Squid not starting/stopping correctly
My new version of squid is not starting or stopping properly. However, it
seems to run OK.
ii squid 2.1.2-1Internet Object Cache (WWW proxy cache)
ii squid-cgi 2.1.2-1Squid cache manager CGI program
ii squidclient 2.1.2-1Command line URL extractor that talks to (a)
[EMAIL PROTECTED]:/usr/doc/modutils# /etc/init.d/squid start
Starting proxy server: squid.
[EMAIL PROTECTED]:/usr/doc/modutils# /etc/init.d/squid stop
Stopping proxy server: Waiting
. Failed.. done.
(Of course, it takes two minutes for the stop script to time out.)
Adding the -x option to the top of the script to show the commands executed:
[EMAIL PROTECTED]:/usr/doc/modutils# /etc/init.d/squid start
+ PATH=/bin:/usr/bin:/sbin:/usr/sbin
+ DAEMON=/usr/lib/squid/RunCache
+ SQUID=/usr/sbin/squid
+ LIB=/usr/lib/squid
+ PIDFILE=/var/run/runcache.pid
+ SPIDFILE=/var/run/squid.pid
+ SQUID_ARGS=-D -sNY
+ '[' -x /usr/sbin/squid ']'
+ echo -n 'Starting proxy server: '
Starting proxy server: + start
+ cd /var/spool/squid
+ start-stop-daemon --quiet --start --pidfile /var/run/runcache.pid --exec
/usr/lib/squid/RunCache -- -D -sNY
+ sleep 2
+ echo squid.
squid.
+ exit 0
[EMAIL PROTECTED]:/usr/doc/modutils# date [exit=0] [tty=p2] [
6:21PM]
Thu Feb 4 18:22:06 PST 1999
Thu Feb 4 18:22:06 PST 1999
[EMAIL PROTECTED]:/usr/doc/modutils# ls -l /var/run/{runcache,squid}.pid
-rw-r--r-- 1 root root5 Feb 4 18:21 /var/run/runcache.pid
-rw-r--r-- 1 root proxy 5 Feb 4 18:21 /var/run/squid.pid
[EMAIL PROTECTED]:/usr/doc/modutils# cat /var/run/runcache.pid
1906
[EMAIL PROTECTED]:/usr/doc/modutils# cat /var/run/squid.pid
1912
[EMAIL PROTECTED]:/usr/doc/modutils# ps aux|egrep '1906|1912'
proxy 1912 1.6 3.9 6896 5008 p2 S N 18:21 0:01 squid -D -sNY
root 1906 0.1 0.6 1640 856 p2 S N 18:21 0:00 sh /usr/lib/squid/Run
root 1944 0.0 0.2 848 348 p2 S N 18:23 0:00 egrep 1906|1912
[EMAIL PROTECTED]:/usr/doc/modutils# /etc/init.d/squid stop
+ PATH=/bin:/usr/bin:/sbin:/usr/sbin
+ DAEMON=/usr/lib/squid/RunCache
+ SQUID=/usr/sbin/squid
+ LIB=/usr/lib/squid
+ PIDFILE=/var/run/runcache.pid
+ SPIDFILE=/var/run/squid.pid
+ SQUID_ARGS=-D -sNY
+ '[' -x /usr/sbin/squid ']'
+ echo -n 'Stopping proxy server: '
Stopping proxy server: + stop verbose
+ start-stop-daemon --stop --quiet --pidfile /var/run/runcache.pid
+ start-stop-daemon --stop --quiet --pidfile /var/run/squid.pid --exec
/usr/sbin/squid
+ sleep 2
++ pidof -o %PPID -o 1945 /usr/sbin/squid
+ '[' '894 893 892 891 890 865 864 863 862 861 836 835 834 833 832 807 806 805
804 803 778 777 776 775 774 749 748 747 746 745 653 652 651 650 649 624 623 622
621 620 595 594 593 592 591 566 565 564 563 562 537 536 535 508 507 506 505 504
468 467 466 465 464 534 533 413 412 411 410 384 383 382 381 380 409 346 345 344
343 342 289 288 287 286 285 254 253 252 251 250 182 181 180 179 178' '!=' '' ']'
+ echo -n 'Waiting .'
Waiting .+ cnt=0
++ pidof -o %PPID -o 1945 /usr/sbin/squid
+ '[' '894 893 892 891 890 865 864 863 862 861 836 835 834 833 832 807 806 805
804 803 778 777 776 775 774 749 748 747 746 745 653 652 651 650 649 624 623 622
621 620 595 594 593 592 591 566 565 564 563 562 537 536 535 508 507 506 505 504
468 467 466 465 464 534 533 413 412 411 410 384 383 382 381 380 409 346 345 344
343 342 289 288 287 286 285 254 253 252 251 250 182 181 180 179 178' '!=' '' ']'
++ expr 0 + 1
+ cnt=1
+ '[' 1 -gt 60 ']'
+ sleep 2
etc.
