>The cause of the high messageSendTime's in the current unstable is: >With PeerHandler, if we queue a message, we queue it to a node rather >than a connection. So we queue a message when no connections are free. >Then we wait for any connection to become free. Since they are all >sending trailers, and since all our attempts to open a connection fail, >either because of misconfiguration, severe overload, or deliberate >sabotage (in order to reduce the effect of vast numbers of incoming >connections), it remains on the queue. The reason it doesn't happen on >builds without PeerHandler (such as stable) is that we wouldn't queue >the message at all if all the ConnectionHandlers were sending trailers, >we would open a new connection THEN queue the message on that - a much >shorter time ON THE QUEUE.
Now, can this also be the thing that causes my node to OOM every single night (for the last x<10 builds)? /N _______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
