Re: TCP hanging on CLOSE?
Romano Giannetti (romano@dea.icai.upco.es)
Thu, 18 Mar 1999 11:16:48 +0100
On Thu, Mar 18, 1999 at 07:18:25AM +0100, Andi Kleen wrote:
> romano@dea.icai.upco.es (Romano Giannetti) writes:
>
> > ...I had a router crash this morning, and fetchmail dropped some
> > connection for timeout. Now I have :
> >
> > Active Internet connections (w/o servers)
> > Proto Recv-Q Send-Q Local Address Foreign Address State
> > tcp 80 0 pern:2280 upco.es:pop CLOSE_WAIT
> > tcp 80 0 pern:2273 upco.es:pop CLOSE_WAIT
>
> Close_wait means it is waiting for the application to call close() - kill the
> process and it'll be fine.
>
> > tcp 0 0 pern:2243 upco.es:pop CLOSE
> > tcp 0 0 pern:2230 upco.es:pop CLOSE
>
> These should not happen (with zero SendQ).
>
> > [...20 more connection]
> > tcp 0 1 pern:2185 upco.es:pop CLOSE
> > tcp 0 1 pern:2184 upco.es:pop CLOSE
> > tcp 0 1 pern:2183 upco.es:pop CLOSE
> > tcp 0 1 pern:2182 upco.es:pop CLOSE
>
> Seems tcp_send_fin does not set a timer in some cases.
>
Well, possibly a bug in fetchmail 4.7.9: when I killed fetchmail,
immediatly all the connection were closed. The CLOSE_WAIT ones, the
CLOSE ones with zero SendQ, and the CLOSE ones with 1 byte SendQ.
I reported it to fetchmail people, too.
Romano
--
Romano Giannetti, Professor - Univ. Pontificia Comillas (Madrid, Spain)
Electronic Engineer - phone +34 915 422 800 ext 2410 fax +34 915 596 569
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/