How interesting, my first crash!

English Install/Boot/Login
#1

OK, so not really a crash because tty 1-6 work fine. I was extracting BitTorrent when things just stopped working quickly, the preview pane wouldn’t close and I was getting KDE .so errors when I tried to move files to the trash. So I figured that something must be wrong with X and I decided that shutting it down and having it reinitialize would probably help it (Can you spot the recent Windows user). When X restarted it came up with a report bug dialogue box. I recently used YaST to try and install the MAD MP3 libraries, but this came back and said it could not do this and gave me options to try. I think it may have been that library install that did it. Anyway, X will not start … well X starts but KDE will not, because I get a blank screen and then my mouse cursor.

When I tried looking at the processes that might be problematic I got a process on tty7 that when I gave it the kill cmd it shifted me right to tty7 and a kde login screen, but it just hung and gave me the report bug screen. Oh and I could not give enough info to make the bug thing worth its time.

The process was:

/usr/bin/Xorg -br -nolisten tcp :0 vt7 -auth /var/lib/xdm/authdir/authfiles/A:0-9IfR4y

I am running OSuSE 11.2 on an AMD 1.1 with 1.5GB RAM and 2 X 160 WD Caviar HDDs and a radeon 9250 video card. If you can help let me know. My linux is set up and running as a learning experiment and I do not have any hard feelings about dumping and reinstalling, but if I can learn how to fix things like this I would like that.

Cheers

#2

You mention a bunch of happenings, not sure what your main problem is. Are you trying to get your video to work normally?

#3

A quick way to add “more information” to help you assess what the problem may be, is to boot to run level 3 (ie a text mode - do this by pressing “3” (no quotes) in the grub/splash initial boot menu/screen) and boot to a full screen text mode. Once there, login, switch users to root, run yast, and add a brand new user.

Then logout and log back in as this new user and check to see if this new user can run KDE. If they can, then you know something is messed up (or set different) in your regular user’s KDE configuration causing the problem.

You can always delete this extra user later.

#4

Pigloo wrote:
> OK, so not really a crash because tty 1-6 work fine. I was extracting
> BitTorrent when things just stopped working quickly,

did you fill up the root partition? to me it sounds like you did (but,
i’ve been wrong before)…

i don’t know where the unpacker you were using puts its transient
files duing unpacking, but i guess you need to clean out /tmp

use the forum’s advanced search to learn how to do that safely…like
windows of old, there are probably files in temp which ARE in
use…but, i think you are gonna find a HUGE file (which was in the
midst of being unpacked) date stamped about the time you crashed…

nuke it (with rm) and try to reboot…if lucky you will be able to
sign in and there won’t much else murdered by running the system out
of disk space…

be advised: no *nix fails gracefully faced with that…remember, to
*nix everything looks like a file…when it can’t write to any file
it dies in agony trying!!

meaning, you may have munged config files, or what not…

morale of the story: disks are cheap and pix/music/movies are HUGE…

btw: yes i spotted “the recent windows user” immediately…no
problem…but wanted to say: that YOU recognized it is
progress…congrats!!

–
palladium

#5

Thanks all, I will try each of these, and get back to you. And as a quick answer to the video thing. I believe it is just the video that is messed up as all the command line tty’s work and I can still see the graphics until the KDE desktop tries to load.

I will let you know, thanks

#6

I have two questions about that:

  1. Is the radeon driver active? If you don’t know, run this command and see if you get a response:
lsmod | grep radeon
  1. Have you installed the proprietory fglrx drivers? If you don’t know, run this command and see if it shows the fglrx RPMs:
rpm -qa | grep fglrx
#7

When I tried looking at the processes that might be problematic I got a process on tty7 that when I gave it the kill cmd it shifted me right to tty7 and a kde login screen

X is configured to be restartet when not running. You have to switch to runlevel 3 with “init 3” to stop it.

#8

OK,
So I ran through all of your suggestions:

oldcpu: I created a user acct and logged in under it. More clues … I got the same black screen with just the mouse pointer, as I got for my normal user; but because it was the first time this user logged in, I got the SuSE splash page, and when I clicked on the link Firefox started up and took me to the OSuse site. So the programs are working in the background, but I cannot see the jumpstart menu or any of the background. So maybe this rings some bells?

palladium: I ran df at the promp as root and things did not look full. 24% and 28% was the most I got. I haven’t downloaded any large files that I know of, except a torrent file of Mac OS X for my kid. but I deleted that under dolfin in kde. A ls -l of /tmp showed 200 files, so I will look through the advanced forums and see if I can dump some of this stuff, I recognize a lot of it as old files so I should be able to dump it.

swerdna: I ran the “lsmod” and got:

radeon_________368000_______________2
drm___________184640_______________3__________radeon

when I ran the rpm cmd you suggested I got no response from the OS, just went to a new command prompt. So it looks like my driver is running, but the fglrx I cannot say for sure.

Thanks very much to everyone, I will check back soon after I try a couple more things.

Cheers

#9

Pigloo wrote:
> palladium: I ran df at the promp as root and things did not look
> full. 24% and 28% was the most I got. I haven’t downloaded any large
> files that I know of, except a torrent file of Mac OS X for my kid. but
> I deleted that under dolfin in kde. A ls -l of /tmp showed 200 files,
> so I will look through the advanced forums and see if I can dump some of
> this stuff, I recognize a lot of it as old files so I should be able to
> dump it.

no, it it wasn’t 100% full–my guess was wrong and there is absolutely
no sense in you picking through there to delete stuff…it is (imo) a
good idea to clean out everything in /tmp at each boot, do that this
way: ‘Clear Temp Files at Boot - openSUSE Forums’
(http://tinyurl.com/yzmzp5b)

ok, back to your problem (scratches head)…i don’t know anything
about those “MAD MP3 libraries” except you think they might be the
problem…have you uninstalled them?

have you yest looked in the logs to see if there are any hints? maybe
in /var/logs/messages you will find a nugget…each line in there
begins with a local date/time…scroll to near the time of the crash
and have a ball looking for key words like “error” or “warning”…

you can do that in (for example) kate, just open her…wait, you
can’t get to KDE…hmmmm, do you know vim, vi, what? i use mc
myself…if you ever used Norton Commander, ‘mc’ (Midnight Commander)
is a wonderful clone, its on the DVD and simple to install…

hmmmmm…failing everything else you should be able to look with:

sudo cat /var/log/messages | less

“less” lets you scroll the list and the ‘end’ button should get you to
the last entry, ‘home’ to the first and the arrows up/down or ‘Page
Up’ and ‘Page Down’ a “page” at a time, and ‘q’ to quit

you could also skinny down the list looking for a particular word,
like error:

sudo cat /var/log/messages | grep -i error
or
sudo cat /var/log/messages | grep -i warning

but then you are gonna get all of them for all of the days messages
has on record…

or you can skinny the list with a word and a date

sudo cat /var/log/messages | grep -i error | grep -i ‘Mar 5’ | less

that “Mar 5” is the format in my messages for a date with two
spaces between the month and the day, if the day’s date is 1 through 9…

see
man sudo
man cat
man grep
man less

Have a lot of fun!

–
palladium

#10

Fglrx has not been installed.

Try to re-establish the video with this command in the console. It must be issued as root:

sax2 -m 0=radeon

then reboot or run this command:

startx

.

Then you can swap over to the fglrx drivers using Yast’s software module to install the RPMs.

#11

Swerdna: No change after running the sax2, but the config tool did start X and give me a very nice looking background until I closed it, and restarted X, then I was back where I started.

palladium: I ran the command:

cat messages | grep error | less > blah/blah/blah error.txt

Where blah … is my home directory’s Documents directory. The file was extensive, most of it I recognized as me trying to copy files from a USB external hdd. I believe that is what it is sr1?? the dates seem right and I did have trouble moving files as my windows system had a little problem with a Trojan. and then didn’t copy files well to the external drive.

