[gentoo-user] 100% CPU usage with no processes to blame?

[gentoo-user] 100% CPU usage with no processes to blame?

[Grant]

I just sat through about 5 minutes of 100% CPU usage and a thrashing
hard disk.  I ran top and it reported 0% idle CPU, but the list of
processes totaled maybe 20% CPU usage.  How can this be?

- Grant
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Alan McKinnon]

On Wednesday 14 February 2007, Grant wrote:
> I just sat through about 5 minutes of 100% CPU usage and a thrashing
> hard disk.  I ran top and it reported 0% idle CPU, but the list of
> processes totaled maybe 20% CPU usage.  How can this be?
>
> - Grant

Most likely you only looked at user cpu % and neglected to list the 
system and niced times as well.

p.s. golden rule: ps lies. top lies. free lies. 
Don't believe the readings they give, rather interpret them in context.

alan

-- 
Optimists say the glass is half full,
Pessimists say the glass is half empty,
Developers say wtf is the glass twice as big as it needs to be?

Alan McKinnon
alan at linuxholdings dot co dot za
+27 82, double three seven, one nine three five
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[brullo nulla]

> Most likely you only looked at user cpu % and neglected to list the
> system and niced times as well.
>
> p.s. golden rule: ps lies. top lies. free lies.
> Don't believe the readings they give, rather interpret them in context.

sob. it's not the first time I hear this. What should I believe to
really know my system state?

m.
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Uwe Thiem]

On 14 February 2007 18:55, Alan McKinnon wrote:
> On Wednesday 14 February 2007, Grant wrote:
> > I just sat through about 5 minutes of 100% CPU usage and a thrashing
> > hard disk.  I ran top and it reported 0% idle CPU, but the list of
> > processes totaled maybe 20% CPU usage.  How can this be?
> >
> > - Grant
>
> Most likely you only looked at user cpu % and neglected to list the
> system and niced times as well.
>
> p.s. golden rule: ps lies. top lies. free lies.
> Don't believe the readings they give, rather interpret them in context.

!!! Golden rule of system administration: All these tools are notoriously 
wrong.

Uwe

-- 
A fast and easy generator of fractals for KDE:
http://www.SysEx.com.na/iwy-1.0.tar.bz2
Proof of concept of a TSP solver for KDE:
http://www.SysEx.com.na/epat-0.1.tar.bz2
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Hemmann, Volker Armin]

On Mittwoch, 14. Februar 2007, brullo nulla wrote:
> > Most likely you only looked at user cpu % and neglected to list the
> > system and niced times as well.
> >
> > p.s. golden rule: ps lies. top lies. free lies.
> > Don't believe the readings they give, rather interpret them in context.
>
> sob. it's not the first time I hear this. What should I believe to
> really know my system state?
>

youc ould try htop. Maybe it lies a little bit less than top.
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Jerry McBride]

On Wednesday 14 February 2007 12:25:27 pm brullo nulla wrote:
> > Most likely you only looked at user cpu % and neglected to list the
> > system and niced times as well.
> >
> > p.s. golden rule: ps lies. top lies. free lies.
> > Don't believe the readings they give, rather interpret them in context.
>
> sob. it's not the first time I hear this. What should I believe to
> really know my system state?
>
> m.

I tracked my 100% cpu usage to FAMD... Killing it instantly freed the cpu...


-- 

--

Jerry McBride
-- 
gentoo-user@gentoo.org mailing list

RE: [gentoo-user] 100% CPU usage with no processes to blame?

[Timothy A. Holmes]

 


-----Original Message-----
From: Hemmann, Volker Armin
[mailto:volker.armin.hemmann@tu-clausthal.de] 
Sent: Wednesday, February 14, 2007 1:14 PM
To: gentoo-user@lists.gentoo.org
Subject: Re: [gentoo-user] 100% CPU usage with no processes to blame?

On Mittwoch, 14. Februar 2007, brullo nulla wrote:
> > Most likely you only looked at user cpu % and neglected to list the 
> > system and niced times as well.
> >
> > p.s. golden rule: ps lies. top lies. free lies.
> > Don't believe the readings they give, rather interpret them in
context.
>
> sob. it's not the first time I hear this. What should I believe to 
> really know my system state?
>

youc ould try htop. Maybe it lies a little bit less than top.
--
gentoo-user@gentoo.org mailing list

If I cannot trust the tools provided by the OS for finding out whats
going on-- than what can I trust -- this is intolerable! - if it is
simply a situation if a non root user not being able to see root
processes that is one thing, BUT if it really is a situation of the
tools not giving accurate asessments, that is a whole other situation,
and one that just cannot be allowed to continue (not that I can do much
other than scream about it, cuz im not a programmer)



Tim Holmes
IT Manager / Webmaster / Teacher

Medina Christian Academy
A Higher Standard... 