--
-Mike Orr, [EMAIL PROTECTED]
Kernelversion command missing
My computer no longer has the kernelversion command, which I used to set the routing in /etc/init.d/network, since it's different in kernel 2.0.x vs 2.1/2.2. The command is not listed in the current Contents-i386.tgz, although the manual page is. Did the command accidentally get dropped out of the package, or is it called something else now? Which package is it supposed to be in? -- -Mike Orr, [EMAIL PROTECTED]
Exim configuration for dialup users (was Re: Exim as default mailer)
On Wed, Nov 11, 1998 at 11:38:16AM +0100, Paul Slootman wrote: > In article <[EMAIL PROTECTED]> you write: > > >As I have a single-user home PC with dial-up access to my ISP, I fall > >squarely into the category of users for whom Exim is "not particularly > >well-suited". Is this a real problem, or is Exim a good mailer for > >dial-up systems? > > I've created two exim.conf files, one for use when I'm online, and > one for when I'm offline. I switch these around via > /etc/ppp/ip-*.d/exim: > I have "queue_remote" in my exim.conf.autodial which means it doesn't > attempt to deliver these directly. Just either don't run the cron job > that flushes the queue periodically if you don't want connections to > happen solely for email delivery (or if you don't have dial-on-demand!). Exim works fine for home users, but there are five areas that are tricky to configure: ** "local_domans" should have ALL the names your local network is known by; e.g., "local_domains = townhouse:localhost". ** don't use the "visible_name" stuff or you'll find that messages to people at your ISP get delivered locally. ** I have all the queue options commended out (queue_smtp, queue_remote, queue_only, queue_run_in_order). Mail delivers immediately, but if it can't (e.g., network is down) it just gets queued, no problem. I run the queue manually when I want to, but you could also do it within ip-up or via cron. If you have dial-on-demand and don't want the link going up for email, edit the diald config file to ignore SMTP (and you'd also need to specify your smarthost via IP#, so it doesn't cause a DNS lookup, which would activate diald). ** You'll need to use your ISP's smarthost, because some sites will reject your mail if you try to deliver it directly, because your local domain name (that exim knows about) is different from the name associated with the (dynamic) IP you're currently connected on, so exim gets the HELO argument wrong and I don't know how to fix it. # In the ROUTERS section of /etc/exim.conf smarthost: driver = domainlist transport = remote_smtp route_list = "* smtp.my_isp.com bydns_a" end ** Rewriting the From: address takes forever to get right. I'm currently using: [EMAIL PROTECTED] [EMAIL PROTECTED] Ff If this is absent, you'll end up being From: [EMAIL PROTECTED], which is not accessible from Internet. Other misconfigrations cause [EMAIL PROTECTED] (somebody else, or a nonexistent user) or [EMAIL PROTECTED] (which your ISP will be real happy about). However, at least it's easier to get this right under Exim than any other MTA. ** It's difficult to funnel multiple local users through one ISP pop account. I've never gotten this to work. Some people use the comment part of the to-address to determine who it's going to. Others use the "-suffix" method, where mail to [EMAIL PROTECTED] goes through your .forward filter based on "-suffix". But I think that requires Qmail and cooperation from your ISP. As another person stated, the problem is really SMTP's fault --- it's inadequate for dialup users with dynamic IPs. But UUCP is frequently not an option because most ISPs don't know what it is and those that do charge extra for it. I've sometimes thought, wouldn't it be better to use Qpopper's upload feature? That way, the authentication depends on knowing your username/ password, which is much more reliable than the domainname/ reverse domainname checks SMTP does. It would be easy enough to write an Exim transport which contacts the remote popserver. Unfortunately, it means the ISP would have to change to Qpopper, and they'd probably say, "But our current system works fine for our Windows users. And we don't support Linux anyway." -- -Mike Orr, [EMAIL PROTECTED]
Python curses import error
I'm trying to write my first Python curses program. The "Curses Programming with Python" HOWTO (A.M. Kychling, April 1998) says there are two modules, "curses" (Lance Ellinghouse) included with the Python distribution, and the enhanced "cursesmodule" (Oliver Andrich, http://www.fga.de/andrich/creations.html#cursesmodule -- but the webpage is gone now). On my Slink system, I can do "import curses" fine. But if I do "import cursesmodule", it says: % python Python 1.5.1 (#1, Oct 15 1998, 01:17:50) [GCC 2.7.2.3] on linux2 Copyright 1991-1995 Stichting Mathematisch Centrum, Amsterdam >>> import cursesmodule Traceback (innermost last): File "", line 1, in ? ImportError: dynamic module does not define init function (initcursesmodule) >>> "cursesmodule" on my computer is actually /usr/lib/python1.5/lib-dynload/cursesmodule.so . "curses" does not exist as curses.so or curses.py or anything, so it must be inside the Python binary or something, I guess. [EMAIL PROTECTED]:~% dpkg -l python\* [exit=1] [tty=p2] [11:35AM] ... ii python-base 1.5.1-5An interactive object-oriented scripting ... ii python-curses 1.5.1-5Curses support libraries for Python. ii python-dev 1.5.1-5Header files and a static library for Python -- -Mike Orr, [EMAIL PROTECTED]
Re: xdm not working
On Mon, Oct 05, 1998 at 02:14:09PM +0100, G. Kapetanios wrote: > > Following an earlier email concerning failure to look up a host adress > which also prevented xdm from running I have included my ISP nameserver > address in /etc/resolv.conf. Now xdm starts, I log in and the system waits > before starting the window manager for some time I still havent waited > long enough for the window manager to start as I close down the xserver. > The question I have is how do you set up xdm for a system not conneted to > the net and with loopback as the only network inteface. I still have my > kernel comiled with etherner support should IU recompile without ethernet > support ? You shouldn't have to recompile the kernel. 1) Do a "route -n" and see if any routes are going through your Ethernet interface (especially the default route). "route del" these and comment out their lines in /etc/init.d/network. The only route that should exist when you're not actively on a network should be the loopback. This will ensure that applications get an immediate "network unreachable" error rather than a delay, if they try to access outside resources like a nameserver. (Ah, if only Win95 had a routing table...) 2) If you use ppp or diald, they should automatically be deleting their routes when they disconnect. (Except that diald keeps its metric-1 slip routes.) Pppd seems reliable in this regard but sometimes diald seems to leave bogus routes in, and I have to stop and restart it. 3) Look at /etc/nsswitch.conf. There should be a line "hosts: files dns". Make sure the "dns" is after "files", not before, or that will also cause a delay. I also have a "networks: files" line in there. 4) /etc/resolv.conf should have your ISP's nameserver, as you said. I don't use xdm much, so I don't know if there's anything in the xdm configuration that would need to be changed. Presumably it would be something in /etc/X11/xdm if so. > Any help will be greatly appreciated > George > > --- > George Kapetanios > Churchill College > Cambridge, CB3 0DSE-Mail: [EMAIL PROTECTED] > U.K. WWW: http://garfield.chu.cam.ac.uk/~gk205/work_info.html > --- > > > > -- > Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null > >
TK applications aborting
I've been having TK applications aborting the past few days. Each time it happens, I get a message like this: X Error of failed request: BadGC (invalid GC parameter) Major opcode of failed request: 70 (X_PolyFillRectangle) Resource id in failed request: 0x26 Serial number of failed request: 22157 Current serial number in output stream: 22206 [2] + exit 1 tkman This has happened repeatedly in two Python applications (PTUI and Grail), and also in Tkman and Tkinfo. I'm using tk8.0 version 8.0.3-1 on a slink system. Is this happening to anybody else? -Mike Orr, [EMAIL PROTECTED]
Re: fvwm2 behaves differently from bo --> hamm?
