*BSD News Article 85035


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!news.eng.convex.com!newsrelay.netins.net!newsfeed.dacom.co.kr!arclight.uoregon.edu!enews.sgi.com!www.nntp.primenet.com!nntp.primenet.com!cs.utexas.edu!geraldo.cc.utexas.edu!piglet.cc.utexas.edu!not-for-mail
From: dougmc@ccwf.cc.utexas.edu (Doug McLaren)
Newsgroups: comp.unix.bsd.freebsd.misc,muc.lists.freebsd.questions
Subject: FreeBSD 2.1.6: Why are things getting reniced automatically ?
Date: 16 Dec 1996 15:34:13 -0600
Organization: Doug's House of Disco
Lines: 30
Message-ID: <594f8l$a3p@piglet.cc.utexas.edu>
NNTP-Posting-Host: piglet.cc.utexas.edu
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:32694

We're running FreeBSD 2.1.6-RELEASE ...

The kernel has been recompiled to allow more tty's, but that's the
only real change I've made.

In any event, after a process accumulates a certain amount of cpu
time, it's reniced to 4.  The value of `certain amount of cpu time'
appears to be around 20 minutes.

It does not occur if the process's nice value is anything other than
0.  If I renice a process back to 0 (as root) it immediately goes back
to 4.  If I renice it to 1 or -1 or anything but 0, it works as
expected, and stays put.

There's no cronjob doing this, not on /var/cron/tabs/root, nor in
/etc/crontab, at least not that I can find.  A cronjob would be too
slow to explain this behavior anyways.

No daemon would appear to be doing it either, but it is running the
standard (update swapper pagedaemon vmdaemon) kernel daemons, perhaps
they have something to do with it.

Is there any way to stop this?  I imagine that this behavior is
deliberate, but I can't find it documented anywhere ...

(A workaround would be to renice everything to `1', of course ...)

-- 
Doug McLaren, dougmc@comco.com
Unsolicited email of a commercial or advertising nature is not welcomed.