Xref: utzoo comp.sys.att:2955 unix-pc.general:529
Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!ncar!oddjob!uwvax!umn-d-ub!umn-cs!bungia!datapg!sewilco
From: sewilco@datapg.DataPg.MN.ORG (Scot E. Wilcoxon)
Newsgroups: comp.sys.att,unix-pc.general
Subject: Re: 7300 uucp trouble with setgetty!
Message-ID: <722@datapg.DataPg.MN.ORG>
Date: 8 Apr 88 03:05:09 GMT
References: <120@slbtoy.UUCP>
Reply-To: sewilco@datapg.DataPg.MN.ORG (Scot E. Wilcoxon)
Followup-To: comp.sys.att
Organization: Data Progress
Lines: 36

In article <120@slbtoy.UUCP> slb@slbtoy.UUCP (Sanford L. Barr) writes:
>...
>	The uucico process spawns a "getoff.sh" which then calls
>"setgetty".  The setgetty process seems to just hang in mid air.  If I
>...

Setgetty has been hanging several times a week here.  I had to make a
daemon which kills any setgetty which has used "too much" CPU time.

In crontab:
6 * * * * /bin/su root -c "/usr/local/bin/setg.clean" > /dev/null

And what does the work (should use setgetty, but it has no documentation):

# Test for a setgetty with a lot of CPU, kill it, and fix inittab

SETGLINE=`ps -ef | fgrep setgetty`
if [ -n "$SETGLINE" ]
then
	CPTIME=`expr "$SETGLINE" : '.*:..:.. ... \(.*\):.. setgetty'`
	if [ ${CPTIME:-0} -gt 9 ]
	then
		echo $SETGLINE | mail root
		PROCID=`expr "$SETGLINE" : '^....... \(.....\) '`
		kill -9 $PROCID
		sleep 5

		cp /etc/inittab /etc/inittab.new
		sed -e '/^:ph1/s/^:/ /' /etc/inittab >/etc/inittab.new
		mv /etc/inittab.new /etc/inittab
		sleep 15
		telinit q
	fi
fi
-- 
Scot E. Wilcoxon  sewilco@DataPg.MN.ORG    {amdahl|hpda}!bungia!datapg!sewilco
Data Progress	  UNIX consulting    +1 612-825-2607      uunet!datapg!sewilco