Re: Newbie mistakes
On Sat, Feb 19, 2000 at 09:21:36PM -0600, Ian Timshel wrote:
> I'm up and subscribed. Thanks to all who were kind enough to tolerate my
> post
De nada.
On request: slow down and space out your paragraphs. Dense text is
hard to read.
> I'm having some problems after a smooth installation of Essential Debian 2.1
> Everything went wonderfully once I felt my way around. I had many attempts
> but the instructions were there to follow and it came off very well....
> except I neglected to note how much room I was using. When I was having
> trouble with X server configuration I checked in Midnight Commander and I
> had 1k left and then I watched it go as well. I'm comfortable with dselect
> and have no trouble loading what I want and gassing what I don't. The
> Midnight Commander has been good for me. I can change permissions and have
> become quite familiar with the directory tree. I'm hammered the FAQ's,
> HOWTO's and man pages.
Trick: http://localhost/doc/ should point to your documentation
directory. You can point a browser -- Netscape/Mozilla for graphics, or
lynx, links, or w3m in console mode.
> I crashed system without enough room left to spool anything and have had
> problems ever since. I can't run startx without problems. Here is the
> list. I thought I sent it earlier but I think I messed up the address.
> Please excuse this if it is a repeat of such a big whack of writing.
If this isn't relevant to your current problems (you've created a fresh
installation without problems), drop the detail, as it makes it unclear
whether this is relevant background or just a personal vent. Not that I
don't understand the frustrations, trust me. <g>
> I've repartitioned, reinstalled, dselected, and commented this and that
> all to no avail. One of the first messages was: "Can't read "messages.
> (phase2.1) : no such variable while executing. "label $w.waiting -text
> $messages(phase2.1) file /usr/X11R6Setup/phase2.tcl line 25.
Context? Boot messages? Output from running a command?
Once your system has booted, you can run the command "dmesg" to capture
the kernel ring buffer output. This should include all boot messages
(other than hardware device output).
You can also read from /var/log/<foo> for other diagnostics. I'd start
with /var/log/messages and /var/log/kern.log
> Now I have the following experience at boot up. Valid xdm at boot
> GNU/Linux logon screen. Login and it leaves me with a textured screen and a
> mouse cursor to play with and that's it.
You have X up and running successfully. Good. You appear to be missing
a window manager. We'll deal with that later.
What you're looking at is the bare-assed X server display. You can
launch something onto it from a console by using the syntax:
$ xterm -display :0 &
...as the same user.
Incidentally, I'd suggest *not* booting a graphical screen (xdm/gdm)
automatically. My own preference is to log on to a console and 'startx'
when I want GUI. Makes debug easier. My X sessions tend to last weeks
anyway.
> ^Alt F1 takes me to the console
> with no error messages. Logon there as root then startx and get "fatal
> server error server is already active for display 0. If this server is no
> longer running remove /tmp/.X0-lock and start again."
You have X up and running successfully -- on display :0. If you want to
launch a second X session (which is what you're trying to do whether you
want it or not <g>), you'll have to specify another display.
Try:
startx twm -- :1 1> ~/.startx.log 2>&1 &
...from console. This starts X, runs twm (a very minimal windowmanager,
if that's not installed, pick another X client -- say fvwm or xterm),
startx X on display :1, and outputs all stdout and sterr to the file
.startx.log. That's pretty much my standard X startup command, except
that I use WindowMaker, and it starts by default (so I don't have to
specify it). The "--" separates X client options (eg: the window
manager) from X server options (eg: the display, color depth, etc.).
I also disable gdm and xdm by deleting the corresponding [SK][0-9][0-9]gdm
files from /etc/rc.d/ You can leave the /etc/init.d scripts alone.
Alternatively, you can run:
/etc/init.d/gdm stop
...and/or whatever is equivalent for your X display manager, to stop the
login screen. This kills the X session on display 0.
> At this point the
> drive is reading like crazy and after a few minutes kicks out: "unable to
> load interpreter /etc/X11/xinit/xinitrc line 3 685 Segmentation fault
> /etc/x11/xinit/xintirc" I had some problems with a file marked red with !
> xserverrc at /etc/X11/xinit/xinitrc but it no longer appears to be there.
How long is "a few minutes"? X often spins for a while before it dies,
I suppose this depends on what clients you're trying to start up. I
haven't tried launching a Gnome or KDE session on top of an active X
session, but it could conceivably take 30-60+ seconds for them to wake
up enough to realize they were dead.
> Any ideas would be greatly appreciated. I messed up and have been
> blown away with the ease of the packaging scheme. I like it and it makes so
> much sense to me. It was flawless until I fell asleep at the wheel! %@$$@**
> I've tried every approach to try and have this system write over and replace
> the file but I just don't know enough to make a go of any of it. Managed to
> get LILO configured to do my bidding. Given that I can't write to my CD
> there shouldn't be a reason that I couldn't start again but I have failed to
> figure out how. I'm tempted to write in a dos partition and format it for
> dos then through the Linux fdisk in and start that way. I'd gladly submit
> to a better way.
You're very close to success -- at least for getting X up. Seems like
you don't have your windowmanager configuration quite right yet, but you
should be able to get up and running with the above info.
> --
> Unsubscribe? mail -s unsubscribe debian-user-request@lists.debian.org < /dev/null
>
--
Karsten M. Self (kmself@ix.netcom.com)
What part of "Gestalt" don't you understand?
SAS for Linux: http://www.netcom.com/~kmself/SAS/SAS4Linux.html
Mailing list: "subscribe sas-linux" to mailto:majordomo@cranfield.ac.uk
Reply to: