From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stan Hoeppner Subject: Re: raid6 low performance 8x3tb drives in singledegraded mode(=7x3tb) Date: Wed, 19 Sep 2012 11:31:33 -0500 Message-ID: <5059F365.8000400@hardwarefreak.com> References: <504FD153.9090602@sandberg-consult.dk> <20120919162557.03df6af2@notabene.brown> <20120919134030.222e4b04@natsu> Reply-To: stan@hardwarefreak.com Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20120919134030.222e4b04@natsu> Sender: linux-raid-owner@vger.kernel.org To: Roman Mamedov Cc: NeilBrown , Kasper Sandberg , linux-raid@vger.kernel.org List-Id: linux-raid.ids On 9/19/2012 2:40 AM, Roman Mamedov wrote: > "/sys/block/md*/md/stripe_cache_size" is absolutely the first thing one should > look at when facing a problem with md RAID5/6 write performance. No. The first thing one should do is verify that the md/RAID stripe is properly aligned to physical sector boundaries on the drives, and that the filesystem is properly aligned to the md/RAID stripe size. If these are not correct then tweaking stripe_cache_size isn't going to help much, if at all. -- Stan