From owner-freebsd-current Fri Nov 10 07:51:13 1995 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id HAA29441 for current-outgoing; Fri, 10 Nov 1995 07:51:13 -0800 Received: from rocky.sri.MT.net (rocky.sri.MT.net [204.182.243.10]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id HAA29432 for ; Fri, 10 Nov 1995 07:51:08 -0800 Received: (from nate@localhost) by rocky.sri.MT.net (8.6.12/8.6.12) id IAA01901; Fri, 10 Nov 1995 08:53:28 -0700 Date: Fri, 10 Nov 1995 08:53:28 -0700 From: Nate Williams Message-Id: <199511101553.IAA01901@rocky.sri.MT.net> To: "Frank E. Terhaar-Yonkers" Cc: current@FreeBSD.org Subject: Re: ijppp wedges system on -stable In-Reply-To: <199511101345.IAA01569@robin.mcnc.org.mcnc.org> References: <199511101345.IAA01569@robin.mcnc.org.mcnc.org> Sender: owner-current@FreeBSD.org Precedence: bulk > I track -stable very closely. Lately I've had several system freezes > when running ijppp. This happens when the phone lines are especially > bad and I have to kill ijppp and restart it. Can't do anything except > hit the reset button. (Can't get to the debugger) I'm running an older -stable at home, but with a brand-new kernel and a slightly modified ppp to do 'ddial' mode (never give up and always re-dial when the line goes down). It hasn't locked up my system, but ppp dies alot when it can't re-connect the line. I have core dumps all over the place, but unfortunately I haven't been able to debug the problem. Are you *sure* the entire system is frozen up? On my box, I've got a getty setup to run on my outgoing ppp line, so when ppp dies I can dial in and re-start it remotely. It works great. > System is an old ISA (no name), dx2-66, 16MB, SCSI, 16550 uarts, > USR 28.8 (with upgraded firmware). This sounds like the same thing as my box, although I'm not running with an updated firmware on my USR 28.8. :( > BTW - I tried running regular PPP a while back but it seemed very broken. > I could establish a connection but almost no TCP traffic would get through. > Pings worked fine. That sounds like both ends aren't agreeing on compression. Try playing with /etc/ppp/options. Because user-mode ppp isn't as robust as the simple script I'm using now #!/bin/sh # # Re-dial when the lines goes down # -detach is necessary since the script wants to know when the pppd # process exits so it can re-try the connection # while ( 1 ); do pppd connection 'chat -v -f /etc/ppp/chat' -detach cuaa0 115200 sleep 15 done Nate