From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vojtech Pavlik Subject: Re: Bcache stuck at writeback of a key, consuming 100% CPU, not possible to detach Date: Mon, 31 Aug 2015 17:54:15 +0200 Message-ID: <20150831155415.GA7394@suse.com> References: <20150830085442.GA31722@suse.com> <20150831163937.00ca3f7a@harpe.intellique.com> <20150831144949.GA3276@suse.com> <20150831170928.56fa9bbd@harpe.intellique.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from mx2.suse.de ([195.135.220.15]:34150 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753125AbbHaPyR (ORCPT ); Mon, 31 Aug 2015 11:54:17 -0400 Content-Disposition: inline In-Reply-To: <20150831170928.56fa9bbd@harpe.intellique.com> Sender: linux-bcache-owner@vger.kernel.org List-Id: linux-bcache@vger.kernel.org To: Emmanuel Florac Cc: kmo@daterainc.com, linux-bcache@vger.kernel.org On Mon, Aug 31, 2015 at 05:09:28PM +0200, Emmanuel Florac wrote: > Le Mon, 31 Aug 2015 16:49:49 +0200 > Vojtech Pavlik =C3=A9crivait: >=20 > > >=20 > > > The bcache_writeback stays at 100% _even_ when in writethrough mo= de, > > > alas. So this looks normal. However dirty_data definitely should > > > drop to zero... =20 > >=20 > > This most certainly isn't normal. The ftrace shows it's looping in = a > > loop doing nothing useful. > >=20 >=20 > Yep, I've had a quick look at the code, however it's not very > inspiring. It very much looks like there's a missing test. From your > trace it looks like it finds a key_bad (?) -- from the struct > bch_extent_keys_ops in bcache/extents.c -- but there is no code anywh= ere > to manage this case apparently. >=20 > Ah, and I can't download your complete trace, too : twilight.ucw.cz > isn't responding. :) Sadly, the bcache array crashed hard earlier today (with a bug in and = the machine doesn't boot anymore. I'm recovering it now. --=20 Vojtech Pavlik Director SuSE Labs