Go Back   nV News Forums > Linux Support Forums > NVIDIA Linux

Newegg Daily Deals

Reply
 
Thread Tools
Old 03-30-08, 08:20 AM   #1
jusbug
Registered User
 
Join Date: Mar 2008
Posts: 1
Default dont hate me for this question

i'm very newbie to linux and ubuntu studio, i tried to install nvidia 169.12 but...
i looked everywhere and always poeple refer to envy utilities, but my probleme is i dont have internet connection.

here's the log file content:

nvidia-installer log file '/var/log/nvidia-installer.log'
creation time: Sun Mar 30 04:25:14 2008

option status:
license pre-accepted : false
update : false
force update : false
expert : false
uninstall : false
driver info : false
precompiled interfaces : true
no ncurses color : false
query latest version : false
OpenGL header files : true
no questions : false
silent : false
no recursion : false
no backup : false
kernel module only : false
sanity : false
add this kernel : false
no runlevel check : false
no network : false
no ABI note : false
no RPMs : false
no kernel module : false
force SELinux : default
no X server check : false
force tls : (not specified)
X install prefix : (not specified)
X library install path : (not specified)
X module install path : (not specified)
OpenGL install prefix : (not specified)
OpenGL install libdir : (not specified)
utility install prefix : (not specified)
utility install libdir : (not specified)
doc install prefix : (not specified)
kernel name : (not specified)
kernel include path : (not specified)
kernel source path : (not specified)
kernel output path : (not specified)
kernel install path : (not specified)
proc mount point : /proc
ui : (not specified)
tmpdir : /tmp
ftp mirror : ftp://download.nvidia.com
RPM file list : (not specified)

Using: nvidia-installer ncurses user interface
-> The file '/tmp/.X0-lock' exists and appears to contain the process ID '5232'
of a runnning X server.
ERROR: You appear to be running an X server; please exit X before installing.
For further details, please see the section INSTALLING THE NVIDIA DRIVER
in the README available on the Linux driver download page at
www.nvidia.com.

please guyz, what's this Xserver and how do you exit and install the driver.

please guyz, i realy need your expertise.

PS: i have 8600 GT, and no internet connection.
jusbug is offline   Reply With Quote
Old 03-30-08, 09:03 AM   #2
zap
Registered User
 
Join Date: Feb 2003
Location: St. Petersburg, Russia
Posts: 20
Default Re: dont hate me for this question

It tells you that the X server (the graphics interface) is running. The generic nvidia drivers for some strange reason require to be installed with a stopped X server.

Here's a howto install nvidia drivers from the generic package:
http://ubuntuforums.org/showthread.php?t=57368
zap is offline   Reply With Quote
Old 03-30-08, 03:46 PM   #3
xbobmx
Registered User
 
Join Date: May 2003
Posts: 191
Default Re: dont hate me for this question

Quote:
Originally Posted by zap
for some strange reason
You're installing graphics drivers that X depends on. If you change the drivers that it's currently using, some things may stop working (OpenGL apps won't start, at least; I've heard reports of X crashing on some systems). Furthermore, if you upgrade the kernel module the old one can't be unloaded -- and the new one can't be tested -- while X is running. So even if you install while inside X and restart X, unless you do `rmmod nvidia` while X is not running it won't work. Or, if for some reason the module doesn't load properly (unresolved symbols or such), then X will just fail to start, and you need to fix the installation while X is not running (as opposed to the installation simply failing without touching your configuration, so you can start X, research/fix things and try again later).
xbobmx is offline   Reply With Quote
Old 04-01-08, 12:54 PM   #4
zap
Registered User
 
Join Date: Feb 2003
Location: St. Petersburg, Russia
Posts: 20
Default Re: dont hate me for this question

Well, here on Fedora I use the livna pre-packaged RPMs and I never had any problems installing or upgrading them directly in X11 (even using the GUI package installer).

Maybe it would make sense to propose the user a reboot after installing the drivers from X11 (and perhaps issuing a warning before starting installation with X11 active), but I don't think forcing the user to install from the text console (many newbie Ubuntu users don't even know about it) is a good idea.

Or alternatively, maybe it makes sense for the nvidia installer to shut down X11 itself (after asking the user about it, of course) and doing everything in text mode without user intervention.

Anyway, it's just a usability question.
zap is offline   Reply With Quote
Old 04-01-08, 01:27 PM   #5
AaronP
NVIDIA Corporation
 
AaronP's Avatar
 
Join Date: Mar 2005
Posts: 2,487
Default Re: dont hate me for this question

Quote:
Originally Posted by zap
Well, here on Fedora I use the livna pre-packaged RPMs and I never had any problems installing or upgrading them directly in X11 (even using the GUI package installer).
The installer needs to be able to load the new kernel module to make sure it works, which requires unloading the old kernel module first, which can't be done while X is running.

Quote:
Maybe it would make sense to propose the user a reboot after installing the drivers from X11 (and perhaps issuing a warning before starting installation with X11 active), but I don't think forcing the user to install from the text console (many newbie Ubuntu users don't even know about it) is a good idea.
This isn't good enough, unfortunately, because there are many cases where Linux kernel modules may appear to build correctly but then fail to load. Currently, the installer will restore the old (working) kernel module if that happens, but if it forces a reboot, it can't do that and the user is stuck in text mode.

Quote:
Or alternatively, maybe it makes sense for the nvidia installer to shut down X11 itself (after asking the user about it, of course) and doing everything in text mode without user intervention.
This would be a good solution, but the mechanism for starting and stopping the X server varies by distribution and it is, in general, impossible for nvidia-installer to kill the X server and make sure that it stays killed.

