KDE - Processor is always at full load

[sekhemty]

Hello everyone, I'm a linux user by some years now, but I'm rather new to KDE.

My last system was Ubuntu 10.04, and with the release of the new LTS versions I decided to switch to Kubuntu. I really like KDE, and I'm using it with satisfaction from nearly a month.

However, it is a couple of days that it started to behave a little strange, and I can't figure why.
The problem is that the system is extremely slow, I noticed that even after a fresh start, the processor load is always at 100%. I mean, I have a dual-core processor at 2.40 GHz, and both of them are at 100% usage. RAM (I have 3 GB plus 1 GB as swap) is not affected by this overloading, its usage is as usual.

In these conditions is really impossible to use the whole system: Plasma Desktop is really slow (it takes a couple of minutes to open the menu or to use anything else), applications don't even start, and basically the only thing I managed to open is the System Monitor; here, some processes are listed as "waiting for disk".

I don't know if this is related, but I tried to open a session with Ctrl+Alt+F1, and without typing anything, I continuously get the messages you can see in the picture. I suspect that the problem is hard-disk related, but I can't figure out if this can be true and how to solve it.

Can anyone help me?

[GreyGeek]

Looks like a possible kernel regression: https://bugs.launchpad.net/ubuntu/+s...ux/+bug/730192

If you can try a previous kernel or install a newer one, see if the problem goes away.

[sekhemty]

Thank you for your answer.

My most up-to-date kernel is 3.2.0-24-pae-generic. The problem was first noticed using this one, and following your advice i tried also with 3.2.0-23-pae-generic, but it persisted. It looks like that the kernel is not the cause.

I don't know if this can help in understanding the problem: I noticed that a couple of processes are always listed as waiting for disk, namely jbd2/sda5-8 and jbd2/sda6-8.
Sda5 and 6 are the partitions where, respectively, are installed the root and the home folder.

[GreyGeek]

The JBD is the journaling block device that sits between the file system and the block device driver. The jbd2 version is for ext4. It writes a log entry every five seconds. If it is "waiting" then the file I/O has problems, which can be either software or hardware. I decided to check the software possibilities first.

A change in kernel didn't affect the problem. You can use the following commands to see how big the logs are, and the last few entries in kern.log. Here is what my setup shows. A log bigger than a few megabytes would suggest a problem. You can open the log and note the time stamps to see if there is a point at which the number of lines increased out of expectations. That may point to the time when the problem appeared. (If your system to too slow to use perhaps booting a LiveUSB or Li veCD and using it may help.)

~$ ls -lS /var/log/*log | head-rw-r----- 1 syslog adm 1303405 May 19 14:39 /var/log/kern.log
-rw-r--r-- 1 root root 552744 May 19 14:34 /var/log/dpkg.log
-rw-r----- 1 syslog adm 316303 May 19 14:55 /var/log/auth.log
-rw-rw-r-- 1 root utmp 292292 May 12 21:33 /var/log/lastlog
-rw-r--r-- 1 root root 225634 May 19 14:37 /var/log/uvcdynctrl-udev.log
-rw-r--r-- 1 root root 173756 May 19 14:38 /var/log/pm-powersave.log
-rw-r----- 1 syslog adm 133787 May 19 14:55 /var/log/syslog
-rw-r--r-- 1 root root 53101 Dec 31 23:27 /var/log/bootstrap.log
-rw-r--r-- 1 root root 34761 May 19 14:41 /var/log/Xorg.0.log
-rw-r--r-- 1 root root 32032 May 12 21:33 /var/log/faillog


