From: Andi Kleen Subject: Re: [PATCH 0/4] (RESEND) ext3[34] barrier changes Date: Sun, 18 May 2008 21:54:48 +0200 Message-ID: <87ej7zcrqv.fsf@basil.nowhere.org> References: <482DDA56.6000301@redhat.com> <20080516130545.845a3be9.akpm@linux-foundation.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Eric Sandeen , linux-ext4@vger.kernel.org, linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, jamie@shareable.org To: Andrew Morton Return-path: In-Reply-To: <20080516130545.845a3be9.akpm@linux-foundation.org> (Andrew Morton's message of "Fri, 16 May 2008 13:05:45 -0700") Sender: linux-fsdevel-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org Andrew Morton writes: > On Fri, 16 May 2008 14:02:46 -0500 > Eric Sandeen wrote: > >> A collection of patches to make ext3 & 4 use barriers by >> default, and to call blkdev_issue_flush on fsync if they >> are enabled. > > Last time this came up lots of workloads slowed down by 30% so I > dropped the patches in horror. Didn't ext4 have some new checksum trick to avoid them? > I just don't think we can quietly go and slow everyone's machines down > by this much. The overhead of journalling is already pretty horrid. > > If we were seeing a significant number of "hey, my disk got wrecked" There used to be a couple of reports of people doing some write workload use automatic power switch to reset turn on again repeat for 24 hours getting consistent corruption after some time <24hours and those went away with barriers. > Do we know which distros are enabling barriers by default? SUSE did for some releases, although reiserfs used to be the default there (but that one uses barriers too if the disk supports them) BTW for a lot of distros which default to LVM install it won't make any difference because DM right now doesn't pass barriers through. I have a patch queued for that for the one-disk case at least, but it didn't make .26-rc. -Andi