lkml.org 
[lkml]   [1999]   [Jul]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: tcp or telnet problem?
"Mike A. Harris" wrote:

> On Sat, 24 Jul 1999, Joe Julian wrote:
>
> >Running redhat 6.0 with redhat's kernel 2.2.5-22.
> >
> >I telnet into my customer's box, they on a DSL line,
> >me on a dialup.
>
> YOUR computer is running RedHat 6.0? Or the customer's box is
> running RedHat 6, or both?

Customer's box is RH6, mine is RH 5.2 or win98 or solaris or sunos 5.3

> >I can login, and do most things just fine. But if I do anything
> >that produces more than about 1/3rd of a screen of info, my
> >output goes away. I can no longet see anything I type, nor any
> >output from any operation. A second telnet's ps shows that my
> >commands are being executed, but all output is going in the bit
> >bucket.
>
> This sounds like what happens in the Win95 telnet program that
> comes with Win95.

No, that will show a full screen of formatted info, then start trashing
things when
it gets to the bottom line if your stty setting is not synchronized to the
number of
lines you have the telnet client configured for and/or your TERM is not set
for vt100.

>
> >I kinda think that this is probably in the kernel
> >somewere because if I take the same output that would
> >kill a session, and instead output it to the other pty
> >(ie. "netstat -t >/dev/pty/1"), it kills that pty.
> >Again, that pty continues to accept input, but the
> >output is never found again.
>
> Hmmm.

And note that netstat does not do any screen formatting, so it's just
plain old scrolling text.

>
> >If i do any of that from their local network, it works
> >fine (a tcp latency problem?).
> >
> >Any ideas, or help with things to do to diagnose this
> >problem would be highly appreciated as I have 5 more
> >of these exact installations to do.
>
> More details about both machines please...

The Customer Machine: Concerto by Micro Star International, featuring
an ne2000 integrated network interface (10base-T) connected to the Fujitsu

DSL modem provided by the phone company, a PCI 3c905b connected to the
100Mbit hubs that form their internal network. That machine was running
RedHat 5.1 for about a week just to test the DSL (in that week the box
was hacked, grumble...) The hard drive was replaced (the customer knocked
the running machine off the desk, grumble again...) and so it is now
running
RedHat 6.0 with all the current updates to packages that are installed,
including the kernel update that fixes a security hole. Under 5.1 there
were
no problems (exept for the fact that 5.1 is as insecure as Linda Tripp's
phone
conversations.)

The remote machines vary, everything from an UltraSparc running sunos to
a K6-3 machine running win98. The one thing that still confuses the issue
to
me is that the phone company monitored the line, and during the packet
retries
shown in a separate post, they saw only an ARP packet on the Frame Relay.
My RH5.2 box never saw the packets, but if you connected a win95 box to
the crossover cable that goes to the DSL, it works perfectly. The DSL line
has been tested (3 times), and comes back clean. ifconfig shows no
dropped packets. My next project is to stick a hub in between the DSL
modem and the computer so that I can run a packet sniffer and see if I can
find out what is happening (any suggestions on a packet sniffing software
to use?)

Hmmm... Looking back at that paragraph... Why would it be doing the ARP
request if it already has the MAC address for the gateway?


-
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/

\
 
 \ /
  Last update: 2005-03-22 13:53    [from the cache]
©2003-2011 Jasper Spaans