However, on 4 Mar, the day this all started, I began to get npviewer.bin segfaults and plasma-desktop segfaults. From 4 mar to now that is what all of my errors are. I know what the plasma-desktop is, but the npviewer is new to me. I expect these are my troublesome twins.

Thanks for your patience.

#12

On 03/06/2010 04:46 PM, Pigloo wrote:
>
> Swerdna: No change after running the sax2, but the config tool did
> start X and give me a very nice looking background until I closed it,
> and restarted X, then I was back where I started.
>
> palladium: I ran the command:
>
> cat messages | grep error | less > blah/blah/blah error.txt
>
> Where blah … is my home directory’s Documents directory. The file
> was extensive, most of it I recognized as me trying to copy files from a
> USB external hdd. I believe that is what it is sr1?? the dates seem
> right and I did have trouble moving files as my windows system had a
> little problem with a Trojan. and then didn’t copy files well to the
> external drive.
>
> However, on 4 Mar, the day this all started, I began to get
> npviewer.bin segfaults and plasma-desktop segfaults. From 4 mar to now
> that is what all of my errors are. I know what the plasma-desktop is,
> but the npviewer is new to me. I expect these are my troublesome
> twins.

Program npviewer is a PDF viewer. I have not had crashes from it, but I have
seen it fail to close when the PDF is closed and have it consume nearly 100% of
one of my CPUs. I have uninsatlled it and now use Adobe Reader. Although it
takes a long time to load, at least it closes gracefully.

#13

I suggest you move on to the fglrx drivers. There’s one more piece of info needed first. What return do you get from this command:

uname -r
#14

Oh, and I forgot, this command too, also as root:

lspci | grep VGA
#15

And sorry I need this too:

uname -p
#16

swerdna: I ran the cmd’s that you asked for:

linux-e329:~# uname -r:
2.6.31.12-0.1-default

linux-e329:~#uname -p:
athlon

linux-e329:~#lspci | grep VGA:
01:00.0 VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200 Pro] (rev 01)

So I read the man page on uname and did not see the -r flag. What does that do and what does the reply mean?

Also when I first entered the lspci cmd I missed the fact that the VGA was in upper case and got no retrun so I ran it without arguments and got the whole list which was not long, but at the end there were two lines dealing with my ATI card, the one you have and this one:

01:00.1 Display Controller: ATI Technologies Inc RV280 [Radeon 9200 Pro] (Secondary) (rev 01)

It seems odd to me to have this. Is that normal?

#17

My wife’s PC has an ATI RV280 (Radeon-9200Pro) with KDE-4.3.5. In her case, to get special desktop effects running, we had to update her opensource ATI driver from the X11 : Xorg repository.

I also note there was an update to the radeonhd driver from Novell/SuSE-GmbH in the last few days, which may theoretically have caused a problem ( ? ) as I did NOT accept that update on my wife’s PC as I did not want to replace the driver that I obtained from the X11 : Xorg repository. I am wondering if some update may have caused this 1st crash hiccup ?

There is guidance in post#11 here where I obtained the driver for my wife’s PC from, and in post#1 on that same thread there is guidance on how to change one’s graphic driver: openSUSE Graphic Card Practical Theory Guide for Users - openSUSE Forums

#18

@Pigloo. I was going to lead you through installing the fglrx drivers. But given what oldcpu said, it might be wrong to install the fglrx drivers before having a last go at radeon to see if you can get a GUI going.