On Mon, Jul 06, 1998 at 02:52:37PM +0200, Dirk Bonne wrote: > I've update from bo to hamm, and now I can't drag a window accross > virtual screens. If you don't have the virtual screens at all, make sure there is a DeskTopSize entry in ~/.fvwm2/post-hook or /etc/X11/fvwm2/system.fvwm2rc I use "DeskTopSize 2 2". -- ** MIKE ORR <[EMAIL PROTECTED]> ** ** Russki Deutsch Esperanto* * ** * * ** * * * (Insert silly quote here) -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null
Leafnode segfaulting in texpire & fetch
Texpire started segfaulting on me a couple days ago, right
at the very beginning of the program. Today, fetch also started
segfaulting. It's working now, but I just wanted to let people know in
case it happens to anyone else.
Purging and reinstalling leafnode (1.4-10) did not help, but deleting
/var/spool/news (230MB) and reinstalling fixed it. I suspect one of the
configuration files got corrupted, but I'm not sure. (I'll have to
re-download all the news, but a lot of it was supposed to have expired
anyway.)
The strace from both programs looks similar:
TEXPIRE:
execve("/usr/sbin/texpire", ["texpire"], [/* 38 vars */]) = 0
brk(0) = 0x805298c
open("/etc/ld.so.preload", O_RDONLY)= -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=10882, ...}) = 0
[2899 lines deleted]
open("/var/lib/leafnode/groupinfo", O_RDONLY) = 3
stat("/var/lib/leafnode/groupinfo", {st_mode=S_IFREG|0664, st_size=7168, ...})
= 0
brk(0x8056000) = 0x8056000
open("/var/lib/leafnode/groupinfo", O_RDONLY) = 4
read(4, "Very 0 1 1 -x-\nalabama 18 17 26"..., 7168) = 7168
close(4)= 0
brk(0x8057000) = 0x8057000
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++
FETCH:
execve("/usr/sbin/fetch", ["fetch"], [/* 46 vars */]) = 0
brk(0) = 0x8054c7c
open("/etc/ld.so.preload", O_RDONLY)= -1 ENOENT (No such file or directory)
[193 lines deleted]
lseek(7, 0, SEEK_CUR) = 0
getdents(7, /* 2 entries */, 3933) = 28
stat(".", {st_mode=S_IFDIR|0755, st_size=1024, ...}) = 0
stat("..", {st_mode=S_IFDIR|0775, st_size=1024, ...}) = 0
lseek(7, 0, SEEK_CUR) = 1024
getdents(7, /* 0 entries */, 3933) = 0
close(7)= 0
open("/var/lib/leafnode/groupinfo", O_RDONLY) = 7
stat("/var/lib/leafnode/groupinfo", {st_mode=S_IFREG|0664, st_size=7168, ...})
= 0
open("/var/lib/leafnode/groupinfo", O_RDONLY) = 8
read(8, "Very 0 1 1 -x-\nalabama 18 17 26"..., 7168) = 7168
close(8)= 0
brk(0x805b000) = 0x805b000
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++
I wondered if the /var/lib/leafnode/groupinfo file could be the culprit,
but I don't have it anymore. I pulled a copy from yesterday morning's
backup and fed it to leafnode, but no segfault.
