Can anybody help me with this problem that has to do with connection
handling?
Well, actually I just want to force an hanging TCP connection to shut
down.
But when a socket is in the FIN_WAIT1/2 state, sometimes I have to wait
up to 15 minutes for the socket to be closed by the system. This is very
annoying if I want to restart my server daemon, because I can't bind to
the port until all sockets are closed correctly.
There is no problem with the server, but sometimes the clients just cut
the connection without sending anything. So if I need to restart the
server (completely, in this case bind () necessary), bind () just fails
since not all sockets are closed, but I know that there can't /
shouldn't be any existing sockets, not even in any wait state, because
the server is startet manually again. Unfortunately I don't have any
handles to theses sockets any more if the server crashes before.
Since the kernel closes such connections after a 15 minute timeout or
so,
I just wondered how I could force to close connections in these
waitstates on startup of my server or whenever.
I don't want to shutdown eth0, so is there maybe a tool, that can kill a
connection like killing a process/task?
cu gerry
Gerhard Kutzelnigg
gerhard@pre.de
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu