Very Computer

I have two Linux boxs. The primary one is connected to the Internet
via a full time dial up connection. The secondary box is in a remote
location connected to the primary via a full time dial up. The
secondary is routed through the primary and onto the Internet.
Everything works OK except with occasional PPP lockups between the
primary and secondary Linux boxs. The dial up connection is never
broken, it simply stops passing data in either direction. By this I
mean I can no longer ping between the two machines. Another thing that
is interesting is if I use the secondary machine and telnet to the
primary, the primary knows someone is attempting a telnet session but
refuses to serve it because it can't identify the remote IP address.
I am running with Slackware and a kernel of 2.0.30. All my connections
are run through inetd. The PPP lockups can sometimes be as close as
four hours apart to many days apart. I have checked the messages and
syslog files looking for errors but there are none.

--
Dan

If you reply via email, remove the SS before dan at starboard dot com

hope that help some.

--
* anti spam address !!! do not reply to this message, use the

I've been having similar problems with a Red Hat 4.1 machine that I use
to Masquerade my net at home.  I use diald to bring the link up on
demand.  Most of the time it works great.  However, at random intervals
the PPP link stops working - much like Dan reports.

The strange thing, is that killing diald doesn't kill the PPP connection
when this happens.  Usually, pppd is killed as diald starts to come down.
Whenever the random lockup occurs, pppd stays active after diald quits.

I'm currently running pppd-2.2.0f, and diald-0.16 on kernel 2.0.30.

Anyone have any ideas?  My next step is to check if there is a new pppd out
the next time things break (assuming I don't find what is causing the
problem...).

        John

: I have two Linux boxs. The primary one is connected to the Internet
: via a full time dial up connection. The secondary box is in a remote
: location connected to the primary via a full time dial up. The
: secondary is routed through the primary and onto the Internet.
: Everything works OK except with occasional PPP lockups between the
: primary and secondary Linux boxs. The dial up connection is never
: broken, it simply stops passing data in either direction. By this I
: mean I can no longer ping between the two machines. Another thing that
: is interesting is if I use the secondary machine and telnet to the
: primary, the primary knows someone is attempting a telnet session but
: refuses to serve it because it can't identify the remote IP address.
: I am running with Slackware and a kernel of 2.0.30. All my connections
: are run through inetd. The PPP lockups can sometimes be as close as
: four hours apart to many days apart. I have checked the messages and
: syslog files looking for errors but there are none.

: --
: Dan

: If you reply via email, remove the SS before dan at starboard dot com

--
John Osmon                 Between the Velvet Lies,
Network Services             There's a truth that's hard as Steel.
Wright State University         - Holy Diver, Ronnie James Dio

I wrote bash script that pings the remote system. If there is no
response and the modem is still connected I capture the
output of ifconfig and cause the modem to disconnect.
The remote system knows the line is dropped and re-acquires.
I am assuming the ifconfig command will determine if there is indeed
a flow control problem. Problem is, the lockup hasn't happened in a
few days now.

#!/bin/sh
#
if [ -r /var/lock/LCK..ttyS1 ]; then
   TESTVAR=`ping -q -c 1 192.168.1.1 | grep -c '100% packet loss'`
   if [ $TESTVAR = "1" ]; then
      date >> /usr/sbin/ppp-ping-starboard.err
      /sbin/ifconfig >> /usr/sbin/ppp-ping-starboard.err
      TESTPPP=`cat /var/lock/LCK..ttyS1`
      if [ ! $TESTPPP = "" ]; then
         kill $TESTPPP
      fi
   fi
fi

--
Dan

If you reply via email, remove the SS before dan at starboard dot com