And I see this entry for your card in the openSUSE Hardware Compatibility List:

Radeon 9200

SUSE Linux Release(s): 10.0 - 10.2 - 10.3 - 11.0 - 11.1 - 11.2 [In 11.1 special desktop effects could be obtained with “radeon” open source driver. On 11.2 for special desktop effects to work well with openGL driver one needs to update to cutting edge drivers from Index of /repositories/X11:/XOrg/openSUSE_11.2 repository (ie mesa, xorg-x11-driver-input, xorg-x11-driver-video and xorg-x11-driver-video-radeonhd). If one does not do this, there will be artifacts on the desktop and one will find it necessary to disable special desktop effects]

Try this one last time: reboot and at the grub prompt select openSUSE (probably automatic) and press the “3” integer key, then the “Enter” key. You should boot to a console where you should log in as root. Then run this command to initialise the radeon cofig, it’s slightly different from the earlier sax2 command:

sax2 -r -m 0=radeon

Then bring the system to runlevel 5 with this command:

init 5

Your graphical login manager should appear. Log in and test the driver.

@oldcpu: I am nearly floundering here, I’m not a video person and find the ati stuff difficult. Would you suggest a different course than radeon followed by radeonhd?

#19

I use the “radeon” driver on my wife’s PC as it works better than the “radeonhd” for her ATI RV280 (Radeon-9200Pro) , BUT we had to update her driver to the one from the X11 : xorg repository. Thats easy to do and is described in post#11 from here: openSUSE Graphic Card Practical Theory Guide for Users - openSUSE Forums

… all one has to do, with PC connected to the internet, is first ADD that repository with root permissions with:

zypper ar http://download.opensuse.org/repositories/X11:/XOrg/openSUSE_11.2/ newx11repository

then go to yast (do not use zypper) and under YaST > Software > Software Management search for and install the X11 : / Xorg version of mesa, xorg-x11-driver-input, xorg-x11-driver-video and xorg-x11-driver-video-radeonhd. Do not be fooled by yast when it says they are already installed. Those are the OSS versions that are already installed and one needs to replace those with the X11 / Xorg versions. If I had $100 for each person who stated those were installed already from X11 / XOrg (when they were not) I could afford a nice vacation to some place exotic :slight_smile:

Then after they are updated, remove the repository with

zypper rr newx11repository

Thats important! It should NOT be forgotten.

Then reboot the PC to run level 3 (don’t go there via “init 3” as there is always a risk that an xorg module will not unload properly). When the splash grub menu appears during the boot press ‘3’ (no quotes) so that a ‘3’ (no quotes) appears in the options line, then boot normally and login as a regular user. Then type “su” (no quotes - enter root password) to get root permissions. Backup the existing /etc/X11/xorg.conf file (if any) and then type:

sax2 -r -m 0=radeon

and configure X under sax2. Then exit sax and then type:

shutdown -r now

to reboot. Do NOT press ‘3’ this time.

Thats what worked on my wife’s PC.

#20

“Sigh” Ok, so I added the newxrepo… and installed the mesa, …driver-input, …driver-video, and the driver-video-radeonhd. tried to remove the repository with zypper but it told me it could not find it. I checked back later with yast and didn’t see it, but … ?

I then ran the Sax2, and accepted the config that X recommended as it all seemed good, accept for the video card it saw was a 5600 series radeon. (thoughts on that??)

after the shutdown command it again booted to “report a bug” and a black screen.

I would like to reiterate (horrible word as iterate means to utter again or repeat) that when I checked the event log the errors that were logged since my problem have been almost exclusively segmentation errors (segfault) of the plasma-desktop. The most recent are today as I am trying the suggestions all of you have made. The error message was extensive with a lot of memory addresses in Hex so I didn`t want to type it all out, but here it is and there are a lot of this one:

kernal: [1023 124450] plasma-desktop[7237]:sefgault at 66 ip b74e53ff sp bfda03bc error 4 in libplasma.so.d.0.0 [b734e000+zb3000]

So with so many of these I am betting the problem is with the libplasma.so.d.0.0 thing.

If I am not seeing something due to my inexperience I fully understand. But when I said I was getting lots of these errors earlier no one said much about it, but it must be significant.

Your opinions enthusiastically encouraged and gracefully accepted.