Re: weird scheduling

[Folkert van Heusden]

James Manning wrote:
> [ Wednesday, February 16, 2000 ] Folkert van Heusden wrote:
> >   PID USER     PRI  NI  SIZE  RSS SHARE STAT  LIB %CPU %MEM   TIME
> >   199 root      20  19 13588  13M   552 R N     0 42.1 10.6   2:50
> >   204 root      20  19   700  700   568 R N     0 32.3  0.5   3:18
> >   118 root      20  19 14100  13M   552 R N     0 24.7 11.0   5:20
> Can you report as to whether the percentages are still the same much
> later? (at least 60 minutes).

They are. That is; most of the time. After a while, precentages
sometimes
are much closer, but they're still jumping between 24% and 38%.
Did I already mention that I user 2.2.14?

> It appears that the processes that have
> used more CPU time overall are getting letting CPU time now, and if
> that's really the case all 3 should settle into a scenario where their
> CPU time used (2:50, 3:18, 5:20 in your snapshot) pretty much "sync up".
> Whether it's desired or not is an orthogonal issue, I just want to know
> if the behavior occurred :)

Nope:

 10:07pm  up 23:48,  0 users,  load average: 3.07, 3.02, 2.93
47 processes: 43 sleeping, 4 running, 0 zombie, 0 stopped
CPU states:  1.7% user,  1.8% system, 96.3% nice,  0.0% idle
Mem:  128076K av, 105068K used,  23008K free,  22704K shrd,  15436K buff
Swap: 144576K av,   5792K used, 138784K free                 29300K
cached
 2052 root      20  19   868  868   688 R N     0 45.1  0.6 239:40
gf_client
 2049 root      20  19 13588  13M   552 R N     0 37.3 10.6 211:59
setiathome
 2045 root      20  19 14100  13M   552 R N     0 16.5 11.0 224:40
setiathome

-- 
------------------------------------------------------------
Folkert van Heusden
http://www.vanheusden.com/
some e-mail addresses: [e-mail], [e-mail]
mobile phone: [phone]