POV-Ray : Newsgroups : povray.off-topic : Highlander Process Server Time
4 Sep 2024 17:19:58 EDT (-0400)
  Highlander Process (Message 1 to 8 of 8)  
From: clipka
Subject: Highlander Process
Date: 16 Nov 2009 20:46:44
Message: <4b020084$1@news.povray.org>
Now there is something wrong here I guess...

-------------------------------------------------------------------------
top - 02:34:39 up 30 min,  2 users,  load average: 1.10, 2.06, 1.70
Tasks: 137 total,   2 running, 128 sleeping,   0 stopped,   7 zombie
Cpu(s):  0.9%us, 22.4%sy,  0.5%ni, 75.1%id,  1.1%wa,  0.0%hi,  0.0%si, 
0.0%st
Mem:   5993696k total,   996736k used,  4996960k free,    29212k buffers
Swap:  7847712k total,        0k used,  7847712k free,   850360k cached

   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
  8727 root      20   0     0    0    0 R  100  0.0  15:38.18 Xorg
  5823 root      20   0  8528 1072  764 R    2  0.0   0:00.01 top
  8751 root      20   0  120m 1344  900 S    2  0.0   0:00.57 rsyslogd
     1 root      20   0  8072  756  624 S    0  0.0   0:00.93 init
     2 root      15  -5     0    0    0 S    0  0.0   0:00.00 kthreadd
     3 root      RT  -5     0    0    0 S    0  0.0   0:00.02 migration/0
     4 root      15  -5     0    0    0 S    0  0.0   0:00.00 ksoftirqd/0
     5 root      RT  -5     0    0    0 S    0  0.0   0:00.00 watchdog/0
     6 root      RT  -5     0    0    0 S    0  0.0   0:00.04 migration/1
     7 root      15  -5     0    0    0 S    0  0.0   0:00.00 ksoftirqd/1
     8 root      RT  -5     0    0    0 S    0  0.0   0:00.00 watchdog/1
     9 root      RT  -5     0    0    0 S    0  0.0   0:00.03 migration/2
    10 root      15  -5     0    0    0 S    0  0.0   0:00.00 ksoftirqd/2
    11 root      RT  -5     0    0    0 S    0  0.0   0:00.00 watchdog/2
    12 root      RT  -5     0    0    0 S    0  0.0   0:00.02 migration/3
    13 root      15  -5     0    0    0 S    0  0.0   0:00.08 ksoftirqd/3
    14 root      RT  -5     0    0    0 S    0  0.0   0:00.00 watchdog/3
-------------------------------------------------------------------------

It's worrying enough that Xorg is eating up 100% cpu power (fortunately 
only a single of four cores).

What's even more worrying, however, is the fact that this was AFTER 
issuing a tirade of "kill -SIGsome 8727" commands from root console... 
AND a few "kill -9 8727" as well... WHAT THE...?!?

Is this some kind of Highlander process, that's immune to any fatal 
wounds, and wants all the power for itself??


Looks like I'm doing something seriously wrong in my attempts to install 
openSUSE. First try went good, but then I tried to resize and move 
partitions around, apparently screwing things up. Now even complete 
re-installs won't seem to screw things back down again, for /some/ 
obscure reason...


Post a reply to this message

From: Nicolas Alvarez
Subject: Re: Highlander Process
Date: 16 Nov 2009 21:00:25
Message: <4b0203b9$1@news.povray.org>
clipka wrote:
> Is this some kind of Highlander process, that's immune to any fatal
> wounds, and wants all the power for itself??

If something can't be killed with -9, you can be sure something in the 
*kernel side* is causing problems.

A hung FUSE filesystem sometimes caused me such unkillable processes :/


Post a reply to this message

From: Darren New
Subject: Re: Highlander Process
Date: 16 Nov 2009 21:34:17
Message: <4b020ba9$1@news.povray.org>
clipka wrote:
> Is this some kind of Highlander process, that's immune to any fatal 
> wounds, and wants all the power for itself??

That usually means it's in a kernel critical section. Altho I've never seen 
it happen with a process that's at 100%. Usually it's blocked on a "fast" 
kernel operation (i.e., anything you can't get back EINTR from) that just 
never completed properly. In this case, I'd suspect it's busy-looping 
probably somewhere in a driver.