jerry@jerry-Aspire-7739:~$ tail /var/log/kern.log
May 19 14:37:43 jerry-Aspire-7739 kernel: [ 25.536681] ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
May 19 14:37:49 jerry-Aspire-7739 kernel: [ 31.145086] postgres (1398): /proc/1398/oom_adj is deprecated, please use /proc/1398/oom_score_adj instead.
May 19 14:37:53 jerry-Aspire-7739 kernel: [ 35.298962] vboxdrv: Found 4 processor cores.
May 19 14:37:53 jerry-Aspire-7739 kernel: [ 35.299301] vboxdrv: fAsync=0 offMin=0x334 offMax=0x2224
May 19 14:37:53 jerry-Aspire-7739 kernel: [ 35.299369] vboxdrv: TSC mode is 'synchronous', kernel timer mode is 'normal'.
May 19 14:37:53 jerry-Aspire-7739 kernel: [ 35.299371] vboxdrv: Successfully loaded version 4.1.14 (interface 0x00190000).
May 19 14:37:53 jerry-Aspire-7739 kernel: [ 35.816185] vboxpci: IOMMU not found (not registered)
May 19 14:37:53 jerry-Aspire-7739 kernel: [ 35.830255] wlan0: no IPv6 routers present
May 19 14:37:56 jerry-Aspire-7739 kernel: [ 38.663683] init: plymouth-stop pre-start process (1730) terminated with status 1
May 19 14:39:20 jerry-Aspire-7739 kernel: [ 121.970219] hda-intel: IRQ timing workaround is activated for card #0. Suggest a bigger bdl_pos_adj.

"tail" defaults to 10 lines. If you want to show more add the parameter "-n", where n is the number you want showing.
tail -25 /var/log/kern.log

Looking at the photo of your display, my first thoughts were either a bug in file system, kernel, or it is hardware. One last check is to determine how much free space your partitions have. If your free space is less than 5% (or is 0%) then that's your problem. If you have plenty of free space then it would suggest that your HD is going south.

[sekhemty]

I tried from a LiveCD.

My kern.log is 7.2 Mb!

This is the result of the first command

-rw-r----- 1 syslog adm  7547261 2012-05-19 20:30 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/kern.log
-rw-r----- 1 syslog adm  5643274 2012-05-19 20:30 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/syslog
-rw-r--r-- 1 root   root 1207704 2012-05-19 12:28 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/dpkg.log
-rw-rw-r-- 1 root   utmp  292292 2012-05-19 20:24 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/lastlog
-rw-r--r-- 1 root   root  287371 2012-05-19 20:21 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/pm-powersave.log
-rw-r----- 1 syslog adm    76300 2012-05-19 20:30 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/auth.log
-rw-r--r-- 1 root   root   60167 2012-05-16 14:13 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/alternatives.log
-rw-r--r-- 1 root   root   50269 2012-05-19 20:31 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/Xorg.0.log
-rw-r--r-- 1 root   root   46891 2012-04-23 12:22 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/bootstrap.log
-rw-r--r-- 1 root   root   35604 2012-05-17 17:45 /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f/var/log/pm-suspend.log

Unfortunately, I'm not able, looking at the content of the files, to tell where is the problem.

My disk space is not full; I have several partitions, and none of them is fully used.

ubuntu@ubuntu:~$ df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/sda5             20161172   6349904  12787128  34% /media/3e9e2201-13e5-4c7f-9b91-48a388949b3f
/dev/sda3             51199148  13424832  37774316  27% /media/DATA
/dev/sda6            108317696  50293868  52521568  49% /media/216b81ed-2e5b-4608-abf6-5fc940e3f624
/dev/sda2             51199148  33489988  17709160  66% /media/ACER

ubuntu@ubuntu:~$ sudo fsck /dev/sda5
fsck from util-linux-ng 2.17.2
e2fsck 1.41.11 (14-Mar-2010)
/dev/sda5: clean, 189657/1281120 files, 1667893/5120710 blocks
ubuntu@ubuntu:~$ sudo fsck /dev/sda6
fsck from util-linux-ng 2.17.2
e2fsck 1.41.11 (14-Mar-2010)
/dev/sda6: clean, 38079/6881280 files, 13005347/27511304 blocks

However, I have also a copy of Windows XP in dual-boot alongside Kubuntu, and it runs without any problem; it makes me think (and hope) that the issue is not hardware related.

