NetBSD Problem Report #43629
From tron@zhadum.org.uk Sat Jul 17 10:38:36 2010
Return-Path: <tron@zhadum.org.uk>
Received: from mail.netbsd.org (mail.netbsd.org [204.152.190.11])
by www.NetBSD.org (Postfix) with ESMTP id BECFD63BAE7
for <gnats-bugs@gnats.NetBSD.org>; Sat, 17 Jul 2010 10:38:35 +0000 (UTC)
Message-Id: <20100717103833.76BD341F5D@mail.zhadum.org.uk>
Date: Sat, 17 Jul 2010 11:38:33 +0100 (BST)
From: tron@zhadum.org.uk
Reply-To: tron@zhadum.org.uk
To: gnats-bugs@gnats.NetBSD.org
Subject: RAIDframe should be able to ignore read errors while replacing a component
X-Send-Pr-Version: 3.95
>Number: 43629
>Category: kern
>Synopsis: RAIDframe should be able to ignore read errors while replacing a component
>Confidential: no
>Severity: non-critical
>Priority: low
>Responsible: oster
>State: open
>Class: change-request
>Submitter-Id: net
>Arrival-Date: Sat Jul 17 10:40:01 +0000 2010
>Last-Modified: Sat Jul 24 17:25:01 +0000 2021
>Originator: Matthias Scheler
>Release: NetBSD 5.1_RC3
>Organization:
Matthias Scheler http://zhadum.org.uk/
>Environment:
System: NetBSD colwyn.zhadum.org.uk 5.1_RC3 NetBSD 5.1_RC3 (COLWYN.64) #1: Thu Jul 1 23:21:22 BST 2010 tron@colwyn.zhadum.org.uk:/src/sys/compile/COLWYN.64 amd64
Architecture: x86_64
Machine: amd64
>Description:
Both disks in the primary RAIDframe RAID 1 volume of my server developed
bad blocks this week (in a single nighty :-( ). RAIDframe dropped one
disk out of the RAID 1 and I'm not stuck with the remaining disk that
has bad blocks as well.
I've just added a replacement disk to my server and it would be really
nice if I could use something like "raidctl -R -f /dev/wd0a raid0" to
force RAIDframe to add that disk to the RAID 1. This would allow me to
migrate the data to the new disk seamlessly. I don't care about the
blocks on the old disk which cannot be read as they are lost anyway.
>How-To-Repeat:
Try to add a new disk to RAID 1 volume with only one disk left which has
bad sectors.
>Fix:
None provided.
>Release-Note:
>Audit-Trail:
Responsible-Changed-From-To: kern-bug-people->oster
Responsible-Changed-By: oster@NetBSD.org
Responsible-Changed-When: Sat, 24 Jul 2021 16:21:19 +0000
Responsible-Changed-Why:
Take.
From: Greg Oster <oster@netbsd.org>
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: kern/43629 (RAIDframe should be able to ignore read errors while
replacing a component)
Date: Sat, 24 Jul 2021 10:25:43 -0600
I'm not sure if this is what you meant with:
raidctl -R -f /dev/wd0a raid0
but I agree that raidctl should grow a flag to basically say "ignore any
read errors on reconstruction and reconstruct as best you can".
(I had a RAID 1 set have both disks develop errors a few years back...
ended up building a custom kernel to ignore the read errors in order to
get a rebuild done (and then replace the failing disk, of course). In
my case the bad blocks were not in any files/directory structures, so
nothing was lost, but it would still be handy to have a 'yes, really do
this' flag...)
Later...
Greg Oster
From: Matthias Scheler <tron@zhadum.org.uk>
To: gnats-bugs@netbsd.org
Cc: oster@netbsd.org, gnats-admin@netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/43629 (RAIDframe should be able to ignore read errors while
replacing a component)
Date: Sat, 24 Jul 2021 18:23:25 +0100
On Sat, Jul 24, 2021 at 04:30:02PM +0000, Greg Oster wrote:
> I'm not sure if this is what you meant with:
>
> raidctl -R -f /dev/wd0a raid0
I cannot remember either, it has been a while. I suppose "-f" is
a force flag here.
> but I agree that raidctl should grow a flag to basically say "ignore any
> read errors on reconstruction and reconstruct as best you can".
Yes, that is exactly the use case that I was thinking of.
Kind regards
--
Matthias Scheler http://zhadum.org.uk/
>Unformatted:
(Contact us)
$NetBSD: query-full-pr,v 1.46 2020/01/03 16:35:01 leot Exp $
$NetBSD: gnats_config.sh,v 1.9 2014/08/02 14:16:04 spz Exp $
Copyright © 1994-2020
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.