--
** MIKE ORR <[EMAIL PROTECTED]> **
** Russki Deutsch Esperanto* *
** * *
** * *
* (Insert silly quote here)
--
Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null
Vim embedded python
[in vim 5.1-0.4] :python import vim Sorry, this command is not implemented :version VIM - Vi IMproved 5.1 (1998 Apr 7, compiled Jun 16 1998 16:49:40) Compiled with (+) or without (-): -python Was there some problem getting vim-with-python to compile, or was it just a feature nobody has asked for yet? I would find it very useful. -- ** MIKE ORR <[EMAIL PROTECTED]> ** ** Russki Deutsch Esperanto* * ** * * ** * * * (Insert silly quote here) -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null
Re: mouseman+ with linux
On Tue, Jun 23, 1998 at 04:14:52PM -0500, [EMAIL PROTECTED] wrote: > Does anybody have any info on using the 4-button/1-wheel logitech > mouseman+ mouse with linux? See Colas Nahaboo's X Mouse Wheel Scroll Page at http://www.inria.fr/koala/colas/mouse-wheel-scroll/ and also also /usr/doc/X11/README.mouse You will have to spend some time tweaking the settings and your XF86Config file and making sure each button does what you want. I have a Logitech MouseMan+, which I had set up specially so that buttons 2 and 3 are reversed, so that the right button does "paste" (same as GPM) and the bottom button does the other thing. (I think I put a "pointer 1 3 2" line or such somewhere, but I can't find it in ~/.xsession or /etc/X11/XF86Config so I'm not sure where I put it.) Anyway, when I tried to set up the mouse wheel thing, my paste button moved to the bottom button, and the Debian Menu (which I use all the time) also moved to the bottom button. This was intolerable, so I backed out the changes until I can find some time to get all the buttons right. Also, the doc mentions that your applications receive the wheel-up and wheel-down events as two extra "buttons" but then makess you remap the button order because X only supports 5 buttons and you now have 6. I'm not sure what the optimal settings should be, and how each application should be set up. If you try this and it works, let me know. Maybe we can put together a "Suggested Debian Solution" if no one else is working on it. One advantage of Linux over Windows is that the Windows driver cannot distinguish between pressing the bottom button and pressing the wheel button. I got one of these mice for my mother (she finds it better for her hand than the MS mice because of her carpal-tunnel syndrome). Anyway, she didn't want to do any dragging, and so we made the wheel button be a "start drag" command. Unfortunately, the function got propagated to the bottom button as well. But because she's used to resting her thumb right where the bottom button is, she was constantly pressing the bottom button inadvertently. So we had to disable the bottom button, which meant we lost the wheel button and the "drag button" function as well. > What about gpm? I haven't heard anything about GPM. -- ** MIKE ORR <[EMAIL PROTECTED]> ** ** Russki Deutsch Esperanto* * ** * * ** * * * (Insert silly quote here) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug in Gnome panel command?
I just installed Gnome (0.12-1), and when I try to run panel it says: _kruto[p1]:~/exp% panel panel: error in loading shared libraries /usr//lib/libgdk.so.1: undefined symbol: XListInputDevices libgtk.so.1 is in package libgtk1 (1.0.3-1). I tried reinstalling gnome and libgtk1 and tried rerunning ldconfig, but it didn't help. Is libgtk1 miscompiled? -- ** MIKE ORR <[EMAIL PROTECTED]> ** ** Russki Deutsch Esperanto* * ** * * ** * * * (Insert silly quote here) _kruto[p1]:~/exp% ldd =panel [mso] 7:04PM libgnomeui.so.0 => /usr//lib/libgnomeui.so.0 (0x4000c000) libgnome.so.0 => /usr//lib/libgnome.so.0 (0x4001a000) libdl.so.2 => /lib/libdl.so.2 (0x40026000) libgtk.so.1 => /usr//lib/libgtk.so.1 (0x40029000) libgdk.so.1 => /usr//lib/libgdk.so.1 (0x400d3000) libglib.so.1 => /usr//lib/libglib.so.1 (0x400f4000) libm.so.6 => /lib/libm.so.6 (0x4010) libX11.so.6 => /usr/X11R6/lib/libX11.so.6 (0x40119000) libXext.so.6 => /usr/X11R6/lib/libXext.so.6 (0x401b7000) libSM.so.6 => /usr/X11R6/lib/libSM.so.6 (0x401c2000) libICE.so.6 => /usr/X11R6/lib/libICE.so.6 (0x401cb000) libc.so.6 => /lib/libc.so.6 (0x401e) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x4000)
Re: MH folders in Pine
On Thu, 21 May 1998, Anders Hammarquist wrote: > >Does anyone know how to get Pine to read mail from MH style folders? > > You give it the magic folder path #mh/, so, for example, your > inbox would be set like > > inbox-path=#mh/inbox > My .pinerc has: inbox-path=#MHINBOX I save my mail in a two-level-deep hierarchy. In order to get pine to show all my folders, I had to list all the top-level (parent) folders explicitly: folder-collections=#MH/[], #MH/DEBIAN/[], #MH/LANG_JAVA/[], #MH/LANG_PERL/[], #MH/LANG_PYTHON/[], #MH/PEOPLE/[], #MH/O/[], #MH/WORK/[] Does anybody know about a mailreader called Mona? I think I saw an announcement for it in comp.os.linux.announce a while ago, but I've never been able to find the program. It's supposed to be a rewrite of Pine which supports colorization. ** MIKE ORR <[EMAIL PROTECTED]> ** ** Russki Deutsch Esperanto* * ** * * ** * * * (Insert silly quote here) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: delurk
On Thu, Jan 01, 1998 at 03:05:18PM -0700, dave mallery wrote: > i have been lurking here for about a month and find this an amazing > list. Welcome. > one simple question: is there any filter with some laserjet4 support > as far as point size and fonts in pcl?? am using magicfilter as was > suggested a few days ago. If your source document is text, 'enscript' has a -F option which lets you pick a font. If the source is already postscript, there might be something in the 'psutils' package which will help (I did a quick look and didn't find anything, but it "must" be there somewhere). The fonts themselves live in /usr/lib/ghostscript/fonts. Sorry to be so vague, but I don't know enough about fonts to know which sizes which fonts support, nor how to match fontnames to the files they come from, so I haven't messed with resizing things, except to install some Cyrillic fonts. -- ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA *Russki * ** Tel: +1 (206) 522-9627, fax: 328-6209 * Deutsch * ** * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Arrgh, its the newbie again...