[GreyGeek]

....
However, I have also a copy of Windows XP in dual-boot alongside Kubuntu, and it runs without any problem; it makes me think (and hope) that the issue is not hardware related.

Well, you've eliminated any hardware problems.

You mentioned you tried from a LiveCD. How did it run (except for being CD slow)?

[sekhemty]

The LiveCD runs as smooth as it can.

Anyway, I think I've managed to find where the problem is located.

Before entering Kubuntu, I tried to rename the .kde directory (in my Home) in .kdebackup, then logged in. A brand new .kde folder was automatically created, and the system was running nicely: no CPU full load, no HD led always on. But, as you can imagine, in this way none of my customization and personal setting was used.

So I think that the problem is in some setting inside that folder, but I don't have a clue on where to look more precisely. In the worst of the hypotheses, I can simply re-set any option and customization in every application, but I hope that there is a way to pinpoint more precisely the source of the issue and solve it.

[james147]

Start with ~/.kde/share/config all settings for kde applications are stored in there. I would start with plasm*rc and kwin*rc (you can rename or move them and restart the associated application or relogin to test).
~/.kde/share/apps contains all the application data (kwallet password and such) and is less likely to cause the issues.
The only other place might be the auto start folder if something is getting auto started that is causing the problems, but I would think that the configs are the most likely place for the problem.

[GreyGeek]

There's a nice little program called "kdiff3" (sudo apt-get install kdiff3) which allows you to compare files, line by line, in ~/.kde with those in ~/.kdebackup.

[sekhemty]

Thanks to both of you. I started with renaming the various folders inside .kde/share and I found that, indeed, the problem is located inside the "config" folder.

I will surely use this tool because the folder contains more than 150 files and checking them one by one will be a monstrous work

[james147]

Thanks to both of you. I started with renaming the various folders inside .kde/share and I found that, indeed, the problem is located inside the "config" folder.

I will surely use this tool because the folder contains more than 150 files and checking them one by one will be a monstrous work

Its likely the plasma or kwin files that are causing the problem, I would start with them.

[andrewsimpson]

Check your hard drive. The errors in the log are almost certainly hardware errors from the hard drive. I have seen these logs before in hard drives about to fail, and I speak from personal experience!

This is why the Live CD is running well for you. I can't answer why Windows running nornally, but it could be that only one, or two, partitions are affected. However these problems tend to spread over more of the disk with time....

[andrewsimpson]

Have a look at this posting on the Ubuntu forums: http://ubuntuforums.org/archive/inde...t-1882996.html

It's an old thread, but the error messages are the same. The second to last post explains what's happening.

Back up now! Investigate later.

[sekhemty]

Thank you for your answer, I'm beginning to worry now.

I did a check with fsck

ubuntu@ubuntu:~$ sudo fsck /dev/sda5
fsck from util-linux-ng 2.17.2
e2fsck 1.41.11 (14-Mar-2010)
/dev/sda5: clean, 189657/1281120 files, 1667893/5120710 blocks
ubuntu@ubuntu:~$ sudo fsck /dev/sda6
fsck from util-linux-ng 2.17.2
e2fsck 1.41.11 (14-Mar-2010)
/dev/sda6: clean, 38079/6881280 files, 13005347/27511304 blocks

and the output seems ok.

Is there any other tool I can use to try to find of the HD has physical errors?

[SecretCode]

There's a GTK+ application called palimpsest (installed via package gnome-disk-utility; appears in the menus as "Disk Utility"). I don't know if there's a KDE version or equivalent.

This will allow you to view "SMART" data and run tests on the drive. On my drive, there are 22 different statistics. You want to look through and see any that show up as red - even then they may not be disastrous, but some certainly are and would indicate a very swift replacement of the disk.