In any case, it's clearly a kernel problem.  Even when it's technically not 
a kernel bug, it's a kernel misdesign wherein something that should be 
interruptable (like a "fast open") winds up being slow because someone stuck 
a network in the middle or something.

-- 
   Darren New, San Diego CA, USA (PST)
   I ordered stamps from Zazzle that read "Place Stamp Here".


Post a reply to this message

From: clipka
Subject: Re: Highlander Process
Date: 16 Nov 2009 21:51:01
Message: <4b020f95$1@news.povray.org>
Nicolas Alvarez schrieb:
> clipka wrote:
>> Is this some kind of Highlander process, that's immune to any fatal
>> wounds, and wants all the power for itself??
> 
> If something can't be killed with -9, you can be sure something in the 
> *kernel side* is causing problems.
> 
> A hung FUSE filesystem sometimes caused me such unkillable processes :/

How can I tell whether a filesystem is running as FUSE?

Some other observations that may hint in a similar direction:

- The lockup occurs when I log in as a normal user, but when I log in as 
root everything is shiny. Now the /home directories are actually on a 
different partition. (fsck -f does not report anything suspicious for 
that partition though.)

- Aside from Xorg eating a full core's power, I typically also see about 
50% of another core's time being eaten by a process named 
"hald-addon-stor", which I think can't be right either.


Post a reply to this message

From: clipka
Subject: Re: Highlander Process
Date: 16 Nov 2009 22:12:24
Message: <4b021498$1@news.povray.org>
Nicolas Alvarez schrieb:
> clipka wrote:
>> Is this some kind of Highlander process, that's immune to any fatal
>> wounds, and wants all the power for itself??
> 
> If something can't be killed with -9, you can be sure something in the 
> *kernel side* is causing problems.
> 
> A hung FUSE filesystem sometimes caused me such unkillable processes :/

Well, whatever the underlying reason was, it is gone, hopefully for 
good; most likely one of the following fixed it:

- Changed screen resolution from 800x600 @ 16bit to 1280x1024 @ 24bit. 
In my first install attempt I had set the resolution to 1280x1024 right 
from the start, but in the subsequent install attempts I was lazy about it.

- Moved all the contents of the user's home directory into a 
subdirectory thereof. As mentioned already, /home resides on a separate 
partition, and wasn't rebuilt from scratch during the subsequent install 
attempts.

I tend to suspect the latter, actually.


Post a reply to this message

From: nemesis
Subject: Re: Highlander Process
Date: 16 Nov 2009 23:00:00
Message: <web.4b021fa6e6e3ddf1a3d6eddb0@news.povray.org>
clipka <ano### [at] anonymousorg> wrote:
> Looks like I'm doing something seriously wrong in my attempts to install
> openSUSE.

why do you even try?  You know Linux doesn't like you a bit. ;)

hey, I wish I had that much memory! :D


Post a reply to this message

From: Nicolas Alvarez
Subject: Re: Highlander Process
Date: 17 Nov 2009 01:38:06
Message: <4b0244ce$1@news.povray.org>
clipka wrote:
> Nicolas Alvarez schrieb:
>> clipka wrote:
>>> Is this some kind of Highlander process, that's immune to any fatal
>>> wounds, and wants all the power for itself??
>> 
>> If something can't be killed with -9, you can be sure something in the
>> *kernel side* is causing problems.
>> 
>> A hung FUSE filesystem sometimes caused me such unkillable processes :/
> 
> How can I tell whether a filesystem is running as FUSE?

Probably you would have started it yourself.

A FUSE filesystem runs in a user-space program. It's often used for network-
based filesystems, like "ftpfs".


Post a reply to this message

From: clipka
Subject: Re: Highlander Process
Date: 17 Nov 2009 03:21:11
Message: <4b025cf7$1@news.povray.org>
Nicolas Alvarez schrieb:

>> How can I tell whether a filesystem is running as FUSE?
> 
> Probably you would have started it yourself.
> 
> A FUSE filesystem runs in a user-space program. It's often used for network-
> based filesystems, like "ftpfs".

Then... probably no. It's a fresh installation, and I didn't knowingly 
start anything exotic. Unless using ext3 instead of ext4 would be 
considered exotic.


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.