2.0.30 network problems...?

Chris Evans (chris@ferret.lmh.ox.ac.uk)
Sun, 25 May 1997 15:02:50 +0100 (BST)

I recently reported a stack of sockets appearing to be stuck in CLOSE
under pre-2.0.31 (with 5 fixes as found on ftp.cistron.nl applied).

Well, a reversion to 2.0.30 (as coming with RH4.2) appears to have a
similar problem.... How long _should_ the CLOSE state hang around and what
does it mean?

When the timers hit 0, they jump back up to ~2.5seconds (was ~10 with our
CLOSE sockets under pre-2.0.31)

Note that the udp socket leakage has not shown itself under 2.0.30, _yet_.

tcp 0 0 ferret.lmh.ox.ac.uk:ftp pbms.demon.co.uk:1172 CLOSE
on (0.29/0)
tcp 0 0 ferret.lmh.ox.ac.uk:ftp pbms.demon.co.uk:2830 CLOSE
on (0.90/0)
tcp 0 0 ferret.lmh.ox.ac.uk:ftp pbms.demon.co.uk:1788 CLOSE
on (2.28/0)
tcp 0 0 ferret.lmh.ox.ac:telnet pbms.demon.co.uk:4878 CLOSE
on (1.33/0)

Please cc: any replies/fixes/patches :-) to me.

Chris