--
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Hans-Werner Hilse]

Hi,

On Wed, 14 Feb 2007 13:31:31 -0500
"Timothy A. Holmes" <tholmes@mcaschool.net> wrote:

> If I cannot trust the tools provided by the OS for finding out whats
> going on-- than what can I trust -- this is intolerable! - if it is
> simply a situation if a non root user not being able to see root
> processes that is one thing, BUT if it really is a situation of the
> tools not giving accurate asessments, that is a whole other situation,
> and one that just cannot be allowed to continue (not that I can do much
> other than scream about it, cuz im not a programmer)

It's mainly not a matter of trust, it's a matter of definitions not
meeting expectations. From an objective POV, the measures shown are
"correct". They are just defined as a more or less complex computation
which isn't exactly what is expected subjectively. People tend to
ignore the definitions and imagine the meaning from the shortened name
of the shown value.

Regarding trusting the OS: Those tools don't come with the kernel. This
is Linux, not the BSDs.

-hwh
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Pongrácz István]

2007. 02. 14, szerda keltezéssel 13.28-kor Jerry McBride ezt írta:
> I tracked my 100% cpu usage to FAMD... Killing it instantly freed the
cpu...

Change to gamin.
The same function in much better.
I did it about a year ago. (Or less, I don't remember).
It solved me some other problems (CD lock down etc.)

In the gentoo documents you can find article about how to change it.

Regards,
István


-- 
Nyílt forráskód azokra az igazán nehéz napokra. BSA
Open source for those really hard days. BSA
http://www.osbusiness.hu

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Jerry McBride]

On Wednesday 14 February 2007 02:48:18 pm Pongrácz István wrote:
> 2007. 02. 14, szerda keltezéssel 13.28-kor Jerry McBride ezt írta:
> > I tracked my 100% cpu usage to FAMD... Killing it instantly freed the
>
> cpu...
>
> Change to gamin.
> The same function in much better.
> I did it about a year ago. (Or less, I don't remember).
> It solved me some other problems (CD lock down etc.)
>
> In the gentoo documents you can find article about how to change it.
>

Thank you, for the tip. I never would have found it, without your help.

Cheers. 

--

Jerry McBride
--
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Hemmann, Volker Armin]

On Mittwoch, 14. Februar 2007, Jerry McBride wrote:
> On Wednesday 14 February 2007 02:48:18 pm Pongrácz István wrote:
> > 2007. 02. 14, szerda keltezéssel 13.28-kor Jerry McBride ezt írta:
> > > I tracked my 100% cpu usage to FAMD... Killing it instantly freed the
> >
> > cpu...
> >
> > Change to gamin.
> > The same function in much better.
> > I did it about a year ago. (Or less, I don't remember).
> > It solved me some other problems (CD lock down etc.)
> >
> > In the gentoo documents you can find article about how to change it.
>
> Thank you, for the tip. I never would have found it, without your help.
>

or just go without it, because both (fam, gamin) are both superfluos
--
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Alan McKinnon]

On Wednesday 14 February 2007, brullo nulla wrote:
> > Most likely you only looked at user cpu % and neglected to list the
> > system and niced times as well.
> >
> > p.s. golden rule: ps lies. top lies. free lies.
> > Don't believe the readings they give, rather interpret them in
> > context.
>
> sob. it's not the first time I hear this. What should I believe to
> really know my system state?

You should use ps, top and free of course! Just realize that they lie...

Actually they don't lie, they give an average. You also need to 
understand what is going on. Without getting into the details of 
exactly what a cpu is doing, it's performing actions millions of times 
a second. In the time it takes top to update (2 seconds), the kernel 
could have started and shut down several thousand forked processes. So 
trying to measure something that changes so fast is a waste of time.

Also, there's the kernels internal timer. It's set by default to 100Hz, 
which means that the kernel updates it's own counters 100 times a 
second. It's quite possible to use and discard lots of memory in that 
gap and for it to never even show up in the kernel's counters. Raising 
the timer frequency for 1000Hz for a desktop gives more accurate 
results, but they are still just an average.

Same with memory - the kernel allocates it out and shares the same 
memory between several processes as it sees fit. And does what it wants 
to with swap as well. Again, trying to find out the current state of 
the system is useless as by the time you read it, it has changed 16 
million times....

When we say that top lies, we really mean that it is giving you an 
average over a looooooong period of cpu time, and you should interpret 
it as such. It's a trend, not an instant value.

alan



-- 
Optimists say the glass is half full,
Pessimists say the glass is half empty,
Developers say wtf is the glass twice as big as it needs to be?

Alan McKinnon
alan at linuxholdings dot co dot za
+27 82, double three seven, one nine three five
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Alan McKinnon]

On Wednesday 14 February 2007, Timothy A. Holmes wrote:
> If I cannot trust the tools provided by the OS for finding out whats
> going on-- than what can I trust -- this is intolerable! - if it is
> simply a situation if a non root user not being able to see root
> processes that is one thing, BUT if it really is a situation of the
> tools not giving accurate asessments, that is a whole other
> situation, and one that just cannot be allowed to continue (not that
> I can do much other than scream about it, cuz im not a programmer)

Calm down, the tools are not outright lying, see my other reply to 
brullo for details. The trend the tools display is accurate though.

Incidentally, you do realize that every OS written in the past 20 years 
suffers from this exact same problem, right? Linux simply doesn't try 
and hide the fact that it is "lying"

-- 
Optimists say the glass is half full,
Pessimists say the glass is half empty,
Developers say wtf is the glass twice as big as it needs to be?

