From mboxrd@z Thu Jan 1 00:00:00 1970 From: Shaohua Li Subject: Re: MD device tinning documentation Date: Wed, 10 May 2017 09:00:19 -0700 Message-ID: <20170510160019.6v2bb34kmu6rvr44@kernel.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-raid-owner@vger.kernel.org To: Vladimir Fedorkov Cc: linux-raid@vger.kernel.org List-Id: linux-raid.ids On Mon, May 08, 2017 at 01:22:53PM +0300, Vladimir Fedorkov wrote: > Hello there! > > I'm running benchmarks and it looks like that md device is only > utilizing single core which it limiting it's performance on concurrent > (4+ threads) writes with 8 SSD drives in software RAID10 > > Kernel: Linux host 4.10.12-1.el7.elrepo.x86_64 #1 SMP Fri Apr 21 > 09:58:15 EDT 2017 x86_64 x86_64 x86_64 GNU/Linux > > Exact numbers are here: http://astellar.com/images/sysbench_rnd_writes.png > > Could you please let me know if there is multi-threaded support on MD > devices in 4.x kernels and point me to the documentation (preferred) > or source code which describes tunable parts of /dev/mdXX devices. There is no knob for tunning. At most time, raid10 directly sends IO to underlayer disks directly from upperlayer thread, so it's multi-threaded. The problem is we have some locking issues in the barrier side. Coly improved the raid1 scalability with a new barrier implementation (fd76863e37fe RAID1: a new I/O barrier implementation to remove resync window), but raid10 is left behind. Thanks, Shaohua