Pages

Search

Multiple sessions (distinct users on the same machine at the sametime) - Forums Linux

Multiple sessions (distinct users on the same machine at the sametime) - Forums Linux

Multiple sessions (distinct users on the same machine at the sametime)

Posted: 07 Jan 2006 04:27 AM PST

kermit <com> wrote: 
 

I'm not sure what you mean - the DISPLAY variable (or -display arg)
controls addressing of an X server by an X application Are you saying
that you don't know if the value should be "0:0" or "1:0"?

(I don't know either, not having the problem :-).


Peter

I/O access over LAN

Posted: 07 Jan 2006 01:28 AM PST

On 01/07/06 10:28, BlueGecko wrote: 

You need some hardware/software to manage everything.
Take a look at < http://www.ltsp.org/ >

Ciao
Giovanni
--
A computer is like an air conditioner,
it stops working when you open Windows.
Registered Linux user #337974 <http://counter.li.org/>

CygwinX

Posted: 05 Jan 2006 08:31 PM PST


"Ken K" <headfog.com> wrote in message
news:IfFvf.5574$..
 

Good. Distinctly up to date versions are available at www.sunfreeware.com: I
find Sun's official releases of software to often lag way, way, way behind
the latest FSF or other open source releases, especially including SSH.

Getting the Suns to play nice with the Linux systems is often quite a lot of
fun, especially because the Linux version of NIS is so much better to work
with.

KVM Switch - Dual Display (DVI) + USB + Audio; 2-port

Posted: 05 Jan 2006 10:31 AM PST


<net> wrote in message
news:googlegroups.com... 
No, it stays DVI the whole way. If you want to convert, you have to add a
DVA-VGA adapter.
 
It all goes through the same bilateral switches. The cable set includes
audio in. The only extra connectors on the box, are the USB ones, and the
audio 'out'. I got mine from kvmdirect.com, and they were able to supply
the correct cable sets. The only problem is that after this unit, Belkin,
then launched the 'DD' version, which supports lower sync rates cheaper.
This had problems, and is currently 'withdrawn', so finding the right unit
and cables, requires talking to a company that knows the Belkin products
better than Belkin do...
 

Best Wishes

define: assertion(heads <256) at disk_dos.c:486 in function probe_partition_for_geon() failed

Posted: 05 Jan 2006 03:49 AM PST


Enrique Perez-Terron wrote: 

thanks a lot...both of u...well..i was installing fedora3 on a disk
which had windows installed on it beforehand..as soon as anaconda
started, i started receving those messages..after ignoring a number of
them over different steps i finally got the disk formatted but the
linux box is running very slow..opening a window on the desktop takes a
long time

me.linuxadmin

Desparate to recover from my stupid mistake with mkdosfs.

Posted: 05 Jan 2006 02:38 AM PST

Miha Verlic wrote:
 


Acronis has a nice tool that works too.

Building a file server - advice please

Posted: 04 Jan 2006 05:09 AM PST

Okay.

Thank you to everyone who has responded. I've had a long chat with Nigel
about what he actually wants, and combined that with what you guys have
suggested.

I've managed to persuade him to free off 1 of the 250gb drives, so that
one along with the 300gb and the 40gb will be going into the new machine
as fixed drives.

The other two 250gb drives he wants to be able to swap out of the linux
box and into his 'doze box. These two therefore I cannot reformat. They'll
have to stay as ntfs discs.

So. If I use the 40gb drive for my /boot / and swap partitions, and then
somehow use lvm for the 300gb and single reformatted 250gb to form my
/home partition. Can I then get a SATAII card to drive the two NTFS discs
so that they can be hot-swapped out?

I've tried to read up on SATA and I'm still a bit hazy on one point. Can I
use the existing 250gb IDE drives with a SATA card? He won't buy any new
hardware beyond a card, SATA, IDE, whatever, to enable him to have all
five drives available all at the same time, but if the SATAII will allow
him to hot swap the IDE drives, then he'll get one for his 'doze machine
as well.

Many thanks,

Dave

--
Dave Stratford ZFCA
http://daves.orpheusweb.co.uk/
Hexagon Systems Limited - Experts in VME systems development

LINUX Server Reboot Frequency

Posted: 03 Jan 2006 06:21 PM PST


"Enrique Perez-Terron" <no> wrote in message
news:home.lan...
 

And what does "free" say? Or /proc/meminfo, when doing such large
allocations? OK, so you'd have to do an allocation large enough to go past
the system RAM and get into swap to really see it happen. But as I
understand it, that memory is still relegated to that program, and is
available *for that program*. At least, that's how I remember building
stacks from scratch: if someone's invented a way to actually release it to
the system instead of merely for that program's use in its own stack, I'm
going to be *VERY* surprised.

