On Thu, Dec 05, 2002 at 05:45:51AM +0800, John Summerfield wrote:
> On Wed, 4 Dec 2002, Post, Mark K wrote:
>
> > Sergey,
> >
> > Nothing is taking all the CPU.  The system is idling.  That's why it's a
> > bug.  :)
>
> Load average doesn't mean CPU. I don't kn0w just what it _does_ mean,
> but during times of stress I''ve seen high loadaverage AND high wait.
> I/O contributes to loadaverage.

Load Average are the number of processes which

- are ready to run (eather on the cpu or in the wait queue for the
  cpu)
- are in uninterruptble sleep state (happens when the process waits
  for resources in the kernel)

So it is possible to have no process on the cpu, but a load average of
one.

Ihno

>
> There was nothing else active when I ran this test, and as you can see
> there was little CPU work involved.
> [root@orange root]# uptime;time dd if=/dev/hda of=/dev/null bs=1024k
> count=1024;uptime
>   5:42am  up 5 days, 10:14, 19 users,  load average: 0.17, 0.25, 0.15
> 1024+0 records in
> 1024+0 records out
>
> real    1m22.574s
> user    0m0.018s
> sys     0m13.588s
>   5:44am  up 5 days, 10:15, 19 users,  load average: 0.87, 0.46, 0.23
> [root@orange root]#
>
>
>
> --
>
>
> Cheers
> John.
>
> Join the "Linux Support by Small Businesses" list at
> http://mail.computerdatasafe.com.au/mailman/listinfo/lssb

--
Ihno Krumreich            [EMAIL PROTECTED]
SuSE Linux AG             S390 & zSeries
Deutschherrnstr. 15-19    +49-911-74053-439
90429 N|rnberg            http://www.suse.de

Reply via email to