From owner-freebsd-geom@FreeBSD.ORG Mon Jan 6 22:07:27 2014 Return-Path: Delivered-To: freebsd-geom@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id D75B619B for ; Mon, 6 Jan 2014 22:07:27 +0000 (UTC) Received: from elsa.codelab.cz (elsa.codelab.cz [94.124.105.4]) (using TLSv1 with cipher ADH-CAMELLIA256-SHA (256/256 bits)) (No client certificate requested) by mx1.freebsd.org (Postfix) with ESMTPS id 93C9C1F01 for ; Mon, 6 Jan 2014 22:07:27 +0000 (UTC) Received: from elsa.codelab.cz (localhost [127.0.0.1]) by elsa.codelab.cz (Postfix) with ESMTP id 9C8D628434; Mon, 6 Jan 2014 23:07:19 +0100 (CET) Received: from [192.168.1.2] (ip-89-177-49-222.net.upcbroadband.cz [89.177.49.222]) (using TLSv1 with cipher DHE-RSA-CAMELLIA256-SHA (256/256 bits)) (No client certificate requested) by elsa.codelab.cz (Postfix) with ESMTPSA id C5E0E2842D; Mon, 6 Jan 2014 23:07:18 +0100 (CET) Message-ID: <52CB2915.7060602@quip.cz> Date: Mon, 06 Jan 2014 23:07:17 +0100 From: Miroslav Lachman <000.fbsd@quip.cz> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9.1.19) Gecko/20110420 Lightning/1.0b1 SeaMonkey/2.0.14 MIME-Version: 1.0 To: Karl Pielorz Subject: Re: Is TRIM working with gmirror? References: <52A90049.7050001@quip.cz> <18C59233568D3A475FF0D843@study64.tdx.co.uk> In-Reply-To: <18C59233568D3A475FF0D843@study64.tdx.co.uk> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-geom@FreeBSD.org X-BeenThere: freebsd-geom@freebsd.org X-Mailman-Version: 2.1.17 Precedence: list List-Id: GEOM-specific discussions and implementations List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 06 Jan 2014 22:07:27 -0000 Karl Pielorz wrote: [...] >> WARNING: /ssd_db: TRIM flag on fs but disk does not confirm that it >> supports TRIM >> >> The filesystem is classic UFS2. >> >> Is TRIM supported throught gmirror, or not? > > I think 8.4 is way too old for this. I have a number of 9.2 boxes - and > even on those, TRIM doesn't propagate through gmirror *yet* (at the > time, which is a few months ago I seem to remember reading 'It should > work soon' - but 9.2 doesn't appear to). > > You can use gstat to show if any calls to BIO_DELETE are being made - i.e. > > gstat -d > > On the systems here the 'd/s' (which afaik is 'BIO_DELETE/second') > remains stubbornly at zero :( Thank you for pointing me to gstat -d. d/s is really zero all the time. > The flipside is with decent SSD's these days TRIM isn't quite so > important... Some may even argue that *not* issuing heaps of BIO_DELETE > on the I/O channel is actually faster - so long as the SSD has time to > play catchup and isn't pegged all the time, I think they may be right :-) The SSDs in question are holding /var/db/mysql with heavily used InnoDB tables (grow only ibdata file, few deletes of MyISAM tables) so maybe there are not much delete commands at all. Thank you. Miroslav Lachman