[986] in linux-net channel archive
Re: rlogins AFTER telnet
daemon@ATHENA.MIT.EDU (Avery Pennarun)
Thu Aug 24 11:24:05 1995
Date: Wed, 23 Aug 1995 16:45:27 -0400 (EDT)
From: Avery Pennarun <apenwarr@foxnet.net>
To: Karl Keyte <kkeyte@esoc.esa.de>
cc: linux-net@vger.rutgers.edu
In-Reply-To: <9508211309.AA04521@art13.dpd.esoc.esa.de>
On Mon, 21 Aug 1995, Karl Keyte wrote:
[I removed the CC: to linux-kernel]
> Can anyone suggest why my rsh and rlogin commands to a 1.3.20
> linux box time out UNLESS I've recently done a telnet, ftp or
> similar to the machine. Telnets and ftps work no problem.
> What am I waking up by doing a telnet or ftp?
I've had this problem too, but since it seemed to start about the time I
upgraded my ARCnet driver on all my systems, I had blamed it on that and put
off looking into it.
The problem is really hard to isolate. I only have two computers locally
(486-33, kernel 1.2.13, and 486-40, kernel 1.3.20). As far as I can tell,
rcp/rsh only fails when the request is made FROM the 1.2.13 system TO
1.3.20. (rcp can be either copying TO or FROM the 1.2.13 system, as long as
the actual command is entered on 1.2.13)
Making requests from 1.3.20 is not a problem. 1.3.17 seems to work the same
way.
The reason it's so hard to track down is that if a telnet is open to the
remote system, the problem doesn't occur. Normally I would try (for
example) rcp'ing from my internet provider to my system, but in order to do
this I have to be logged into my internet provider via telnet, which defeats
the purpose :)
It seems really strange. Starting and closing a telnet will cause rcp to
work once or twice before giving up again. It may have to do with the
number of sockets currently existing. During testing, these two computers
are idle net-wise except for the rcp session, which may hide the problem on
busier sites.
I would do more testing, but with only two computers readily accessible it's
hard to tell what's going on. Hopefully this information will help, though.
Have fun.
Avery
