Date: Fri, 31 May 1996 08:51:46 -0700 From: Jason Fesler <jfesler@calweb.com> To: "Miguel A.L. Paraz" <map@iphil.net> Cc: freebsd-isp@freebsd.org Subject: Re: detecting and killing CPU hogs Message-ID: <2.2.32.19960531155146.0071577c@jpop.calweb.com>
next in thread | raw e-mail | index | archive | help
At 03:27 PM 5/31/96 +0800, you wrote: >Hi! > >Is there a way to do kill programs like this via scripts: > >> 1989 p0- RN 1688:47.17 lynx >> 7488 p0- RN 1477:42.69 lynx >> 15296 p0- RN 5359:54.88 lynx >> 17227 p0- RN 2380:00.85 lynx >> 19982 p1- RN 429:47.18 pine >> 21068 p1- RN 419:02.95 pine I nearly did so, but found too many problems with killing the wrong ones :-). People get upset about that ;-) Everyone who works here, though, has the current network and load status built into their prompt (ala tcsh's magic). Any time something goes down, or the load gets high, they know about it. I *do* kill off "guest" entries where the CPU is high - guesst will never load pine, tin, etc [which spike the CPU usage for their processes pretty easily]. >I think these are caused by people who run these from the shell via >a telnet session, which gets disconnected. Am I right? Bingo. -- Jason Fesler jfesler@calweb.com Admin, CalWeb Internet Services jfroot@calweb.com I like my Usenet over ice, please. http://www.gigo.com Disclaimer: My /dev/null can beat your /dev/null any day.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?2.2.32.19960531155146.0071577c>