We still have a mix of timer and non-timer kernels and I don't see anything
like this on any of them.

Are you sure this is 2.4.17?  As far as I know, "officially", SLES7 is
locked into variations of 2.4.7.  The latest kernel upgrades from SuSe are:

k_timer-2.4.7-9
k_deflt-2.4.7-44

both dated 5/21/02.

> -----Original Message-----
> From: John Campbell [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, August 13, 2002 8:53 AM
> To: [EMAIL PROTECTED]
> Subject: Re: [LINUX-390] ksoftirqd_CPU0 takes 100%
>
>
> >>>>>>>>
>
>
>   # top
>
>
>   2:00pm  up  5:15,  2 users,  load average: 1.00, 1.00, 1.00
> 41 processes: 35 sleeping, 3 running, 3 zombie, 0 stopped
> CPU states:  0.3% user, 99.4% system,  0.0% nice,  0.1% idle
> Mem:   501980K av,  494408K used,    7572K free,       0K
> shrd,  136120K
> buff
> Swap: 1464376K av,       0K used, 1464376K free
>     56796K
> cached
>
>   PID USER     PRI  NI  SIZE   RSS SHARE STAT %CPU %MEM   TIME
> COMMAND
>     4        root       20   19      0         0           0
>         RWN
> 98.6        0.0       300:11       ksoftirqd_CPU0
>
>
> # uptime
>   2:05pm  up  5:20,  2 users,  load average: 1.03, 1.03, 1.00
>
>
> What does it mean?
>
> btw: SLES7 64bit on LPAR with 2.4.17-timer kernel.
>
> WBR, Sergey
> <<<<<<<<
>
>       Could this be due to the lack of the 100Hz timer pop?
>
>       With a truly "idle" Linux instance you're measuring utilization
>       from *inside* a rubberized room-  this is kind of a "software"
>       example of the Lorentz-Fitzgerald contraction-  your ruler is
>       changing it's length as the virtuality takes effect.  Even if,
>       in this instance, the ruler is a clock  :-)
>
>       AlanA, how far off am I?
>
> --------------------
> John R. Campbell, Speaker to Machines (GNUrd)      {813-356|697}-5322
> Adsumo ergo raptus sum
> IBM Certified: IBM AIX 4.3 System Administration, System Support
>

Reply via email to