hello everyone,
i am using slackware 11 and have been for the past 2 years now. it is a server machine so i dont fiddle around with it much...

then last week i realised (while listening to music from another pc) that there were silent gaps while playing the songs on the playlist..
looking around into what happened, i connected a monitor, mouse and keyboard directly to the server and tried to log on from a kdm greeter window... well everything was frozen there... no mouse movement either...
i switched back to the other pc and connected to the server through ssh just to realise (using top command) that the X was consuming all processor resources..

a restart later the pc was working fine again for a few mins... then same thing happens... i really dunno where to start looking for a problem... i mean i ve been listening to music everyday ... and i didnt have these silent gaps until midweek... then X starts hanging.. no wonder there are no resources for samba to share in 'real time' even an mp3 on the network...

i swtched to runlevel 3 for now and among the last resort solution would be to reinstall slack 12 on top (what do u think? too risky?)... but i would like to figure this out instead of just covering it all up...
any hints would be appreciated! where should i start from
thank you in advance!
nass

Is there any attackers? Try blocking them using iptables:

You may wish to scan for rootkits. Just a thought.

how do i scan for rootkits?

i have ports 6000-6002 open intentionally... and would like to avoid losing the option to use them...

You could install rkhunter. It is at slackbuilds.org.

http://slackbuilds.org/repository/12.2/system/rkhunter/

To run rkhunter after you install it you navigate to: /usr/bin

then update it with: # rkhunter --update

to scan: # rkhunter --checkall

so i did run rkhunter and got the following warnings:

i am using slackware 11. Is it possible to get some MD5 checksums on the commands ldd, whatis, adduser and groups, just to verify that there is nothing that have been tampered with?
i guess i could get those checksums from the installation CD but i dont have it here and will not for the next few hours..

The source disks have md5sums but as far as I know they are just md5sums for the packages and not the programs. These warnings are normal from rkhunter however if you are worried why not just replace the programs mentioned with the originals from the source media ?

The md5sums for the packages are available from a mirror for example here are the sums for Slackware 11.

if you tell me this warning as normal then i will look no further...
i either way had very few worries about a rootkit... i try to keep the system well out of intruders reach....

which brings us back to the beginning... what could be wrong with the X and it juices up processor resources...?

You can get Slackware-11.0 packages from various mirrors. e.g. http://mirror.switch.ch/ftp

I have a simple and stupid idea: Reinstall the X packages and see if it cures. If not, we try other ways.

well, i am reviving this topic, after i have done some research but have not found a suitable solution..

my problem, which i managed to dig from dmesg has to do with the nvidia driver rather than with some problem with X..
i came across similar posts several times:

http://www.nvnews.net/vbulletin/show...0&postcount=14
http://www.linuxquestions.org/questi...errors-385117/
http://www.fedoraforum.org/forum/sho...08&postcount=3
this last one is more informative
http://www.nvnews.net/vbulletin/arch...p/t-92255.html

better yet the problem lies with the kernel as nvidia states
http://www.nvnews.net/vbulletin/showthread.php?t=58498

as for me i tried all the kernel boot options that nvidia suggests (nommconf, maxcpus=1,idle=poll)

since my gpu (Nv GF7100 GS) is connected to the pci-e bus, i don't have a NvAGP option to try out... and of course the rest of the options are 64-bit bus related..and as such don't concern me.

so i startup in runlevel 3 at this point ... and if i try to 'startx' i have approx 10-15 secs until X freezes... if that doesn't do the job and i attempt to run 'glxgears', the gears won't even start to rotate and im already frozen ... crap....

what im basically getting at is, anyone knows whats going on with the kernel and nvidia ? if this is a known kernel bug, anyone knows when will it be dealt with?

as a last resort i'm willing to switch back to a rather old nvidia driver i have. the ' NVIDIA-Linux-x86-1.0-9755-pkg1.run'... cause things were working with this driver and its not like ill try to play a game ... but how to i unistall the current (180.29) gpu driver??

thank you in advance for your help
nass

Just out of curiosity, since no one has addressed it yet, what kind of server is it, and does it absolutely require X?

well its a home server and im using it as a file, proxy, web, vpn even vnc server etcetc.... no it doesnt need to have an X server running... especially since i have the vnc server... but still would you like to know the system is not working 100% ? especially a slackware system!

Have you looked through all the usual log files to see if anything is amiss? Can you run top and post the output?

Run the Nvidia installer with --uninstall


Wonder why everything was working, then now not working correctly. Did you recently update any packages or the Nvidia driver. Something had to change between working --- not working.

/var/log/Xorg.0.log might contain some info.

manwichmakesameal, well ill haven't been @ the servers location lately and i will avoid starting X remotely, cause i will not be able to reboot the machine (yes the crashing seems to render rebooting impossible)..
basically top shows that X is using up 99% - 100% of CPU... if it is smth else you would like to know though, ill post top output asap...

disturbed1, likewise i 'll do the same with the output of dmesg... but its very similar to the ones i found in the posts i attached.. the Xorg.0.log contains nothing after a crash, probably because X is in an state that is impossoble to log anything