And in fact, you can even get away with doing that for quite a while: the
memory allocated will wind up being heavily in swap, and if it's not
actually being used for much, most of it will be swapped out at any given
time.
 

OK, that makes more sense: by freeing up unused stuff off what is
effectively the end of the stack, that could work for releasing the
resources.
 

Who knows? It's closed source, very proprietary, multi-platform, and clearly
written for big iron, not casual implementation by any means. Using its own
memory management makes sense, but at some point you sort of *have* to talk
to glibc unless you want to rewrite a lot of very, very basic functions from
scratch.

Just want to create a bootable CD!!!!!!!!!!!!!

Posted: 03 Jan 2006 12:18 PM PST

James Kimble wrote: 
Here's the overview, it's really not hard. The BIOS will only boot off a
floppy (image), so on the CD or DVD there's a floppy image, the 2.88MB
flavor in most cases. You can roll your own, or D/L the Slackware
install disk and use that.

Copy the floppy image to a file, loop mount it...
mount -o loop flpimg.dat /mnt/temp
now cd to it and edit the startup scripts, etc.

Opinion: I like Slack better than SYSLINUX, I used it for years, I ran
BSD at one time, and it's totally easy to understand and edit.

When you're done, run mkisofs with the -b option (READ the man page!) to
create the CD.

AFAIK there is no general purpose software to do this, most are special
purpose, like mondo, and at least the old version, on which I gave up,
really didn't want to be general purpose. If you write a tool, please
make it human readable, no XML, no GUI, understandable config, etc.

--
bill davidsen
SBC/Prodigy Yorktown Heights NY data center
http://newsgroups.news.prodigy.com

SuSE 10.0 Install problems

Posted: 03 Jan 2006 10:25 AM PST


Jim Castle wrote: 

have you ever used Maxblast setup tools on that 120GB HDD ?
You likely have a non-standard partition table byte set.
Suggest you zero out the disk prior to installation (i forget the exact
'dd' command to do this in linux, but even a FDISK /mbr from a win98
boot disk should do it) - and then wipe the unintended install on the
unintended 250GB SATA - proceed with setup onto 120GB.

GRUB Geom Error

Posted: 01 Jan 2006 07:35 AM PST


"Enrique Perez-Terron" <no> wrote in message
news:home.lan... 

I like Enrique. He actually gives useful answers. Michael? What were you
doing just *before* that reboot? Anything that would touch the disks, or
modify grub settings, like manipulating your partitions from Windows or
Linux?

Xorg.conf

Posted: 01 Jan 2006 06:40 AM PST

On Tue, 03 Jan 2006 21:41:49 +0100, jon <co.uk> wrote:
 

I fully inderstand your statement that something is wrong, if you set the
some bounds for the refresh rate, and X does something else.

Now that it appears confirmed that there is a problem, and that there
does not seem to be a way to make X behave as specified, given the
default modes, the question is 1) is it possible to write a custom mode,
and 2) how serious is it for the monitor to be driven faster than the
recommended rate.

I guess that it is not a problem for the hardware, but that the higher
rate gives the crystals little time to turn, and some visual artifacts
may happen in rapidly changing scenes. But don't sue me if your
monitors goes up in flames with a POFF!

For the question of a custom modeline, I never felt the descriptions I
found were anywhere nearly clear. I did play a little with them some
years ago, and well... I think I got behaviours I could explain,
*mostly*. I don't feel assured that I had the right interpretation of
the stuff. Perhaps others here know more. If you first do some basic
reading, and try to compose a modeline and post it here. Explain how
you arrived at it, and we can probably contribute som lateral thinking.

This route may be easier, if the ATI card is so problematic (something I
have no knowledge about, positive or negative).

-Enrique

? Whats best for a newbie?

Posted: 01 Jan 2006 05:13 AM PST

Happy New Year!

On Sun, 01 Jan 2006, in the Usenet newsgroup comp.os.linux.setup, in article
<ZqRtf.242$eR.60@fed1read03>, MuNkEeSuSELV wrote:
 
 

You seem to be running some form of *nix to start with, I guess you are
asking about learning, rather than the usual "what distribution should
I use" type of question. A good answer would be "The Linux Documentation
Project". One could start here

http://ibiblio.org/pub/linux/docs/HOWTO/
http://en.tldp.org/HOWTO/HOWTO-INDEX/howtos.html

or here

http://tldp.org/guides.html

and find enough reading material to keep you busy for several weeks. Last
I looked, there were over 470 HOWTOs covering a huge number of subjects.
Relatively recent copies should have been included with your Linux distro.
There are more than 25 guides at the that are available for download if they
aren't part of your distribution. Examples would be "The Linux Network
Administrator's Guide" which is also published by O'Reilly and Assoc. for
US$40. Another example is the "solrhe" which is the first edition of
"Securing-Optimizing-Linux-RH-Edition" (version 1.3 actually talks about
Red Hat 6.2, but the concepts are unchanged - there is a commercial version
2.0 which is not free, but is more recent).
 

