linux-raid.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: NeilBrown <neilb@suse.de>
To: "Arkadiusz Miśkiewicz" <a.miskiewicz@gmail.com>
Cc: linux-raid@vger.kernel.org
Subject: Re: Bad block management - which mdadm?
Date: Tue, 25 Oct 2011 09:46:16 +1100	[thread overview]
Message-ID: <20111025094616.78384241@notabene.brown> (raw)
In-Reply-To: <201110242212.14934.a.miskiewicz@gmail.com>

[-- Attachment #1: Type: text/plain, Size: 1820 bytes --]

On Mon, 24 Oct 2011 22:12:14 +0200 Arkadiusz Miśkiewicz
<a.miskiewicz@gmail.com> wrote:

> 
> kernelnewbies.org reports such feature for 3.1 kernel
> 
> 1.9. Software RAID: Bad block management
> The MD layer (aka "Multiple Devices", aka software raid) has gained bad block 
> management support: bad blocks will be added to a list, and the system will 
> try not to use them. This feature requires an updated mdadm version.
> 
> Which mdadm supports that?
> 

mdadm-3.3 will support bad block management.
However it is not released yet.

You can get current code from

   git://neil.brown.name/mdadm  devel-3.3


It isn't released yet because

  1/ I haven't found/made time to work on mdadm and integrate support
     properly.
  2/ The kernel isn't really quite read for full support.

If you have a bad block list, then a write error will not fail the drive but
will record the location of the write error.
This might be what you want.  But if you have a hot-spare available it might
not be what you want - the current kernel code will never use a hot spare
until a drive fails really badly.

What it *should* do is "hot-replace".  i.e. recover the data on to a spare
without first removing the device with the bad block.  But the hot-replace
code isn't ready yet.

I expect hot-replace to be in linux-3.3 (for RAID4/5/6 at least) and I hope to
release mdadm-3.3 before then with full support for bad blocks and hot
replace.

For now if you use the devel-3.3 code to create a new array it will place a
bad-block-log on the array and the kernel will use that bad block log to
record failed blocks rather than failing the whole device.
If you want a device to be failed, you might have to explicitly do that
yourself with "mdadm /dev/mdX --fail /dev/sdY"

NeilBrown


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 828 bytes --]

      reply	other threads:[~2011-10-24 22:46 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-10-24 20:12 Bad block management - which mdadm? Arkadiusz Miśkiewicz
2011-10-24 22:46 ` NeilBrown [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20111025094616.78384241@notabene.brown \
    --to=neilb@suse.de \
    --cc=a.miskiewicz@gmail.com \
    --cc=linux-raid@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).