Hello, We use two timers: one for the idle case and one for the request/response started case. Both are required because the client may simply power off. Winsock normally does not timeout for reasons unknown to me...
Regards, SZ On 2/6/08, S.Korotky <[EMAIL PROTECTED]> wrote: > > Hello, All! > > It's a little bit offtopic but relates to an ICS-based software, and I > suppose ICS gooroos > may shed some light on this. > > Imagine you have a public web-server (runs on Windows 2003). From time to > time, > some strange client software opens a connection(s) on the server but never > issues > a http request. Such connections may stay open quite a while, even longer > than > 10 minutes which is equal or greater than usual timeout for sockets (to my > belief). > That happens several times a day. There is no any info about these clients > except for > ip-addresses. The only thing that can been deduced from the log is that > the same > ip-addresses are used (in other connections) by clients with IE browsers > (if we can > believe given values for the Agent header). > > Does anyone know what this means? > Why don't the dummy sockets get always closed by server OS upon a timeout? > I'd suppose > that there should be the same timeout for all such dummy connections, but > they live > different times, even for days, without data transfer, waiting until I > kill them explicitly > in backend UI. > > Thanks in advance. > > Best wishes, > Stanislav Korotky. > > -- > To unsubscribe or change your settings for TWSocket mailing list > please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket > Visit our website at http://www.overbyte.be > -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