A few minutes at google will normally uncover more information that you
have time to read.
 

Post here - read here. Additionally, all microsoft net space is blackholed
here, and I couldn't connect even if I wanted to.

Old guy

GDM Login Screen Problem

Posted: 31 Dec 2005 08:18 PM PST

On Mon, 02 Jan 2006 03:01:11 +0100, Chen Yang <com> wrote:
 

Hm, I was wrong, I triggered on the word "audit", which occurs in the
log messages when selinux prevents something.

I just searched the gdm sources for the string AUDIT and rejected, and
none of them occurs in a string in the source. Probably the strings
come from some library used by gdm (there is 35 of them, and thre may be
more if a library calls another).

Is there any surrounding text in /var/log/gdm/:0.log? Are there other
files than :0.log* in the directory?

I suspect that gdm is capable of outputing more error messages.

First I assume that you have this line in /etc/inittab:

x:5a:once:/etc/X11/prefdm -nodaemon

In the file /etc/X11/prefdm you will find this

if [ -n "$preferred" ]; then
$preferred "$@" >/dev/null 2>&1
if [ $? -ne 127 ]; then
exec $0 "$@"
exit $?
fi
fi

Replace "/dev/null" with "/tmp/gdm.stderr".

Open a virtual terminal (Ctrl-Alt-F1), login as root and do

telinit 3; sleep 10; telinit 5

Just wait for the gdm screen to not showup. When it should have been there,
switch back to the vt1 console again. Check that the /var/log/gdm/:0.log
has the error message, and restore the original /etc/X11/prefdm.

Check if there are any more meaningfull messages in /tmp/gdm.stderr.

If this does not help, I can only suggest you use "strace" to find
out what is going on.

In /etc/X11/prefdm, change the line with /dev/null above, to:

/usr/bin/strace -o /tmp/gdm.strace -ff $preferred "$@" >/dev/null 2>&1

Open a virtual terminal (Ctrl-Alt-F1), login as root and do

telinit 3; sleep 10; telinit 5

Just wait for the gdm screen to not showup. When it should have been there,
switch back to the vt1 console again. ps-aux and note the pid of the *second*
(child) gdm. Check that the /var/log/gdm/:0.log has the error message, and
restore the original /etc/X11/prefdm. Repeat the above telinit commands. Or
just kill -2 the strace process. That should make it detach all processes.
Login the usual way (kill -1 the gdm-binary) and inspect
/tmp/gdm.strace.<pid of second gdm>. Search for AUDIT or other pieces of
the error message. Hopefully you will find something like

write(4, "AUDIT ... rejected from local host\n", 35) = 35

Then look at the lines above that line. They tell what the program was doing
before it wrote the error message. Look particularly for lines having
"= -1 E" in them like this:

connect(3, {sa_family=AF_FILE, path="/var/run/nscd/socket"}, 110) = -1 ENOENT (No such file or directory)

There is some chance that this will reveal what is going wrong.

I am not sure if the error will be in the second gdm process. There will
probably be a host of files in /tmp, with different pids, since some of
the programs involved are shell scripts that invoke various sed and similar
commands. "grep execve /tmp/gdm.strace*" will give some indication of
what programs are run by what processes. Look for gdmgreeter-binary,
that seems a likely candidate.

-Enrique

new to Linux, where are they/how do you download files

Posted: 31 Dec 2005 04:30 PM PST

Thank you. I had downloaded a bittorrent client but that doesn't work
either. I had suspected that those files might be some sort of tracking
files, so thanks for confirming that. The only other thing I can think
of is my computer simply doesn't like those files, it's rather
mystifying. I think the most reasonable thing for me is to not even
bother figuring it out, but just have an actual C.D. mailed to me. I
wanted to get it installed over the weekend but I've alread spent half
the weekend just trying to download it.

Thanks.

X is broke... freeing multiple contexts (2)

Posted: 31 Dec 2005 02:49 PM PST

On Sat, 31 Dec 2005 23:49:01 +0100, Wayne Dernoncourt <com> wrote:
 


Trying to sum up the info with some comments:

Distro FC4
HW: Laptop, IBM Thinkpad
Ram 640 MBytes (so much? 512 + 128 ?)

Some problems with DSL, attributed to the ISP, replaced splitter.
Comment:
no details about failure modes. OK, assume laptop disconnected
in practice, i.e. failure outside laptop.

Sound card works for programs running as user, but game "blinkem"
run as root claims unable to open sound card.
Comment:
Assume for now sound card is basically OK, kernel drivers OK,
hotplug/udev stuff OK.

