEP_ALIVE_DURATION), so 30 seconds might be a good timeout.
The text was updated successfully, but these errors were encountered: