From owner-cvs-src@FreeBSD.ORG Thu Mar 9 17:01:29 2006 Return-Path: X-Original-To: cvs-src@freebsd.org Delivered-To: cvs-src@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 4131816A420; Thu, 9 Mar 2006 17:01:29 +0000 (GMT) (envelope-from marck@rinet.ru) Received: from woozle.rinet.ru (woozle.rinet.ru [195.54.192.68]) by mx1.FreeBSD.org (Postfix) with ESMTP id 72E1C43D4C; Thu, 9 Mar 2006 17:01:27 +0000 (GMT) (envelope-from marck@rinet.ru) Received: from localhost (localhost [127.0.0.1]) by woozle.rinet.ru (8.13.4/8.13.4) with ESMTP id k29H1PYZ077005; Thu, 9 Mar 2006 20:01:25 +0300 (MSK) (envelope-from marck@rinet.ru) Date: Thu, 9 Mar 2006 20:01:25 +0300 (MSK) From: Dmitry Morozovsky To: Ruslan Ermilov In-Reply-To: <20060309091410.GO54826@ip.net.ua> Message-ID: <20060309195433.O69649@woozle.rinet.ru> References: <200603080827.k288RXLb076843@repoman.freebsd.org> <20060308125450.I68519@woozle.rinet.ru> <20060309091300.GN54826@ip.net.ua> <20060309091410.GO54826@ip.net.ua> X-NCC-RegID: ru.rinet MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-2.0.2 (woozle.rinet.ru [0.0.0.0]); Thu, 09 Mar 2006 20:01:25 +0300 (MSK) Cc: cvs-src@freebsd.org, src-committers@freebsd.org, Pawel Jakub Dawidek , cvs-all@freebsd.org Subject: Re: cvs commit: src/sys/geom/mirror g_mirror.c X-BeenThere: cvs-src@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: CVS commit messages for the src tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 09 Mar 2006 17:01:29 -0000 On Thu, 9 Mar 2006, Ruslan Ermilov wrote: RE> > YMMV of course, but at work we prefer to mirror slices not disks. RE> ^^^^^^^^^^^^^ RE> > This way it's much more flexible -- you can have mirrored root, RE> > unmirrored swap, stripped+mirrored data, etc. RE> > RE> Err, I meant "mirror partitions" of course. :-) So do I. Hence, in standard mirror situation I have /dev/mirror/m0a from ad4a+ad6a for root, /dev/mirror/m0b from ad4b+ad6b for swap etc. Troubles was with crashdumps, so earlier I'd reserve ad4d without file system to dump to. Now it should be a bit simplier. To pjd@: idea from wild space: if one have to replace faulted disk he could have serious performance problems with syncing more than one gmirror/graid3 at a time. Maybe introduce a mechanism to suspend one or more sync process(es) when other are in progress? Or even max concurrency sync level? Sincerely, D.Marck [DM5020, MCK-RIPE, DM3-RIPN] ------------------------------------------------------------------------ *** Dmitry Morozovsky --- D.Marck --- Wild Woozle --- marck@rinet.ru *** ------------------------------------------------------------------------