From owner-freebsd-fs@freebsd.org Tue Dec 6 09:04:12 2016 Return-Path: Delivered-To: freebsd-fs@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id DAFE0C69694 for ; Tue, 6 Dec 2016 09:04:12 +0000 (UTC) (envelope-from lexa@lexa.ru) Received: from mx3.lexa.ru (mx3.lexa.ru [198.27.68.102]) by mx1.freebsd.org (Postfix) with ESMTP id BD2521A3F for ; Tue, 6 Dec 2016 09:04:12 +0000 (UTC) (envelope-from lexa@lexa.ru) Received: by mx3.lexa.ru (Postfix, from userid 66) id F228C224A5F; Tue, 6 Dec 2016 04:04:11 -0500 (EST) Received: from [193.124.130.130] (unknown [193.124.130.130]) by home-gw.lexa.ru (Postfix) with ESMTP id 02FF4B0583 for ; Tue, 6 Dec 2016 12:02:49 +0300 (MSK) Subject: Re: 12-CURRENT: ZFS single stream/large files read performance degradation To: "freebsd-fs@freebsd.org" References: <477264f9-9263-6df8-4566-2911e84f2ad8@lexa.ru> <20161206084537.GJ54029@kib.kiev.ua> From: Alex Tutubalin Message-ID: <84782619-1a39-edb9-a04d-c3642ddc869e@lexa.ru> Date: Tue, 6 Dec 2016 12:02:49 +0300 User-Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:45.0) Gecko/20100101 Thunderbird/45.5.1 MIME-Version: 1.0 In-Reply-To: <20161206084537.GJ54029@kib.kiev.ua> Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 06 Dec 2016 09:04:13 -0000 On 12/6/2016 11:45 AM, Konstantin Belousov wrote:O > Are you using memstick image from the download site ? > Note that all HEAD images produced by re@ have full debugging enabled, > among which both kernel-side INVARIANTS and WITNESS, and usermode malloc > debugging, have potentially huge impact on the system performance. Yes, I've used memsticks from download site. Anyway, I see this single-stream read speed degradation on my 12-current compiled without INVARIANTS/WITNESS. Also, it looks like 11-STABLE is affected too, I'm compiling releng/11.0 now to see the difference, to be reported when I'll get results with this version. Alex Tutubalin