From owner-freebsd-current@FreeBSD.ORG Sun Mar 21 01:24:07 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 12CE216A536; Sun, 21 Mar 2004 01:24:07 -0800 (PST) Received: from sccrmhc11.comcast.net (sccrmhc11.comcast.net [204.127.202.55]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8BE3843D3F; Sun, 21 Mar 2004 01:24:04 -0800 (PST) (envelope-from julian@elischer.org) Received: from interjet.elischer.org ([24.7.73.28]) by comcast.net (sccrmhc11) with ESMTP id <2004032109235901100icb3he>; Sun, 21 Mar 2004 09:24:03 +0000 Received: from localhost (localhost.elischer.org [127.0.0.1]) by InterJet.elischer.org (8.9.1a/8.9.1) with ESMTP id BAA18946; Sun, 21 Mar 2004 01:23:58 -0800 (PST) Date: Sun, 21 Mar 2004 01:23:56 -0800 (PST) From: Julian Elischer To: Scott Long In-Reply-To: <405D4F09.5050804@freebsd.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: Craig Rodrigues cc: Andy Hilker cc: current@freebsd.org Subject: Re: libpthread / mysql41-server X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 21 Mar 2004 09:24:07 -0000 On Sun, 21 Mar 2004, Scott Long wrote: > Daniel Eischen wrote: > > > > > > Mysql uses scope system threads which are limited to > > kern.threads.max_groups_per_proc. System scope threads > > seem to have unfair scheduling also. I use SCHED_4BSD > > and this patch to mysql40-server to make it use process > > scope threads and it works much better: > > > > http://people.freebsd.org/~deischen/mysql40-server.diffs > > > > I haven't made a patch to other versions of mysql, but I > > have sent the maintainer the patch. > > > > Having MySQL work 'out of the box' is a critical item for the success > of FreeBSD 5.3 and beyond. Should your patch be incorporated into the > port? What are the tradeoffs between system and process scope threads > in this situation? Since issues with system scope threads seem to come > up alot, what will it take to make them no longer be a problem? system scope threads use more resources They are artificially limmited in the kernel to 50 per process. I guess we should increase that to 500 or something. heck I just committed that change.... ULE also doesn't treat KSE threaded processes well, and there is a patch floating around to fix that but no-one is willing to commit it to ULE and Jeff has not commented on it. it is reported to make a BIG difference to KSE programs.