On Wed, 17 Dec 1997, Will Lowe wrote: > On Wed, 17 Dec 1997, Rick wrote: > > > dpkg -i /dos-d/linux/packages/the_name_of_file.deb > > Why don't you move them into /tmp? Careful, /tmp gets wiped at reboot. So if you want to make sure the files stay around till you're done with them, keep them out of /tmp. ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA *Russki * ** Tel: +1 (206) 522-9627, fax: 328-6209 * Deutsch * ** * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Python-tk giving Xpm errors, can't find tkinter [unstable]
When I try to run a python script that uses Tk, I get: ### BEGIN INCLUDE _kruto[p4]:~% grail& [1] 25187 _kruto[p4]:~% python: can't resolve symbol 'XpmCreateXpmImageFromData'] python: can't resolve symbol '_IO_putc' python: can't resolve symbol 'XpmFreeXpmImage' python: can't resolve symbol 'XpmReadFileToXpmImage' Traceback (innermost last): File "/usr/local/bin/grail", line 33, in ? import Tkinter # Do this first to avoid confusing ni on Mac/Win File "/usr/lib/python1.4/tkinter/Tkinter.py", line 11, in ? import tkinter ImportError: No module named tkinter [1] + exit 1 grail ### END INCLUDE This just happened this week after a bulk upgrade. (Other than that the upgrade went very smoothly.) By "bulk upgrade" I mean letting dselect upgrade everything it thought necessary. A similar problem happened to me several months ago. I don't remember the error messages, but all python TK programs refused to run. It was solved then by reinstalling tcl & tk. But this time around, I tried reinstalling python-tk, tcl, tk, xpm and xlib, and none of them helped. [python-tk] Depends: python-base (= 1.4.0-4), libc5 (>= 5.4.0-0), tcl76 (>= 7.6-1), tk42 (>= 4.2-1), xlib6 (>= 3.2-0) ii python-base 1.4.0-4An interactive object-oriented ii python-tk 1.4.0-4Tk support modules for Python. ii libc5 5.4.33-7 The Linux C library version 5 (run-time ii libc6 2.0.5c-0.1 The GNU C library version 2 (run-time ii tcl76 7.6p2-7The Tool Command Language (TCL) v7.6 - ii tk424.2p2-7The Tk toolkit for TCL and X11 v4.2 - ii xlib6 3.3.1-2Shared libraries required by X clients ii xpm4.7 3.4j-0.5 X Pixmap libraries (for libc5) - ii xpm4g 3.4j-0.5 X Pixmap libraries (for libc6) - I think the real problem is those Xpm messages, because Tkinter.py tries to import first _tkinter.py and if that fails, tkinter.py. If I import _tkinter directly, I get fewer error messages: _kruto[p4]:/usr/lib/python1.4/tkinter% python Python 1.4 (Mar 4 1997) [GCC 2.7.2.1 Objective-C snapshot 960906] Copyright 1991-1995 Stichting Mathematisch Centrum, Amsterdam >>> import _tkinter python: can't resolve symbol 'XpmCreateXpmImageFromData' python: can't resolve symbol '_IO_putc' python: can't resolve symbol 'XpmFreeXpmImage' python: can't resolve symbol 'XpmReadFileToXpmImage' Traceback (innermost last): File "", line 1, in ? ImportError: Unable to resolve symbol >>> ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA *Russki * ** Tel: +1 (206) 522-9627, fax: 328-6209 * Deutsch * ** * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
German Debianites please read
[English -- German part follows] I may have a chance to go to Germany later this month. If so, I would like to meet any Debianites and Linux activists who might have a moment to say hi, and also attend a Linux users' group or two if there are any. I'll be in Unna first, then travelling maybe to Duesseldorf, and possibly also to Cologne and Bielefeld. And of course I may go through Frankfurt, depending on which flight I get on. The reason for this unusual itinerary is that I'll be following a friend who isn't sure where he'll be staying, but he knows it will be somewhere in the Rhine-Ruhr area, and I have another friend in Bielefeld. If anyone has any information, please email me. [deutsch] Es mag vorkommen, dass ich die Chance habe, spaeter in Dezember nach Deutschland zu reisen. Wenn's gelingt, wuerde ich gern irgendeine Debianisten und Linux-Anhaenger treffen, die einen Moment haetten, sich mir vorzustellen. Ich moechte auch in einer oder zwei Linux-Benutzer-Gruppen teilnehmen, wenn sie passieren. Ich bin zuerst in Unna, und dann vielleicht in Duesseldorf. Vielleicht fahr' ich auch nach Koeln und Bielefeld. Der Grund fuer die untypishce Reiseroute ist, dass ich einem Freund folge, und er weiss nicht, wo er stecken wird, sondern weiss nur, dass es irgendwo im Rhein-Ruhrgebiet wird, und ich hab sonst einen Freund in Bielefeld zu besuchen. Wenn jemand was fuer Information hat, bitte senden Sie mir einen Email! Danke. ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA *Russki * ** Tel: +1 (206) 522-9627, fax: 328-6209 * Deutsch * ** * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Linux, FAT32 and a mighty odd MBR.
