> Harald Koenig <koenig@tat.physik.uni-tuebingen.de> writes:
> >
> > we've been using this patch to run "real" background jobs, which shouldn't
> > compete even with "nice 19" user processes, for a long time now (both SMP
> > and non-SMP) with no problems at all.
> >
> Um, right, but nice-19-processes shouldn't compete (much) with nice-0
> processes either. Unfortunately, they do, you get 50% CPU time for both if
> you wait long enough.
I was thinking of nice-19 process from one user (running as "background"
or real batch job using e.g. NQS) which shouldn't need to compete at all with
my "idle sucking" jobs like rc5 (which would suck 50% or real usefull
nice-19 work without this patch...)
> nova:~ 40$ ps axlr
> F UID PID PPID PRI NI SIZE RSS WCHAN STAT TTY TIME COMMAND
> 40 202 7176 7003 20 0 1304 740 0 R p5 0:49 -bash
> 100040 202 7175 7170 20 9 1288 832 0 R N p9 0:29 -bash
>
> Note the identical 'PRI' values...
hmm, I don't know how you started this and what these bash'es do.
note the "NI" value of 9, which is weird too.
using a small C program with just an idle loop gives the following for me:
# ps uaxp 8566 8567
USER PID %CPU %MEM SIZE RSS TTY STAT START TIME COMMAND
root 8566 7.8 0.1 820 56 p3 R N 01:28 0:08 idle
root 8567 92.0 0.1 820 56 p5 R 01:28 1:44 idle
Harald
--
All SCSI disks will from now on ___ _____
be required to send an email notice 0--,| /OOOOOOO\
24 hours prior to complete hardware failure! <_/ / /OOOOOOOOOOO\
\ \/OOOOOOOOOOOOOOO\
\ OOOOOOOOOOOOOOOOO|//
Harald Koenig, \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik // / \\ \
koenig@tat.physik.uni-tuebingen.de ^^^^^ ^^^^^