Suze freezes..

I installed 12.3 recently, having some problems lately with Suse freezing completely for shorts periods of time, longest is around 25 seconds… then 2 or 3 seconds, then it stops freezing… no error message is displayed or any warning at all.

How to determine the cause ? since can’t start system/monitor.

OS version: Linux 3.7.10-1.16-default
KDE: 4.10.5 “release 1”

P4 2.8 Ghz, 2 GB RAM
GPU: Radeon HD 4670 1 GB (AGP)

3d accelerator: unknown
Renderer: Gallium 0.4 on AMD RV730
OpenGL/ES extension 3.0 Mesa 9.0.2
Kernel module radeon

open a konsole and run top and just leave it where you can see it. When freeze happens see what is eating cycles.

On Mon, 29 Jul 2013 17:26:01 +0000, mhunt0 wrote:

> I installed 12.3 recently, having some problems lately with Suse
> freezing completely for shorts periods of time, longest is around 25
> seconds… then 2 or 3 seconds, then it stops freezing… no error message
> is displayed or any warning at all.
>
> How to determine the cause ? since can’t start system/monitor.

Start the monitor before you run into the problem - I’d use htop in a
terminal window myself, sort by CPU % as a start.

When it locks up, see what it’s doing using that window.

Do you see disk activity when the freeze happens? If you do, that could
indicate insufficient physical memory (which means it’s swapping to swap
space). I’ve seen this on one of my systems with less memory than it
really should have. Solid disk activity is a sign that this may be
happening.

Jim


Jim Henderson
openSUSE Forums Administrator
Forum Use Terms & Conditions at http://tinyurl.com/openSUSE-T-C

The system monitor and the ‘top’ command are not useful because is a total ‘stop’ (including mouse pointer), which means both of those tools ‘froze’ before indicating any process taking the CPU to 100%. The only time some indication or message was right after a long ‘stop’ of 25 seconds that the desktop pop up a message near the task bar that reads ‘Bell in Session ’ Root Shell’ '. At that time I was using a ‘Second Life’ viewer, but the problem has happened without using that software.

I’ll check on disk activity… didn’t think of that, but I did check RAM usage, I have 2GB of ram, last time, right before the ‘stop’ , ram use was of only 370 MB since I wasn’t using any ‘heavy’ software such as as graphics or even the web browser.

On 2013-07-29 22:36, mhunt0 wrote:
>
> The system monitor and the ‘top’ command are not useful because is a
> total ‘stop’ (including mouse pointer), which means both of those tools
> ‘froze’ before indicating any process taking the CPU to 100%. The only
> time some indication or message was right after a long ‘stop’ of 25
> seconds that the desktop pop up a message near the task bar that reads
> ‘Bell in Session ’ Root Shell’ '. At that time I was using a ‘Second
> Life’ viewer, but the problem has happened without using that software.

Check for bad sectors in disk. smartctl long test.


Cheers / Saludos,

Carlos E. R.
(from 12.3 x86_64 “Dartmouth” at Telcontar)

On Mon, 29 Jul 2013 20:36:01 +0000, mhunt0 wrote:

> The system monitor and the ‘top’ command are not useful because is a
> total ‘stop’ (including mouse pointer), which means both of those tools
> ‘froze’ before indicating any process taking the CPU to 100%. The only
> time some indication or message was right after a long ‘stop’ of 25
> seconds that the desktop pop up a message near the task bar that reads
> ‘Bell in Session ’ Root Shell’ '. At that time I was using a ‘Second
> Life’ viewer, but the problem has happened without using that software.

Those tools will tell you what was taking CPU up right before the freeze.

Otherwise, look at disk activity as I suggested before.

Jim


Jim Henderson
openSUSE Forums Administrator
Forum Use Terms & Conditions at http://tinyurl.com/openSUSE-T-C

ok, just happened again, this time desktop did show several messages when it got out of the ‘frozen’ state.

Here are some…


Message from syslogd@linux-1ogx at Jul 30 21:39:41 … kernel: 9008.088004] BUG: soft lockup - CPU#0 stuck for 22s! [konqueror:3174] Message from syslogd@linux-1ogx at Jul 30 21:39:41 … kernel: 9008.088015] Process konqueror (pid: 3174, ti=f3c0a000 task=efd292c0 task.ti=efd2a000) Message from syslogd@linux-1ogx at Jul 30 21:39:41 … kernel: 9008.088015] Stack: Message from syslogd@linux-1ogx at Jul 30 21:39:41 … kernel: 9008.088015] Call Trace: Message from syslogd@linux-1ogx at Jul 30 21:39:41 … kernel: 9008.088015] Code: 00 00 8b 15 e0 dd b7 c0 c1 ea 09 83 e2 07 83 fa 03 0f 87 c9 03 00 00 0f b7 53 34 39 c2 77 2d 8b 87 5c 02 00 00 3e 80 00 01 fb 90 74 26 00 8b 44 24 24 e8 1c 07 fe ff 8b 87 5c 02 00 00 83 c4

Message from syslogd@linux-1ogx at Jul 30 21:50:44 … kernel: 9596.088008] BUG: soft lockup - CPU#0 stuck for 23s! [Chrome_HistoryT:3896]

Message from syslogd@linux-1ogx at Jul 30 21:50:44 … kernel: 9596.088008] Code: 8b 15 c0 02 a8 c0 64 a1 10 c0 a7 c0 89 d5 89 44 24 20 c7 44 24 1c 0a 00 00 00 8d 74 26 00 64 c7 05 c0 02 a8 c0 00 00 00 00 fb 90 74 26 00 bf 00 5a 9b c0 eb 07 83 c7 04 d1 ed 74 5b f7 c5 01


What was I doing ? using chromiun (not konqueror).

oh, no disk activity before or after the ‘frozen’ since it’s not just one but several of different lenghts until it stops happening.

