From owner-freebsd-questions Sat Jul 10 16: 4:16 1999 Delivered-To: freebsd-questions@freebsd.org Received: from implode.root.com (root.com [209.102.106.178]) by hub.freebsd.org (Postfix) with ESMTP id 65C4714CC5 for ; Sat, 10 Jul 1999 16:04:12 -0700 (PDT) (envelope-from root@implode.root.com) Received: from implode.root.com (localhost [127.0.0.1]) by implode.root.com (8.8.8/8.8.5) with ESMTP id QAA21179; Sat, 10 Jul 1999 16:03:07 -0700 (PDT) Message-Id: <199907102303.QAA21179@implode.root.com> To: Chris Fedde Cc: junkmale@xtra.co.nz, freebsd-questions@FreeBSD.ORG, mckusick@mckusick.com Subject: Re: Uptime basics!!! In-reply-to: Your message of "Sat, 10 Jul 1999 01:49:26 MDT." <199907100749.BAA98281@fedde.littleton.co.us> From: David Greenman Reply-To: dg@root.com Date: Sat, 10 Jul 1999 16:03:07 -0700 Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG >Digging into my bookshelf I find a definition of the load average >in _The Design and Implementation of the 4.3BSD UNIX Operating System_ >in the section on Process Scheduling. > > ...where /load/ is a sampled average of the length of the run queue > over the previous 1 minute interval of system operation. > >It is also defined in the glossary of the same book. Unfortunately >my copy of the 4.4 book is out on loan so I can't confirm that the >text is there too. The BSD 4.3 book is wrong then. I just looked at the 4.3 sources and it also includes short term waits in the load. > >On 8 Jul 99, at 19:41, David Greenman wrote: > > > >> Uh, no, that is not what the load average means. The load average is > a > >> composite number that includes both runnable processes and processes tha > t > >> are blocked in a short term wait (usually disk I/O). This means that for > >> machines that are doing heavy disk I/O, the load average could be quite > >> high even when the CPU is 95% idle. On wcarchive, for example, the load > >> average typically runs around 40-50 with 50% CPU idle time. This may sou > nd > >> high, but there are 38 disk drives on the machine, so although the drive > s > >> are fairly busy, the I/O is spread out over all of them - keeping the > >> interactive response time low and overall performance quite high. -DG David Greenman Co-founder/Principal Architect, The FreeBSD Project - http://www.freebsd.org Creator of high-performance Internet servers - http://www.terasolutions.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message