On Thu, Oct 30, 1997 at 01:15:06PM +1100, Dale Harrison wrote: > Until it came to LILO. Installed LILO, booted Linux fine. Added the lines > to boot Win95, it took 2 goes for it to install the new record [it never > complained the first time about failing]. Problem is, it simply doesn't > boot Win95. It says "booting Win95", then returns to the LILO prompt. > > The problem remains, I can't seem to overwrite the MBR, not with LILO, not > with fdisk. I found out that the Windows partition was FAT32, but surely I bet FAT32 is the problem. It is not DOS-compatible, so existing disk utilities cannot operate on it. We had that problem at work: a department got a DTP computer with FAT32 and we can't work on it, because our scandisk and f-prot won't work on it. Under Linux there's a kernel patch to recognize FAT32 partitions, but it's still experimental. Since Lilo works differently than any other Linux utility (it writes the physical address of the partition's bootstrapping code into its menu, because when Linux is not booted yet and names such as "/dev/hda2" have no meaning yet. Perhaps FAT32 has a different disk header format or something. If not too inconvenient, you may want to reformat the Win95 partition to FAT16. That may be less of a headache in the long run until all the utilities have had time to catch up. This is all based on secondhand knowledge -- I've never used FAT32 myself. -- ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: dselect infinite loop?
On Mon, Oct 27, 1997 at 06:07:51PM -0500, Will Lowe wrote: > If you note carefully your package selections within the dselect > "selections" screen, you'll note that dselect does NOT unselect packages > after it installs them -- if you select xforms, install it (via dselect), > and then run dselect, it remains selected as "install". You want to > change this from "install" to "leave in current state"; this is probably > done by pressing "=" when the package line is selected -- at this point > the indicator should change from > *** to **= or something similar. Huh? When I install a package, the indicator goes to ***, but it never tries to reinstall that package again until a newer version appears. In fact, if I *want* to reinstall it, I have to remove it first. "=" is more meant for special occasions, such as when there is a newer version but you don't want to upgrade yet. At least that's what it's supposed to do. Somebody else will have to figure out why yours keeps on reinstalling. -- ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
End of device [was Re: kernel messages]
On Sun, Oct 26, 1997 at 09:14:56PM -0500, Thalia L. Hooker wrote: > Hi, > > I was just noticed in my log files that I have these messages in my > /var/adm/kern.log: > attempt to access beyond end of device > 03:02: rw=0, want 707406379, limit 1433376 > > This message is repeated several times and ultimately the computer locks up. > Does someone have an idea what is causing these messages? > > btw, the only process being run by a user at this point is myself trying to > build a new kernel. I don't know if this is related or not, but the last time I tried to make a partition on my disk, I got hundreds of consecutive seek errors from mkfs. This is on a 1.3G Western Digital IDE that has never had any bad sectors. I had just deleted some partitions to resize them, and the last new partition failed to build. (I think there was some free space at the end of the disk before, and this time around I tried to use up all the space.) This was a couple months ago. I just chalked it up to either my disk really is going bad (although it's remained reliable ever since), or else e2fsck really is overrunning the end of the disk for some reason, or cfdisk reported more blocks than were actually there. This is with a 2.0.28 kernel. I just gave up trying to use that portion of the disk. -- ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Iomega Ditto Tape drive on parallel port?
On Wed, Oct 01, 1997 at 10:23:28PM +0200, Johann Spies wrote: > I want to by an external tape drive and am considering an Iomega Ditto > that connects to the parallel port. > > However, the Howto's say that such drives (that use the parallel port) are > unsupported. Is that also the situation with the 2.0.30-kernel? I just saw a reference to that yesterday, at http://www.torque.net/ditto.html by someone who's working on it. But the page is dated last April. -- ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: debian-user needs a charter!
On Fri, 22 Aug 1997, Jim Pick wrote: > I just came back from a week's work, and I've discovered that the nature > of debian-user has changed dramatically. That's unfortunate, since ... > > Now it has been hijacked for use as a batteground for a flame-war. > > Bruce, could you please write up something simple - perhaps stating that > debian-user is a "user support" list only, and that from now on, > political discussions should go to debian-debate? (or some other list). This is what debian-talk was supposed to be for -- a place to take noisy discussions to. Now's a good time to put such a list on the "official" site. If debian-user is seen as solely a Q&A help forum, and debian-devel as Q&A + debates on how to extend Debian technically, then debian-talk could be chartered as "everything that does not belong on the other lists, such as non-technical discussions regarding Debian, peripheral discussions that are only incidentally about Debian but mostly about something else, discussions about the administrative structure of Debian, complaints about how some developer is doing things, advocacy for & against Debian vis-a-vis other distributions, etc." Right now there is no other place for these discussions, so they end up in the "default" list, debian-user. GNU has gnu.misc.discuss, Linux has c.o.l.misc and c.o.l.advocacy, why doesn't Debian have the same? Or else leave debian-user alone and move all Q&A to debian-help with the charter, "Newbies, here's your place to ask questions! Oldsters, sysadmins and Debian developers, your questions also belong here." (The last part is to give an incentive for more experienced people to subscribe, to increase the quality of the answers.) Debian-devel, of course, would remain the forum for questions of interest only to developers: e.g., problems building a new package, which directory a conffile should go in, etc. Homage must also be made to Lars W's (I think it was his) suggestion several months ago for a debian-politics list. ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Debian Newsletter Proposal
