From owner-freebsd-questions Mon Oct 28 13:28:15 1996 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id NAA04339 for questions-outgoing; Mon, 28 Oct 1996 13:28:15 -0800 (PST) Received: from dyson.iquest.net ([198.70.144.127]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id NAA04332 for ; Mon, 28 Oct 1996 13:28:11 -0800 (PST) Received: (from root@localhost) by dyson.iquest.net (8.8.2/8.6.9) id QAA00722; Mon, 28 Oct 1996 16:27:49 -0500 (EST) From: John Dyson Message-Id: <199610282127.QAA00722@dyson.iquest.net> Subject: Re: server death when swap space is all gone. To: robh@imdb.com Date: Mon, 28 Oct 1996 16:27:48 -0500 (EST) Cc: freebsd-questions@freebsd.org In-Reply-To: <199610281957.TAA10074> from "Rob Hartill" at Oct 28, 96 07:57:03 pm Reply-To: dyson@freebsd.org X-Mailer: ELM [version 2.4 PL24 ME8] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-questions@freebsd.org X-Loop: FreeBSD.org Precedence: bulk (The "humor" in this message is not directed to the original poster, but perhaps is slightly misguided Linux oriented friend :-)). > > On a related note, a Linux using friend takes pleasure in telling me > that FreeBSD is brian-dead w.r.t memory management because it can't diff > a couple of 20Mb files on a machine with ample memory and swap (combined). > Didn't know that diff used mmap? Linux is brain-dead because it doesn't push unused pages out to swapspace, until memory needs to be freed. :-). You know that it is faster to free up the memory, thereby letting the system utilize the ram, as opposed to tying it up with stuff that should be on swap? :-). Many of the complaints are due to people not understanding this bit of info (or perhaps Linux fervor)... > > Are there any kernel tweaks in the area that could be useful ? > No tweaks, but you do need enough swap space. The swap space calculation is most easily described as the total amount of program space that you have... That is unlikely to change without a serious impact on performance. John