Quote:
Anyway, it's just a usability question.
It's a usability question that it's ultimately up to the various distributions to solve.
AaronP is offline   Reply With Quote
Old 04-04-08, 08:55 AM   #6
Niva
iBURNu
 
Niva's Avatar
 
Join Date: Jan 2008
Location: Houston, Tx
Posts: 26
Send a message via ICQ to Niva
Default Re: dont hate me for this question

Aaron thanks a lot for replying here, I read this earlier and thought to myself exactly as stated earlier: well I didn't have to shut down X in Fedora when installing the livna packages... I've had no issues at all and the packages have even updated twice since then using regular system update to the 169.12 driver but I'm questioning whether there are underlying issues which I simply haven't noticed.

Obviously since my system is stable and everything runs great I'm not concerned about it. Given what you said I'm questioning why and how it all worked great... so I have a question about the installer in general:

Rather than trying to kill X can't the nvidia installer simply check if x is alive and require it be killed before the install?

This would avoid the fiasco of trying to kill x in different ways to accomodate all the different distributions out there. A pop up message could display a few different ways to kill x to help out with the most common distros.

There has to be a somewhat elegant method to get this done and reduce install issues for everyone involved especially new folks.
__________________
GTS 512 + AMD Phenom + 8 Gigs of RAM + 64 Bit Linux = Goodness!
http://www.eternalseven.com/
Niva is offline   Reply With Quote
Old 04-04-08, 10:47 AM   #7
pe1chl
Registered User
 
Join Date: Aug 2003
Posts: 1,026
Default Re: dont hate me for this question

Indeed it has been discussed before.
Technically, there is no need to shutdown X to replace the files that need to be replaced during a driver update. In Linux, it is perfectly possible to replace a file that is open. In Windows there are issues with that, but they are not relevant to Linux.
What Nvidia (AaronP) claims is correct too. The driver installer cannot test if the new configuration is going to be working, and it can happen that the installed driver crashes the X server the next time it is started. That would mean the user starts out with a working X server, then does a driver update, restarts his machine and finds himself in textmode or with a black screen.
Of course this is not really desirable either.

The packages distributed with Linux distributions like SuSE or RedHat are not affected by this problem because they can do the installation steps that are most likely to fail (compiling the sources and linking the loadable module) on the maintainer's system and ship the readymade module as part of the package.
This means the chance that your X server would not work after the upgrade is less with that method than with the full Nvidia driver installer.

It would be possible to check for some of the problem areas, e.g. by compiling a "test module" and trying to load it to see if the components required for that (compiler, headerfiles, linker etc) are present on the system and have the correct version.
It could probably not be made 100% foolproof.

So in the end, the decision is between:
- require everyone to learn how to shutdown X and perform the driver operation in textmode, with full recovery in case of installation problems
- perform a "hope for the best" installation from within X, and require those that encounter a failure to learn how to remedy it without re-installing their entire system

Apparently Nvidia have chosen the first alternative and the package distributors have chosen the second.
pe1chl is offline   Reply With Quote
Old 04-04-08, 08:37 PM   #8
txf
Registered User
 
Join Date: Feb 2008
Posts: 272
Default Re: dont hate me for this question

Any chance that in the future the nvidia installers use dkms? It seems like a pretty decent solution. When I upgrade my kernel dkms automatically builds the module before loading the xserver.

On my ubuntu system envy and the repo packages use dkms. Its pretty neat watching it automatically building the kernel module when I upgrade or install new kernels
txf is offline   Reply With Quote

Old 04-04-08, 10:12 PM   #9
kwizart
Registered User
 
Join Date: Feb 2005
Location: Paris, France
Posts: 129
Default Re: dont hate me for this question

Quote:
Originally Posted by txf
Any chance that in the future the nvidia installers use dkms? It seems like a pretty decent solution. When I upgrade my kernel dkms automatically builds the module before loading the xserver.

On my ubuntu system envy and the repo packages use dkms. Its pretty neat watching it automatically building the kernel module when I upgrade or install new kernels
Fedora 9 (actually, via the rpm.livna.org repository ) will use something quite similar named akmods (for automatic kmods) see here: http://rpmfusion.org/Packaging/KernelModules/Akmods
This not only permit to build a kernel module but also a rpm "on the fly" when there is a new kernel update. It allow to redistribute the rpm with all workstations in the neighborhood. (no need to install a compiler and kernel-devel on many host).
This remains compatible with our pre-built packages scheme (kmod-nvidia etc..).

The main advantage is to allow to have packaged nvidia.ko with either custom kernels or kernel-rt (available at planetccrma) and others.
The pre-requirement remains to have a kernel packaged (with rpm). And as it is easier to have rpm'd kernel (IMO)... But the framework is portable for others distributions (using kmodtool).


About what I've been said above. I never had any problem when upgrading the driver version on a session. But indeed, I haven't waited too much time to have to see what was the improvements(/regression) with the new driver which leads to a quick reboot.
I will try to see what can be improved in this domain. (update on next boot/shutdown handled by yum/rpm ?! I think something similar might be on schedule ).
kwizart is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


Similar Threads
Thread Thread Starter Forum Replies Last Post
PhysX question Spartan 117 NVIDIA GeForce 600 Series 5 11-09-12 09:45 PM
Dragon Age Origins question Cylphid Gaming Central 3 05-30-12 07:15 PM
1st post - XP Restore question Needa916 Microsoft Windows XP And Vista 2 09-12-02 05:24 PM
Newby question about nforce.... superconti NVIDIA Linux 2 09-07-02 11:46 PM
Dual boot question.. Tactics General Hardware 15 09-07-02 10:39 AM

All times are GMT -5. The time now is 05:17 PM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.