From mboxrd@z Thu Jan 1 00:00:00 1970 From: Phil Turmel Subject: Re: Unable to (un)-grow raid6 Date: Tue, 8 Dec 2015 22:12:30 -0500 Message-ID: <56679C1E.1040908@turmel.org> References: <4E26C335-DEDB-489E-B54E-A285273569A1@alkaline-solutions.com> <465BF0F5-B5C7-4C02-8EB5-552B959B8D29@alkaline-solutions.com> <56649C04.9000601@turmel.org> <9F08EE4D-B926-4F8F-9F9D-F9C7234959B4@alkaline-solutions.com> <5664A6E2.4070900@turmel.org> <5664D042.9080401@turmel.org> <10980501-F73B-4A63-AF7D-D7100C9C6B72@alkaline-solutions.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <10980501-F73B-4A63-AF7D-D7100C9C6B72@alkaline-solutions.com> Sender: linux-raid-owner@vger.kernel.org To: David Waite Cc: Linux-RAID List-Id: linux-raid.ids On 12/08/2015 06:42 PM, David Waite wrote: > Thanks for your help - it is now reshaping the array with sdia3 and > sdib3 as hot spares. > > It appears the version of mdadm on the NAS (3.1.4) is pretty old, and > does not have a replace command - just marking disks as faulty. I > will build and load in a newer version of mdadm. I assume there is no > issue in doing this while a reshaping is going on? No issue. You don't even have to install it. However, like most features, there's a kernel part and a userspace part. If the new mdadm that supports --replace can't talk to the matching kernel part, it won't work. Either way, you can't actually do the --replace until the reshape finishes. {Convention on kernel.org lists is to reply-to-all, trim replies, and bottom post. Please do.} Phil