Hello, I haven't mixed SUSE kde with stock kde... killing won't work... kill -9 <proc number> doesn't work... in fact it is impossible to kill any of the X processes... Yesterday I updated every rpm-package I have installed... Just now It happend once again. This time it was kopete, which caused the trouble (at least I think so): I was sending a message over the icqprotocol and suddenly the the window turned pale and freezed. Then the keyboard got stuck and I did a ctrl-alt-backspace, since all the other keys got barred. The ctrl-alt-backspace let the X disappear (but the process was still running) but after that it didn't restart. Then the usual: changing runlevels, kill -9 and some other useless stuff. Eventually a reset did the "trick".. grrrr! How can kopete upset the kernel in such a way, that one can no more kill any X processes? The keyboard got freezed and I had to do a restart (which didn't work -> reset on the mainboard) For now I will try to install a new kdenetwork3-InstantMessenger package (I already have kdenetwork3-InstantMessenger-3.5.5-41.2). Probably SuSE 10.3 will provide a new one - including kopete 0.12.4.... Any suggestions? New Kernel (the one shipping with opensuse10.3)? It's got to be a kernel problem!?? What might encourage the system to no more react to kill signals? Tasklet/Softirq problem? After all I have a standard openSUSE 10.2 system here on my Athlon X2, it might be possible, that other folks get irritated by the same problem (I am just wondering whether I did something naughty here with my OS. Actually I am certain not to have done anything special)! Is there a big difference between stack kopete and kdenetwork3-instantmessenger? PS: Thank you for the message M Harris! On Monday 30 April 2007 05:13, M Harris wrote:

On Sunday 29 April 2007 09:56, Siegfried Wolkenstein wrote:

...

sometimes kde or x just hangs.... I can not even kill any kde programme after it happens. I am able to kill X by doing a "ctrl-alt-backspace", but after that, X doesn't restart - even if I change the runlevel from 5 to 3 and then to 5 again. After reinitialising runlevel 5, X simply won't start...

This is very unusual. toggling between runlevel 3 & 5 normally fixes things that go crash in the X night.

...

If I try to restart the machine it hangs while trying to kill any programme started in the X-Shells...

You will need to find the process number that has everything tied up... killing them thoughtfully one by one (from the bottom) is one way to handle this... a little experience will tell you which to try first... but obviously anything having to do with X, kde, or anything that may touch the video in any obscure way (gaming, etc). If the normal kill doesn't get it then a hard kill will be required (use with caution): kill -9 <proc number> as root.

Have you mixed and matched SUSE kde with stock kde?

-- Kind regards,

M Harris <>< -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Hi Stan, great to read this! I just found out, that there is a new driver available for NVIDIA and I will install it immediately... I use Nvidia 6600. It could be a graphics driver problem, but it used to work before... How can the graphics-instructions - issued by kopete (not directly of course) - bring down X? How does that work? And also it doesn't happen on a regular basis. Thanks! On Monday 30 April 2007 16:14, S Glasoe wrote:

On Monday April 30 2007 8:41:34 am Siegfried Wolkenstein wrote:

...

Hello,

I haven't mixed SUSE kde with stock kde...

killing won't work... kill -9 <proc number> doesn't work... in fact it is impossible to kill any of the X processes... Yesterday I updated every rpm-package I have installed... Just now It happend once again. This time it was kopete, which caused the trouble (at least I think so): I was sending a message over the icqprotocol and suddenly the the window turned pale and freezed. Then the keyboard got stuck and I did a ctrl-alt-backspace, since all the other keys got barred. The ctrl-alt-backspace let the X disappear (but the process was still running) but after that it didn't restart. Then the usual: changing runlevels, kill -9 and some other useless stuff. Eventually a reset did the "trick".. grrrr!

This sounds like a video driver issue. Where the system totally locks up while doing anything and only a reboot will bring it back. I suggest you re-evaluate your video card driver, monitor settings and sax2 settings in general.

For example, if you are using ATI's fglrx driver, go back to the radeon driver and see if the system is stable. -- Stan -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tuesday 01 May 2007 13:52:32 peter nikolic wrote:

That sounds very much like a failing hard drive rather than an software problem maybe you should get a little more intensive and give that hard disc a real goo test cus it sounds like a garbadge can merchant ..

Pete

Hello Pete, This (hard drive failure) is what I thought as well. But consider this: Kaffeine is recording in this home partition happily even after the crash. So writing to the partition is still possible. It's just opening and/or closing of inodes or files that doesn't work anymore.... Also other partitions lying on the same hard drive work perfectly (reading is doing well (executing ls and so on)). How is this possible... Processes that wait for one specific broken sector get stuck in kernelspace and thus suspending or killing is impossible? Pretty poor behaviour I must say (to be honest I couldn't do it any better (kernelprogrammingwise)). But still, how can i verify that it indeed is a hard drive problem? It could be a reiserfs problem, couldn't it? How can I test my hard drive thoroughly? How can I tell the manufacturer of the drive that it is broken? They are going to test it and they won't find anything wrong. I tried to delete some files and immediately afterwards the crash happend... After the reboot deleting the same files did work without a fuss. And I can imagine the files still dwell on the same sectors or on the same location... off topic: I am very pleased with this mailing list... We seem to get closer to the problem, thank you very much! I can imagine this is a rather boring problem, because it could be some sort of "personal" thing not concerning SUSE (or Linux in general). But still it would be very kind of the kernel or the programmes to provide the user with some useful information, when something like this happens. This is what we can extract of this matter, just as an excuse we don't waste all our time on this rubbish ;-) To this is of course vital... so thanks again!

On Tuesday 01 May 2007, Siegfried Wolkenstein wrote:

...

Hello again,

this is very interresting I believe:

I updated the system successfully to openSUSE 10.3 Alpha 3. Everything except java works well. No problems at all.

But then I started recording Musicvideos... After 4 hours I removed 4 GB of my home directory (rm /home/name/kaffeine/somefiles*) and the system got stuck!

I switched to tty and found out that I could no more access my home directory! Logging in with the username I usually use didn't work anymore... logging in as root did work.. I was able to do everything except killing X processes (of course most of them access the /home/user* folders) and accessing the folders in my home partition didn't work either...

/dev/sda6 on /home type reiserfs (rw,acl,user_xattr,usrquota)

After all I had to do a reset!

So what can we read into this?

Thanks

-- Suse 10.3 alpha2 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tuesday 01 May 2007 18:01, peter nikolic wrote:

On Tuesday 01 May 2007, Siegfried Wolkenstein wrote:

...

I tried to delete some files and immediately afterwards the crash happend... After the reboot deleting the same files did work without a fuss. And I can imagine the files still dwell on the same sectors or on the same location...

Hi .

I think the best way would be to get the makers test suit off there web site ,

Which filing system are you using ? I use Reiser and find it solid untill you get a slight problem with the disc then the fun starts and these crashes you are getting are typical BUT having said that i have EXT3 on the laptop and i am just waitingfor an excuse to can it and get back to sanity of Reiser

Curious.......Just what is it about EXT3 that makes you want to return to the "sanity" of Reiser? You just said that Reiser starts to give trouble when there is a slight problem with the disk. Bob S -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tuesday 01 May 2007 23:18, M Harris wrote:

On Tuesday 01 May 2007 21:40, Bob S wrote:

...

Curious.......Just what is it about EXT3 that makes you want to return to the "sanity" of Reiser?

bob can answer for himself...

... but the reiserfs is organized on a b-tree and allocates only as much actual disk-space as it needs... so no wasted space--ie. small files residing in large fixed-size clusters which take longer to find... ext3 is journalled, but its slower, and it wastes space... esp on systems with thousands of small files. typically reseirfs is typically more stable than ext3, and it also doesn't have the long file check times associated with ext(2) ext(3). The old ext filesystems (for general purpose) are "insane".

Thanks for your input M Harris. I guess that you think Peter has the same reasons you do. Just thought I would ask him. I guess that the Rieser file system would have been great for saving space on my 250MB disk years ago, Probably not a significant worry these days with 250GB disks. As far as speed is concerned with today's super fast CPU's and ram I doubt anybody could see a speed difference. Don't know where you get the "more stable" statement from; must be personal experience. I've been on this list for many years and have seen many posts about problems with Rieser. Don't ever recall seeing anything about ext3 causing stability problems. Anyway, not my intention to argue with anybody over the merits of this vs that. To each his own. Whatever floats your boat they say. That's what Linux is all about. I am just curious as to why anyone would make statements about ext2 &3 being an insane FS. Bob S -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Wednesday 02 May 2007 22:58, M Harris wrote:

On Wednesday 02 May 2007 21:35, Bob S wrote:

...

I am just curious as to why anyone would make statements about ext2 &3 being an insane FS.

heh, insane is of course an extreme word... mostly intended for humor... like calling M$ e v i l .....

You will notice the the speed difference under load of course.

However, the *big* insanity of ext2 ext3 is the wasted disk space. Now, it must also be said (in fairness) that the disk partition must be 30-40 megs large just to hold the journal (for reiserfs). Aside from that overhead... the fixed cluster size of ext2 ext3 makes the fs insane... from a modern perspective.

Stable was a poor word choice... sorry. The better word would be reliable. The reiserfs is much more reliable from a recovery standpoint. The ext2 ext3 is more likely to get hammered than reiserfs, and recovery for a not clean disk shutdown is faster (way faster esp for large disks) for reiserfs. The only advantage that I can see for ext3 is that the fs journals both the meta data and the data, whereas reiserfs only journals the meta data... which of course is usually what gets clobbered. Bottom line, size and speed make ext3 less desireable... ok, maybe not insane. :-)) OK ....Points taken. Thanks for your opinions.

Bob S. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Wednesday 02 May 2007, peter nikolic wrote:

On Thursday 03 May 2007, Bob S wrote:

...

On Tuesday 01 May 2007 23:18, M Harris wrote:

...

On Tuesday 01 May 2007 21:40, Bob S wrote:

...

Curious.......Just what is it about EXT3 that makes you want to return to the "sanity" of Reiser?

I go to a customers site turn on the laptop and hey presto i am stuck there not able to use it because ext3 has deciede it needs to do an fsck oj the filing system now it only an 80Gb drive in a 64 bit machine

Quite true. I've had this happen to me on a 160 gig drive which was quite full. It took forever. I went back to reiserfs. All the stories about the beauty of "Open Source" were proven wrong by Suse's dropping Reiserfs because a) nobody was maintaining it, and b) hans got thrown in the slammer, and c) it doesn't "scale well". The fact that it was good enough for 5 generations of Suse before they decided to revert to something clearly inferior on the weight of three lame excuses seems to fall on deaf ears. Reiserfs has been rock stable for me for years, I've never lost any data with it, but I have with ext2 and fully expect to with ext3. -- _____________________________________ John Andersen -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Thursday 03 May 2007, Jerry Feldman wrote:

. But, that is a configurable parameter you can change with tune2fs(8). The command 'tune2fs -c -1 /dev/hda1' will turn off the max-counts for that partition. The -i option is a time interval. All file systems should be checked periodically.

The point is that turning these OFF in ext2 file system will sooner or later bite you. Not so with reiserfs, where the check happens each boot but is so fast its not a problem. You don't need to read each and every block and compare pointers forward and backward to discover disk problems. Other ways have been invented to detect these. -- _____________________________________ John Andersen -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org