Linux RAID subsystem development
 help / color / mirror / Atom feed
From: Andreas Klauer <Andreas.Klauer@metamorpher.de>
To: "Patrik Dahlström" <risca@powerlamerz.org>
Cc: Brad Campbell <lists2009@fnarfbargle.com>, linux-raid@vger.kernel.org
Subject: Re: Recover array after I panicked
Date: Sun, 23 Apr 2017 17:42:40 +0200	[thread overview]
Message-ID: <20170423154240.GA3494@metamorpher.de> (raw)
In-Reply-To: <f353b7fc-a63a-5503-5cda-6a62698f69ae@powerlamerz.org>

On Sun, Apr 23, 2017 at 05:11:08PM +0200, Patrik Dahlström wrote:
> * I try to find 0x53EF (ext4 magic) within the first 128M of
> /dev/sd[abcde]. Not after? This will be an indication of my "New
> Offset". I need to adjust the offset a bit since the ext4 magic is
> located at 0x438 offset.

Yes, that should be about it.
 
> * I create a 5 and a 6 drive raid set and try to find an offset where
> they both carry the same raw data. With some overlays, I should be able
> to create both these raids at the same time, correct?

Yes, two sets of overlays.

So overlay A is your 5 disk raid5, overlay B is your 6 disk raid5, 
and then you'll just have to take a stab at it with hexdump.

So kind of like,

hexdump --skip $((17*1024*1024*1024)) --length 4096 /dev/md42
hexdump --skip $((17*1024*1024*1024)) --length 4096 /dev/md43

If that produces the same random-looking data then your reshape 
might have progressed 17G-ish and you could try to use that as 
a starting point for a linear device mapping that uses the 
first 17G of the 6 disk raid and everything else from the 5disk.

The further the grow processed the larger a zone of overlap there 
should be since more data ends up on the additional drive so 
the original representation of the same data isn't yet written 
into on the old drives.

Does that make sense?

5 disks: a b c d e   : f g h i j   : k l m n o   : p q r s t : ...
6 disks: a b c d e f : g h i j k l : m n o p q r : s t       : ...

If it stopped at "t", both have "r s t"... and that "r s t" 
would be what you have to find. You have to be wary of false 
matches though (such as zeroes or other common patterns of data 
you might find anywhere).

And there's one more thing, you mentioned a disk had a bad sata cable.
If that disk got kicked and reshape went on for a while afterwards, 
you should take that disk out of consideration. (Specify as "missing" 
when creating the arrays.) It will have outdated/unreshaped data on 
it that would be hard to incorporate into your recovery attempt...

Regards
Andreas Klauer

  parent reply	other threads:[~2017-04-23 15:42 UTC|newest]

Thread overview: 63+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-04-23  9:47 Recover array after I panicked Patrik Dahlström
2017-04-23 10:16 ` Andreas Klauer
2017-04-23 10:23   ` Patrik Dahlström
2017-04-23 10:46     ` Andreas Klauer
2017-04-23 11:12       ` Patrik Dahlström
2017-04-23 11:36         ` Wols Lists
2017-04-23 11:47           ` Patrik Dahlström
2017-04-23 11:53             ` Reindl Harald
2017-04-23 11:58           ` Roman Mamedov
2017-04-23 12:11             ` Wols Lists
2017-04-23 12:15               ` Patrik Dahlström
2017-04-24 21:04                 ` Phil Turmel
2017-04-24 21:56                   ` Patrik Dahlström
2017-04-24 23:35                     ` Phil Turmel
2017-04-23 13:16         ` Andreas Klauer
2017-04-23 13:49           ` Patrik Dahlström
2017-04-23 14:36             ` Andreas Klauer
2017-04-23 14:45               ` Patrik Dahlström
2017-04-23 12:32     ` Patrik Dahlström
2017-04-23 12:45       ` Andreas Klauer
2017-04-23 12:57         ` Patrik Dahlström
2017-04-23 14:06 ` Brad Campbell
2017-04-23 14:09   ` Patrik Dahlström
2017-04-23 14:20     ` Patrik Dahlström
2017-04-23 14:25     ` Brad Campbell
2017-04-23 14:48   ` Andreas Klauer
2017-04-23 15:11     ` Patrik Dahlström
2017-04-23 15:24       ` Patrik Dahlström
2017-04-23 15:42       ` Andreas Klauer [this message]
2017-04-23 16:29         ` Patrik Dahlström
2017-04-23 19:21         ` Patrik Dahlström
2017-04-24  2:09           ` Brad Campbell
2017-04-24  7:34             ` Patrik Dahlström
2017-04-24 11:04               ` Andreas Klauer
2017-04-24 12:13                 ` Patrik Dahlström
2017-04-24 12:37                   ` Andreas Klauer
2017-04-24 12:54                     ` Patrik Dahlström
2017-04-24 13:39                       ` Andreas Klauer
2017-04-24 14:05                         ` Patrik Dahlström
2017-04-24 14:21                           ` Andreas Klauer
2017-04-24 16:00                           ` Patrik Dahlström
2017-04-24 23:00                         ` Patrik Dahlström
2017-04-25  0:16                           ` Andreas Klauer
2017-04-25  8:44                             ` Patrik Dahlström
2017-04-25  9:01                               ` Andreas Klauer
2017-04-25 10:40                                 ` Patrik Dahlström
2017-04-25 10:51                                   ` Patrik Dahlström
2017-04-25 11:08                                   ` Andreas Klauer
2017-04-25 11:37                                     ` Patrik Dahlström
2017-04-25 12:41                                       ` Andreas Klauer
2017-04-25 18:22                                       ` Wols Lists
2017-04-27 19:57                                     ` Patrik Dahlström
2017-04-27 23:12                                       ` Andreas Klauer
2017-04-28  7:11                                         ` Patrik Dahlström
2017-04-28  9:52                                           ` Andreas Klauer
2017-04-28 10:31                                             ` Patrik Dahlström
2017-04-28 11:39                                               ` Andreas Klauer
2017-04-28 22:46                                         ` Patrik Dahlström
2017-04-29  9:56                                           ` Andreas Klauer
2017-05-02 13:08                                             ` Patrik Dahlström
2017-05-02 13:11                                               ` Brad Campbell
2017-05-02 15:49                                               ` Anthony Youngman
2017-04-25 23:01                 ` Patrik Dahlström

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=20170423154240.GA3494@metamorpher.de \
    --to=andreas.klauer@metamorpher.de \
    --cc=linux-raid@vger.kernel.org \
    --cc=lists2009@fnarfbargle.com \
    --cc=risca@powerlamerz.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