From mboxrd@z Thu Jan 1 00:00:00 1970 From: Neil Brown Subject: Re: raid0 to raid10 takeover bug Date: Fri, 17 Dec 2010 10:24:57 +1100 Message-ID: <20101217102457.3ec8c8da@notabene.brown> References: Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-raid-owner@vger.kernel.org To: "Wojcik, Krzysztof" Cc: "linux-raid@vger.kernel.org" List-Id: linux-raid.ids On Thu, 16 Dec 2010 14:50:42 +0000 "Wojcik, Krzysztof" wrote: > Hi, > > I would like to report a bug connected with raid0 to raid10 takeover operation. > When we add a spare drive to array after takeover, recovery process has been triggered but it is not finish- it stops at about 90%. > "resync" process is in uninterruptable sleep state (D). Deadlock? > It is reproducible on native and imsm metadata. > Any ideas, suggestions? > I would start by getting a stack trace of all relevant processes, with by echo t > /proc/sysrq-trigger or cat /proc/XXXX/stack Find out where the processes is waiting. An 'D' stands for 'Disk wait' rather than 'Deadlock'. It is a by if an anachronistic name as it is really any uninterruptible wait, which can be for things other than waiting for Disk. NeilBrown