From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kent Overstreet Subject: Re: [bcachefs][tier] INFO: task bch_tier_read:21414 blocked for more than 120 seconds. Date: Wed, 12 Oct 2016 04:11:05 -0800 Message-ID: <20161012121105.ym3sed4tgc2vx3dh@kmo-pixel> References: <20161012012600.s2cq5ac37hwwrp3e@kmo-pixel> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Return-path: Received: from mail-pa0-f52.google.com ([209.85.220.52]:32804 "EHLO mail-pa0-f52.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752053AbcJLMLJ (ORCPT ); Wed, 12 Oct 2016 08:11:09 -0400 Received: by mail-pa0-f52.google.com with SMTP id vu5so27814685pab.0 for ; Wed, 12 Oct 2016 05:11:09 -0700 (PDT) Content-Disposition: inline In-Reply-To: Sender: linux-bcache-owner@vger.kernel.org List-Id: linux-bcache@vger.kernel.org To: Marcin =?utf-8?B?TWlyb3PFgmF3?= Cc: linux-bcache@vger.kernel.org On Wed, Oct 12, 2016 at 01:07:46PM +0200, Marcin Mirosław wrote: > W dniu 12.10.2016 o 03:26, Kent Overstreet pisze: > Hi, > [...] > > You're not the only one who's seeing this, I've been chasing this bug for the > > past several days... > > > > Is something that can be done on my side to help with it? > I can add information that tier0 device is a device with short time of > writing requests - it's a HW controller with RAM. Second bling shot is > that using btree-node=256k was problematic some time ago: > http://www.spinics.net/lists/linux-bcache/msg04178.html , maybe this is > somehow related to the btree-node size? It turns out it's not a bcache bug - FUA is buggy... If you want, you could confirm by changing journal_flushes_device() to always return true - journal.h line 85. Gonna try and figure out what the actual FUA bug is tomorrow...