From mboxrd@z Thu Jan 1 00:00:00 1970 From: Phil Turmel Subject: Re: mdadm stuck at 0% reshape after grow Date: Tue, 5 Dec 2017 21:51:27 -0500 Message-ID: References: <1865221512489329@web5g.yandex.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1865221512489329@web5g.yandex.ru> Content-Language: en-GB Sender: linux-raid-owner@vger.kernel.org To: 002@tut.by, Jeremy Graham , "linux-raid@vger.kernel.org" List-Id: linux-raid.ids On 12/05/2017 10:55 AM, 002@tut.by wrote: > A well known reason for this behavior are bad blocks in device's BBL, > which you happen to have: [trim /] > This feature generally shouldn't be used, because its implementation > is unfinished. Empty BBL's can be removed from every device by giving > "--update=no-bbl" option to mdadm on assemble, but before that you > must manually regenerate content for each block in BBL's and then > manually zero the lists in superblocks. I endorse this opinion. The BBL should never have been merged as-is. Without a block reallocation system like a real hard drive, a BBL entry simple drops redundancy on the block in question without any other corrective action. Completely, utterly, brain-dead. At the very least, it should not be enabled by default on new arrays. Phil