Could not log out as root, after playing "blinkem".
Comment:
Was this a graphical X session? a console bash session? Is "blinkem"
an X application? A framebuffer application?
No account of how "log out". Exit command in a console? Control-D?
Logout gui button in X session manager?

message GDB not able to start/exit - uncertain memory about details.
Comment:
Did the message appear in a console? In a terminal window? In a popup?
GDB? Sounds like a a desktop or session manager application, with a
watchdog function of sorts that tries to start a debugger when a session
process dies unexpectedly. GDB installed? Probably not. Unexpected
process death, could be effect of logout attempt, or "forced a quit" in
"blinkem" What is "forced quit"?
 
Comment:
This message appears to be harmless, but related to a kernel bug fixed in
later kernel releases. Probably unrelated to the other problems.
 
Comment:
Display :17.0 ???? Compputers usually have a "display.screen" of ":0.0".
Some are dual-head, with additional :0.1 or run independent "displays"
showing as :1.0 etc. When running ssh with X forwarding, the display is
typically set to "10.0". Where could this numbe 17 come from?
The window manager is a process separate from X. I don't know how
to trace how the display manager (the program that starts X) determines
what other programs to run for a user that has just logged in. The
windows manager would be one of them, directly or indirectly.


On Sun, 01 Jan 2006 20:49:15 +0100, Wayne Dernoncourt <com> wrote:
 

This seems to verify that the audit message (freeing multiple...)
is unrelated.

It is not clear to me how "gibberish" looks like, or how I could make
any inferences from this statement.
 

Which log file? /var/log/Xorg.0.log ?

This message looks significant. Is the xfs program running?
 

That is right, it's in /tmp as you note ...
 

Yes, this is it.
 

This seems to indicate that the xfs font server just started. Did it die?
The default config file is /usr/X11R6/lib/X11/fs/config.
 

There is no file "fixed" on my system either. As Peter noted, it is
the last resort fallback, but that does not say where its data comes
from. I guess it will make the font server give you *any* font if it
has any at all, but preferring one that is marked monospaced or fixed-width
or something like that.

You have not told if you did "ps -ef | grep xfs", and what the outcome was.
If the server is running, does it not have any fonts? (I suppose you are
using a virtual console (Ctrl-Alt-F1 through F4, for example) to execute
commands when the X server fails to run.

The message "Fatal server error" sounds like it is the X server proper
that is deciding to quit since it does not have a font. No, that is not
likely. Could it be xfs that was outputing this message? Where did you
find it? If you found it in /var/log/messages, there is probably
"xfs[1234]:" or something similar in front of it.

The xfs config file, has among others this:

# where to look for fonts
#
catalogue = /usr/X11R6/lib/X11/fonts/misc:unscaled,
/usr/X11R6/lib/X11/fonts/75dpi:unscaled,
/usr/X11R6/lib/X11/fonts/100dpi:unscaled,
/usr/X11R6/lib/X11/fonts/misc,
/usr/X11R6/lib/X11/fonts/Type1,
/usr/X11R6/lib/X11/fonts/cyrillic,
/usr/X11R6/lib/X11/fonts/TTF,
/usr/share/fonts/default/Type1,
,
/usr/lib/openoffice/share/fonts/truetype,
/usr/share/fonts/japanese/misc:unscaled,
/usr/share/fonts/japanese/misc,
/usr/share/fonts/japanese/TrueType,
/usr/share/fonts/chinese/TrueType,
/usr/share/fonts/korean/misc:unscaled,
/usr/share/fonts/korean/misc,
/usr/share/fonts/korean/TrueType

on my system. How about yours? Do these directories exist? What do
they contain?

In a mail outside the newsgroup:

On Thu, 05 Jan 2006 12:01:00 +0100, Wayne Dernoncourt <com> wrote:
 

I see. :)
 

Yes.
 

Does not seem usefull to me (but no offense, it's OK to provide
things in the hope they be useful to others :) )
 

So you are using Gnome (the Fedora default) and it's probably running.
 

Yeah, but I am not sure if this file will disappear once the xfs server quits,
so I would prefer "ps -ef | grep xfs to know if it is running.
 

OK

I'm dying to hear about what causes the window manager problem (and the :17
message), and also what is going on with xfs. I suggest you try to esablish
as many facts as possible about everything concerning xfs first. If the process
is there and running, "lsof -p <xfs-process-id> shows it is using the file
/tmp/.font-unix/fs7100, the font files are there, as detailed in the config
file, then you could switch to the window manager issue.

My approach to debugging the window manager issue is to use "strace" (install
it from the CDs if you don't have it). You may have to change the file
/etc/X11/prefdm temporarily to put "strace -ff -o /tmp/gdm.strace "
in front of "$preferred "$@" >/dev/null 2>&1". Yes, also change
/dev/null to /tmp/gdm.out" in this line. Examining the strace files is
kind of hard. Write here to get help, I don't know yet if you will need
to do this, depending in the xfs issue.

-Enrique