From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from amazon.4net.rs ([159.69.148.70]:54812 "EHLO amazon.4net.rs" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1730269AbeLOPgP (ORCPT ); Sat, 15 Dec 2018 10:36:15 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit Date: Sat, 15 Dec 2018 16:36:10 +0100 From: =?UTF-8?Q?Sini=C5=A1a_Bandin?= Subject: Re: XFS and RAID10 with o2 layout In-Reply-To: <23572.5681.810087.381496@quad.stoffel.home> References: <20181213220533.GH6311@dastard> <3e0e8def-2bb4-5e8a-56c1-d010ede059d4@4net.rs> <5C13694A.8070303@youngman.org.uk> <23572.5681.810087.381496@quad.stoffel.home> Message-ID: <1a5643d5f68d3dfe1aaa472aac6bd398@4net.rs> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: linux-raid@vger.kernel.org, linux-xfs@vger.kernel.org, linux-raid-owner@vger.kernel.org On 14.12.2018 21:44, John Stoffel wrote: >>>>>> "Wols" == Wols Lists writes: > > Wols> On 14/12/18 07:03, Sinisa wrote: >>>> And another MD flush thread waiting on a MD sync barrier. >>>> >>>> Basically, this looks and smells like a MD sync barrier race >>>> condition, not an XFs problem. >>>> >>>> Cheers, >>>> >>>> Dave. >>> >>> But why don't we see the same issue with other filesystems? > > Wols> Possibly because, iirc, xfs is aware of the underlying raid? > > I don't think it's that in touch with the raid. It can/does query for > settings so it can optimize stuff, but I don't think it actually uses > it other than as a block device. > > Wols> I don't know, but I seem to remember that from earlier > discussions about > Wols> "xfs over mdraid". > > Have you tried doing this with the absolute latest kernel? Well, not "absolute" but relative :) Absolute latest being 4.20-rc6, I have tried with 4.20-rc2 or -rc3 -- Srdačan pozdrav/Best regards, Siniša Bandin