Alan McKinnon
alan at linuxholdings dot co dot za
+27 82, double three seven, one nine three five
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Michael Crute]

On 2/15/07, Alan McKinnon <alan@linuxholdings.co.za> wrote:
> On Wednesday 14 February 2007, brullo nulla wrote:
> > > Most likely you only looked at user cpu % and neglected to list the
> > > system and niced times as well.
> > >
> > > p.s. golden rule: ps lies. top lies. free lies.
> > > Don't believe the readings they give, rather interpret them in
> > > context.
> >
> > sob. it's not the first time I hear this. What should I believe to
> > really know my system state?
>
> You should use ps, top and free of course! Just realize that they lie...

Seeing this thread reminded me of a blog article I saw on Virtual
Threads a while back...
http://virtualthreads.blogspot.com/2006/02/understanding-memory-usage-on-linux.html

He does a pretty good job of explaining where top gets its numbers and
how to properly interpret them. HTH

-mike

-- 
________________________________
Michael E. Crute
http://mike.crute.org

God put me on this earth to accomplish a certain number of things.
Right now I am so far behind that I will never die. --Bill Watterson
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[b.n.]

Michael Crute ha scritto:
>> You should use ps, top and free of course! Just realize that they lie...
> 
> Seeing this thread reminded me of a blog article I saw on Virtual
> Threads a while back...
> http://virtualthreads.blogspot.com/2006/02/understanding-memory-usage-on-linux.html
> 
> 
> He does a pretty good job of explaining where top gets its numbers and
> how to properly interpret them. HTH

Nice post. Thanks for the link.

m.
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Mick]

[-- Attachment #1: Type: text/plain, Size: 706 bytes --]

On Thursday 15 February 2007 22:02, b.n. wrote:
> Michael Crute ha scritto:
> >> You should use ps, top and free of course! Just realize that they lie...
> >
> > Seeing this thread reminded me of a blog article I saw on Virtual
> > Threads a while back...
> > http://virtualthreads.blogspot.com/2006/02/understanding-memory-usage-on-
> >linux.html
> >
> >
> > He does a pretty good job of explaining where top gets its numbers and
> > how to properly interpret them. HTH
>
> Nice post. Thanks for the link.

Indeed, I hadn't used the pmap command until now!  I was shocked to discover 
the direct memory used by Firefox-bin vs Opera vs Konqueror (just for a 
laugh).
-- 
Regards,
Mick

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Grant]

> When we say that top lies, we really mean that it is giving you an
> average over a looooooong period of cpu time, and you should interpret
> it as such. It's a trend, not an instant value.

Shouldn't top have provided some kind of info for why the CPU usage
was 100% for 5 minutes straight?  If it does display trends, shouldn't
it have picked up on that one?

- Grant
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Alex Schuster]

Grant writes:

> Shouldn't top have provided some kind of info for why the CPU usage
> was 100% for 5 minutes straight?  If it does display trends, shouldn't
> it have picked up on that one?

I bet it was updatedb. This is what top shows me when it is running:

Cpu(s): 15.6% us, 12.6% sy,  0.3% ni,  0.0% id, 70.2% wa, 1.3% hi,  0.0% si
  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 1563 root      18   0  1936  928  528 R  8.9  0.2   0:09.68 updatedb

Whatever this "wa" entry is, probably something with I/O related waiting,
it seems is it not being taken into account for the %CPU culumn entry of the 
process.

I think whis was different with kernel 2.4, it did not show such a high 
load, while performing similar as now when updatedb was running.

	Alex
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] 100% CPU usage with no processes to blame?

[Alan McKinnon]

On Saturday 17 February 2007, Alex Schuster wrote:
> > Shouldn't top have provided some kind of info for why the CPU usage
> > was 100% for 5 minutes straight?  If it does display trends,
> > shouldn't it have picked up on that one?
>
> I bet it was updatedb. This is what top shows me when it is running:

or eupdatedb, or prelink... :-)

> Cpu(s): 15.6% us, 12.6% sy,  0.3% ni,  0.0% id, 70.2% wa, 1.3% hi,
>  0.0% si PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+
>  COMMAND 1563 root      18   0  1936  928  528 R  8.9  0.2   0:09.68
> updatedb
>
> Whatever this "wa" entry is, probably something with I/O related
> waiting, it seems is it not being taken into account for the %CPU
> culumn entry of the process.

"wa" is "wait" - a process is trying to do IO and it is being blocked as 
something else is using a resource the process wants to use. SO it sits 
and does nothing, much the same way you sometimes sit in your car and 
do nothing at the McDonalds drive-through queue.

Usually a blocked process will consume no cpu time (as it's doing 
nothing), but it can slow the machine down and make it less responsive 
while many processes sit and wait

alan

-- 
Optimists say the glass is half full,
Pessimists say the glass is half empty,
Developers say wtf is the glass twice as big as it needs to be?

Alan McKinnon
alan at linuxholdings dot co dot za
+27 82, double three seven, one nine three five
--
gentoo-user@gentoo.org mailing list