home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!dtix!darwin.sura.net!wupost!waikato.ac.nz!comp.vuw.ac.nz!canterbury.ac.nz!cantua.canterbury.ac.nz!cantsc.canterbury.ac.nz!cctr127
- From: cctr127@cantsc.canterbury.ac.nz (Jason Haar)
- Newsgroups: comp.infosystems.gopher
- Subject: Re: Gopher, Ultrix and netstat
- Message-ID: <BuHyKz.6FL@cantua.canterbury.ac.nz>
- Date: 13 Sep 92 03:19:47 GMT
- References: <1992Sep12.160620.8052@bernina.ethz.ch>
- Reply-To: j.haar@csc.canterbury.ac.nz
- Organization: University of Canterbury, Christchurch, New Zealand.
- Lines: 46
- Nntp-Posting-Host: cantsc.canterbury.ac.nz
- X-Newsreader: Tin 1.1 PL3
-
- karrer@bernina.ethz.ch (Andreas Karrer) writes:
- >
- > is this a bug in the TCP implementation that does not properly close a
- > TCP connection after a timeout?
- >
- > i've seen similar things with nntp connections originating from a
- > particular (broken?) newsreader, on an ultrix machine.
- >
- > anyway, i think a TCP server should be able to close a TCP connection
- > cleanly even with mis-behaving clients.
-
- It's a problem with TCP/IP - although some developers claim it's a feature
- - not a bug ;-)
-
- To kill them off, either:
-
- a) reboot ;-))
-
- or
-
- b) follow this arcane advise I got from a mailing-list (it works at least
- under SunOS 4.1.*)
-
- do a 'netstat -A'
-
- >connection, as evidenced by the output from netstat -A:
- >
- >ff68000c tcp 0 96 stis.70 192.139.2.152.15669 CLOSING
- >
- >Apparently, Matt Crawford (matt@oddjob.uchicago.edu) is the expert on
- >this, because he immediately sent me the following cryptic, but
- >effective, advice:
- >
- > adb -w /vmunix /dev/kmem
- > ff68000c+8/w 0
- > $q
- >
- > then wait up to 2 minutes.
-
-
-
- Hope that helps
-
- Cheers
-
- Jason Haar, Network Consultant
-