Ted Unangst wrote:
On 8/11/08, Daniel Ouellet <[EMAIL PROTECTED]> wrote:
Any idea on how it might be possible to boot the system step by step to get
an idea of where this bug might be isolated?
The real bug is looking at load average and pretending it means anything.
Not saying it's a huge deal, but looking at:
/sys/kern/sched_bsd.c
The update priority and some process do use it:
One example may be from the above file:
fixpt_t loadfac = loadfactor(averunnable.ldavg[0]);
and some process are decay based on it as well in schedcpu() and
calculations made from it.
I sure don't pretend to understand the implications of not having the
proper informations to deal with, but if the scheduler based some
decisions from that, doesn't it mean it may have some impact never the less?
That's where I was/am looking anyway to see how this might be cause.
Or may I not understanding this the right way?
If it wasn't important what so ever, then the load average wouldn't be
use in scheduler decisions? I sure agree that I may be 100% off track in
understanding how it's really done here. I am trying to understand it.
Meaning anything or not, still is wrong however isn't it?
I got more interested in it when I try to use the cronjob based on the
load average and in some cases, never run because the load average was
never lower then 1, witch granted, I can adjust for sure, but then it's
wrong.
So, even if it doesn't mean anything then, any suggestions as to how I
could dig it up more to actually fix it anyway if possible?
Best,
Daniel
