From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christian Pernegger Subject: Re: unbelievably bad performance: 2.6.27.37 and raid6 Date: Sat, 31 Oct 2009 20:59:39 +0100 Message-ID: References: Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: Sender: linux-raid-owner@vger.kernel.org To: Jon Nelson Cc: LinuxRaid List-Id: linux-raid.ids > md0 : active raid6 sda4[0] sdc4[5] sdd4[4] sdb4[6] > 613409536 blocks super 1.1 level 6, 64k chunk, algorithm 2 [4/4]= [UUUU] Why would you use a 4 disk raid6? If 50% of raw capacity is enough just go with raid10. The default 64KiB chunk size has always been very slow for me, try something larger, perhaps even 1MiB. > While rsyncing a file from an ext3 filesystem to a jfs filesystem, rsync isn't really a good diagnostic, how's performance with something simple, dd, bonnie++, whatever? > I looked at stripe_cache_size and increased it to 384 Much too low, I usually set it to 8192. Take care you don't run out of RAM, though, it's in pages/disk (16KiB in your case). > blockdev --getra reports 256 for all involved raid components. In my experience only the top component of a layered block device matters at all, that would be the LV. 256 sectors / 128 KiB seems awfully low, try something higher, 2MiB at least. > Am I crazy? =C2=A0Is 12.5MB/s (average) what I should expect, here? No, it's probably just that the default tunining options are not very g= ood. Cheers, C. -- To unsubscribe from this list: send the line "unsubscribe linux-raid" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html