Arch Linux

kjozsa

Member

Registered: 2008-06-05

Posts: 147

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

I experience the same (load constantly over 1.0) running 4.1.2-2-ARCH on an up-to-date arch install (as of 2015-07-25) and running Xorg with vanilla i3wm. I have no NFS mounts and see no processes in htop with any load over 5% (including system processes).

Where should I look for the cause of this load?

ulke

Member

Registered: 2011-04-29

Posts: 50

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

Exactly the same issue here as kjozsa. The LTS kernel (3.14.48-1) is not affected on my machine.

I see no other problems except the high reported load.

tsh

Member

From: Munich

Registered: 2014-07-25

Posts: 41

Website

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

I recently upgraded linux from 4.0.4-2 to 4.1.3-1 and now I see that the load average is always > 1.  It wasn't like this before.

Like others, I do not have any CPU hungry programs running.  The processor fan does not seem to run.

I'm using btrfs, vfat, and no NFS mounts.

Last edited by tsh (2015-07-30 20:51:44)

kjozsa

Member

Registered: 2008-06-05

Posts: 147

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

I'm not at my box at the moment, but did anyone tried the ck kernel to see if it behaves differently?

ewaller

Administrator

From: Pasadena, CA

Registered: 2009-07-13

Posts: 19,804

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

I have a sneaky feeling that this is related to my thread here.

Things were okay with 3.14.  After that, the ACPI subsystem was rewickered to start ACPI services earlier in the boot process.  I think this broke a few things.  There are about three relevant files that were touched in the acpi module.  This weekend, I intend to build a kernel 4.1 kernel with those changes backed out.  I am doing this with a custom kernel with modules built in.  If  I have success, I'll rebuild the Arch kernel with the reversions to the acpi files and will post the module on line somewhere and will post the hash here.  Anyone who trusts me will be welcome to try the substitute modules.   I am busy the rest of today, so look for something maybe Sunday.

---

Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

dRbiG

Member

Registered: 2015-08-04

Posts: 1

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

Can confirm. ThinkPad x220, Sandy Bridge i7 - 4.0.7-2 works fine, 4.1.x minimum loadavg of 1.00. Seems yet again ThinkPad/Sandy Bridge triggers some weirdness :/

kjozsa

Member

Registered: 2008-06-05

Posts: 147

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

I also have a Sandy Bridge CPU (i7-2630QM) but in a HP laptop - so it might be something Sandy Bridge related..

ewaller

Administrator

From: Pasadena, CA

Registered: 2009-07-13

Posts: 19,804

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

This weekend, I grabbed 4.2.0-rc4 from kernel.org, built, and installed it.  It solves a multitude of sins related to ACPI with out an additional patches

edit

I also have a Sandy Bridge CPU (i7-2630QM) but in a HP laptop - so it might be something Sandy Bridge related..

I've an i7-4720HQ processor in an HP Envy 15 touch.  What is your HP model?

Last edited by ewaller (2015-08-04 13:58:11)

---

Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

kuminamoya

Member

Registered: 2013-10-29

Posts: 3

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

fyi: Seems like I am having the same problem (load permanently >1 for no obvious reason). Lenovo Thinkpad T440s with an i7-4600U CPU (Haswell, no sandy bridge). Just noticed it today, I can´t tell since when this occurs.

kjozsa

Member

Registered: 2008-06-05

Posts: 147

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

My HP is an Elitebook 8560w.

eparis

Member

From: NC, USA

Registered: 2015-08-20

Posts: 9

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

This issue appears fixed for me on linux 4.1.5-1. System load is back to 0 when nobody is using the system, as opposed to the constant 1.0 system load.

Last edited by eparis (2015-08-20 01:13:16)

kuminamoya

Member

Registered: 2013-10-29

Posts: 3

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

Yes, fixed in 4.1.5-1 for me as well.

It wasn´t just a cosmetic thing, the fan was running constantly, now everything is back to normal.

tsh

Member

From: Munich

Registered: 2014-07-25

Posts: 41

Website

Re: (Very) high loadavg - probably bug like in kernel 3.3.5?

I too confirm that it is fixed in 4.1.5-1 for me.