On Sun, 10 Aug 1997, Brandon Mitchell wrote: > I've put together some ideas for the debian newsletter I proposed last > week. > 1) This will be a weekly summary of info from debian-user, debian-devel, I tried to do something like this last spring. I quickly gave up because it takes a *lot* of hours to read through a week's worth of messages, decide which threads are worthy to summarize, and then write the summaries. Also because a good summary, even after cutting out all the fluff, still ends up being several pages long and thus pushes the limit of what people are willing to read through. In case it may be helpful, here's how it worked: the idea sprang out of a discussion on debian-talk (an unofficial list run by Mark Constable <[EMAIL PROTECTED]> -- I don't know if it still exists), regarding how all these great ideas on the lists just get lost because it's impossible to find that one vital piece of information you need to restore/enhance your system, the tip you read last month that is now buried somewhere in a thousand messages and unfindable by the search engines. Anyway, two other people and I decided to write weekly summaries of important information gleaned from debian-user and debian-devel. At first I just had one window open with pine, sorted by subject, and another window with the vi document I was writing. That led to a python program that would read the messages in an MH folder and create a blank template listing each subject-thread, the name/address of the originator, and the number of replies, and a blank line. Then you just read the messages and write a one or two sentance summary of the threads you want to keep (above the blank line). I can send you the program and an archive of the discussion if that would help, since I don't think any of it got onto debian-user. Actually, the original idea came from something I saw the year before that, called Digital Espresso. DE was a web newsletter sponsored by a company that made weekly summaries of the discussions in comp.lang.java. I found that to be a very valuable resource -- you just read the summaries and go back to the originals just for the threads you want to find out more about. It's a way of surviving the onslaught of so many messages. However, that zine was supported by a company, and it also no longer exists. Around the time I was writing my summaries, the Debian Documentation Project (debian-doc) came into being. We immediately saw our effort as really a kind of documentation, and were trying to figure out how to incorporate it into the rest of the documentation repitoire that the debian-doc folks were creating. However, since I burned out on writing summaries, we had no content to offer debain-doc, even if we did come up with a structure. Also around that time, Jim Pick <[EMAIL PROTECTED]> came up with the idea of having a debian-tips mailing list and/or web database that would allow people to contribute their own "this is how I did it" hints. However, this idea also has never been implemented. Anyway, good luck with your newsletter. It sounds like you're going a little more general (reporting only on the major trends in the Debian world) than I was doing (trying to be a comprehensive FAQ of all our collective wisdom), which may make the difference in terms of making something small enough to be manageable. If you can pull it off, it's definitely something that's needed. By the way, for trivia's sake: over an eleven day period last March, there were 728 messages (294 threads). I don't know what the number has increased to by now. My debian-user MH archive folder, which has been collecting messages since March 23, has reached 35.3MB (10535 messages), and my debian-devel folder (since March 14) has gotten up to 30.7MB (8612 messages). ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Toshiba laptop resume mode
On Mon, 7 Jul 1997, Bill Moran wrote: > I am trying to install debian on a Toshiba laptop 410 CS and am having > trouble making resume mode work. When I try to resume the computer > hangs. I have a Toshiba Satellite T2130CT; I don't know how much that differs internally from yours. Suspend/resume works wonderfully. I just close the cover and it suspends. I had to go back to the Dos partition and enable it with the Windows Control Panel. Mine also has that function on the keyboard, but I forget which of the function-key icons is for suspend (could it be the lock? or the water faucet? or the suitcase?). Does the screen show where you left off and then it hangs, or what does it show on the screen when you restart? If I leave my computer suspended and unplugged for three days or so, I get a blank screen saying "RESUME FAILURE". I then press a key and it reboots. Naturally the filesystems don't get unmounted cleanly when it dies, so fsck complains, but I have not had any major problems with it. ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: help with exmh
On Thu, 3 Jul 1997, Mark Longair wrote: > On Wed, 2 Jul 1997, Peter S Galbraith wrote: > > > Clint Adams wrote: > > > > In psckf's man page: > > > > BUGS > >Packf doesn't handle the old UUCP-style "mbox" format > >(used by SendMail). To pack messages into this format, > >use the script /usr/lib/mh/packmbox. > > > > This might apply here, but since I've never done it I'm not sure. > > Pine was able to understand both of these formats, though I think packmbox > produced the format I had originally. Many thanks for the help... Yes, packmbox produces the standard Unix mail format, and is the proper way to undo the conversion to MH. Packf is an alternate mh format. ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Dselect won't download
Now that contrib and non-free have moved inside the distributions, I can no longer access them via dselect. What stupid thing am I doing wrong? I'm trying to get unstable's main, contrib and non-free areas. I tried both dpkg 1.4.0.8 in stable and 1.4.0.17 in unstable, combined with dpkg-ftp 1.4.8 (which is in both stable and unstable). If I set the access method Debian directory to "/debian" and the distributions to "unstable", I get only the Packages file from the main area (via the old link /debian/unstable, I assume). If I set the Debian directory to "/debian/dists/unstable" and distributions to "main contrib non-free", I get all three Packages files. The Install routine lists all the files I want in this form: "want: dists/unstable/main/binary-i386/doc/dwww_1.4.1-1.deb (46K)" ^^ An extra level of "dists/unstable" directories appears out of nowhere in the file names, even though dselect has already cd'd *inside* /debian/dists/unstable. All the downloads fail with "No such file or directory". Thinking that perhaps the "dists" part is now implied by dselect and I don't actually have to mention that directory, I try "/debian" and "unstable/main unstable/contrib unstable/non-free". Then it can't find any of the Packages files. ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: modem speed
On Wed, 11 Jun 1997, George Bonser wrote: > On Wed, 11 Jun 1997, Timothy Phan wrote: > > > Hi, > > > > Anyone knows how to determine the modem speed after the PPP has > > connected to the ISP? > > You might set the modem to report the link data rate rather than the > serial port rate then tail /var/log/messages and grep for CONNECT after > login. > > Another way to do this is to use the REPORT option in your chatscript: REPORT CONNECT ABORTBUSY ABORT"NO CARRIER" etc... Then have pppd invokes chat with the -r option + filename (e.g., in /etc/ppp/peers/provider for ppp 2.3b3-2): connect "/usr/sbin/chat -v -f /etc/chatscripts/uw -r /var/log/ppp-speed-uw.log" defaultroute /dev/ttyS1 38400 user mso noauth noproxyarp (all on one line of course). Chat will then write the entire CONNECT line into the file. ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
How to get new uploads?