On 07/31/2013 06:06 AM, mhunt0 wrote:
> no disk activity before or after the ‘frozen’ since it’s not just
> one but several of different lenghts until it stops happening.

it would really be nice to know what application/utility/process is
causing the “soft lockup”…that should be able to be learned if you
will open top in an x-term and arrange your desktop so that you can
see top even if you can not move your mouse…like here
http://susepaste.org/93533969 i can see the top three most CPU/memory
hungry processes at just a glance…without moving or clicking
anything…and if i felt like i needed more i could adjust my desktop
to see the top 15 processes, or more…

oh, and turn off desktop effects and see if that solves all the
freezes (i think maybe it is a video problem)…

eventually you probably need to log a bug…


dd
http://tinyurl.com/DD-Caveat
http://tinyurl.com/DD-Complaints

On 2013-07-31 06:06, mhunt0 wrote:

> Here are some…

Please use code tags for printouts and commands (the ‘#’ button).
Posting in
Code Tags - A Guide


> Message from syslogd@linux-1ogx at Jul 30 21:50:44 ... kernel:
> 9596.088008] BUG: soft lockup - CPU#0 stuck for 23s!
> [Chrome_HistoryT:3896]

Seeing the word “bug” in there, I would try Bugzilla.

You, you could try a newer kernel version.


Cheers / Saludos,

Carlos E. R.
(from 12.3 x86_64 “Dartmouth” at Telcontar)

Will do the things mentioned when it happens again. Lucky, it’s not so frequent or I already would return to 12.2.

I normally turn off all desktop effects and the desktop search to improve performance.

thank you

What file system are you using?
(Running blkid as root will display each partition and fs type)

I ask this as I’ve seen a very similar issue with btrfs where certain kernel btrfs threads would cause a complete system hang lasting from a few to many seconds. Granted, the soft lockup error message don’t reference btrfs-endio or other threads, but I still wonder.

If you are using btrfs, do you have lots of old snapshots? (snapper list will display them)

Problem still exits… the ‘desktop alerts’ have not returned, having the system monitor on, I can see ‘Xorg’ taking up to 80% of the CPU when it gets frozen, then ‘watchdog/0’ the it’s released, there are occasions where it gets frozen constantly, and there are days when it doesn’t happens at all.

I check the HD, so disk activity or disk errors after checking all partitions.

Not using btrfs… using blkid I have:

/dev/sda1: UUID=“A4C0E395C0E36BCC” TYPE=“ntfs”
/dev/sda5: TYPE=“swap”
/dev/sda6: UUID=“b55aa797-6b8e-497b-aea7-8d7859992f00” TYPE=“ext4”
/dev/sda7: UUID=“61960605-499b-47e7-9ec6-fcc363b22505” TYPE=“ext3”

The ntfs is from an incomplete/failed XP partition whose installation failed and haven’t touched again.


OS version: Linux 3.7.10-1.16-default
KDE: 4.10.5 “release 1”

P4 2.8 Ghz, 2 GB RAM
GPU: Radeon HD 4670 1 GB (AGP)

3d accelerator: unknown
Renderer: Gallium 0.4 on AMD RV730
OpenGL/ES extension 3.0 Mesa 9.0.2
Kernel module radeon

On 08/05/2013 07:06 PM, mhunt0 wrote:
> using blkid I have:
>
> /dev/sda1: UUID=“A4C0E395C0E36BCC” TYPE=“ntfs”
> /dev/sda5: TYPE=“swap”
> /dev/sda6: UUID=“b55aa797-6b8e-497b-aea7-8d7859992f00” TYPE=“ext4”
> /dev/sda7: UUID=“61960605-499b-47e7-9ec6-fcc363b22505” TYPE=“ext3”

interesting, could we also please see the terminal output and
input, as well as the beginning prompt and exit prompt, from


sudo /sbin/fdisk -l

copy/paste the in/output back to this thread using the instructions
here: http://goo.gl/i3wnr


dd
http://tinyurl.com/DD-Complaints

Are you by chance using Dropbox?

Even if not, there is an issues with KDE which can cause Xorg high CPU utilization due to animated icons in the task bar. It plagued me for months, with XOrg showing 80-90% CPU usage. I finally found the bug report and solution at:

https://bugs.kde.org/show_bug.cgi?id=314919

https://forums.opensuse.org/english/get-technical-help-here/applications/480831-xorg-high-cpu-usage-sluggish-performance-fresh-install-4.html

However, this solution does not resolve high Xorg CPU usage for all people, but it may be well worth examining.

Running xrestop as outlined in the links above may help in identifying the offending application(s). In my case is was [FONT=arial]plasma-desktop continually allocating more memory when Dropbox sync was running.
[/FONT]

… I didn’t got an ‘alert’ on this post…

anyway, thank you for the information, I don’t use DropBox. I still have the problem, I have disabled every ‘effect’ on KDE, any indexer or automatic system update and lately, ‘watchdog’ using Yast but it seems something activates it again…

I couldn’t reproduce the situation mentioned on those threads with blinking icons.