I used to find new uploads on ftp.debian.org in debian/Incoming, but now this directory now has a message in it but no other files: "Do NOT upload files here! This is the WRONG directory to upload files into Upload into /home/Debian/ftp/private/project/Incoming" Unfortunately, this other directory is not visible to anonymous logins, AFAICT. Sometimes you need a package or update right away, and packages often get stuck in Incoming for two weeks or longer before making their way into unstable. A few packages are available on the maintainer's ftp sites, but not most of them. Would it be possible to mirror the real Incoming directory back into debian/Incoming? ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
lprm won't let user remove his own print jobs
I seem to have the permissions messed up for lpd, because user's can't remove their own jobs. /var/spool/lpd and subdirs are set to "drwxrwxr-x root.lp". I tried drwxrwxrwx, drwxrwsr-x and drwxrwsrwx but none of these helped. Script started on Mon Mar 24 11:29:48 1997 $ lpr Finding_Fonts.txt $ lpq lj is ready and printing Rank Owner Job Files Total Size active mso82 Finding_Fonts.txt 9045960 bytes $ lprm - cfA082Aa09695: Permission denied cfA082Aa09695: Permission denied $ lprm 82 cfA082Aa09695: Permission denied cfA082Aa09695: Permission denied $ ls -ld /var/spool/lpd drwxrwxr-x 6 root lp 1024 Jan 12 21:19 /var/spool/lpd $ ls -lR /var/spool/lpd total 4 drwxrwxr-x 2 root lp 1024 Mar 24 11:31 hpljet4l drwxrwxr-x 2 root lp 1024 Feb 05 17:59 ibmpro drwxrwxr-x 2 root lp 1024 Jan 12 21:20 lp drwxrwxr-x 2 root lp 1024 Jul 29 1996 remote /var/spool/lpd/hpljet4l: total 8873 -rw-rw 1 daemon lp132 Mar 24 11:31 cfA082Aa09695 -rw-rw 1 mso lp9045960 Mar 24 11:31 dfA082Aa09695 -rw-r--r-- 1 root root 19 Mar 24 11:31 lock -rw-rw-r-- 1 root root 25 Mar 24 11:31 status /var/spool/lpd/ibmpro: total 2 -rw-r--r-- 1 root lp 4 Mar 22 12:49 lock -rw-rw-r-- 1 root lp 25 Feb 05 17:58 status /var/spool/lpd/lp: total 0 /var/spool/lpd/remote: total 0 $ su -c "lprm 82" Password: dfA082Aa09695 dequeued cfA082Aa09695 dequeued $ lpq no entries $ Script done on Mon Mar 24 11:32:53 1997 ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here)
Re: Connection closed by foreign host.
Perhaps telnet is working properly but something else is causing your login shell to exit immediately. I get a similar response, not with telnet, but with script. I would get "Script started, output file is typescript" and then it would immediately exit with "Script done, output file is typescript". This happens with bash at the console, but it does not happen in an rxvt, and it doesn't happen with zsh at all. I don't notice it anymore since I use zsh all the time now, but a quick check revealed that it's still doing that. Try putting an "echo Hello telnet world" as the first line of you .bashrc (or whatever shell initialization file you use), to verify whether your telnet session gets at least that far. If it doesn't, try switching login shells and see if it that works. ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here)
Re: Connecting to a Windows NT server
On Sat, 15 Mar 1997, Christopher C. Pitts wrote: > Leandro Asnaghi-Nicastro wrote: > > > > Has anyone managed to succesfully connect a Windows NT server? THe > > How-To's I've looked up and the books I have bought assume I use a > > normal server (Unix/Linux) but unfortunately at work we run Windows ^^ :) ;):) ;):) > > NTfor our servers .. > I currently must use '95 to connect to my ISP... he runs NT...I'm Connecting pppd to an NT server is tricky, but do-able. My friend got it up and running immediately; for me it took several days. I was able to get it to work by putting " * " into BOTH pap-secrets and chap-secrets (in /etc/ppp), and by including the "user " option in the options file (/etc/ppp/peers/provider for pppd 2.3b3-2, another file for some earlier versions). Remember to do "tail -f /var/log/ppp.log" so you can watch for error messages as they happen. I used to get a "serial line is looped back" error connecting to NT, which the manual was extremely unhelpful in troubleshooting. I think it means that the server has hung up on you. In any case, the error mysteriously disappeared when I got my configuration files right. You can also set a "verbose debugging" option for pppd, which shows more detail about what is sent & received. I don't remember the option offhand, but it's in the docs somewhere. Which version of pppd are you using? I upgraded to 2.3b3-2 last December. I don't remember if that was because of my problems connecting to NT, or whether it was after I had gotten it working. More recent versions of pppd do better at handling NT servers, and also have NT-specific sections in their documentation. However, it takes some time to get used to the changes to the configuration file structure in new versions. If you decide to upgrade, make a backup of /etc/ppp.*. /etc/ppp/* and /etc/chatscripts/* if you have anything in your old configuration files you might want to look at later after making changes. Good luck. ** MICHAEL SCOTT ORR <[EMAIL PROTECTED]> * ** 1405 NE 56th Street, Seattle, WA 98105 USA * English * ** Tel: +1 (206) 522-9627, fax: 328-6209 *Russki * ** Work: [EMAIL PROTECTED] * Esperanto * * (Insert silly quote here)
