From: Roy Sigurd Karlsbakk <roy@karlsbakk.net>
To: Jorge Nunes <jorgebnunes@gmail.com>
Cc: Linux Raid <linux-raid@vger.kernel.org>
Subject: Re: RAID 1 to RAID 5 failure
Date: Tue, 5 Apr 2022 11:17:10 +0200 (CEST) [thread overview]
Message-ID: <336816279.3605676.1649150230084.JavaMail.zimbra@karlsbakk.net> (raw)
In-Reply-To: <987997234.3307867.1649118543055.JavaMail.zimbra@karlsbakk.net>
I re-did these tests this morning, since I was unsure if I could have made some mistake last night - I was tired. There results were about the same - complete data loss.
As for curiousity, I also tried to skip the expand phase after creating the initial raid5 on top of the raid1. After creating it, I stopped it and recreated the old raid1 with --assume-clean. This worked well - no errors from mount or fsck.
So I guess it was the mdadm --grow --raid-devices=4 that was the final nail in the coffin.
I just hope you find a way to backup your files next time. I'm quite sure we've all been there - thought we were smart enough or something and the shit hit the fan and no - we weren't.
Vennlig hilsen
roy
--
Roy Sigurd Karlsbakk
(+47) 98013356
http://blogg.karlsbakk.net/
GPG Public key: http://karlsbakk.net/roysigurdkarlsbakk.pubkey.txt
--
Hið góða skaltu í stein höggva, hið illa í snjó rita.
----- Original Message -----
> From: "Roy Sigurd Karlsbakk" <roy@karlsbakk.net>
> To: "Jorge Nunes" <jorgebnunes@gmail.com>
> Cc: "Linux Raid" <linux-raid@vger.kernel.org>
> Sent: Tuesday, 5 April, 2022 02:29:03
> Subject: Re: RAID 1 to RAID 5 failure
>> Didn't do a backup :-(
>
> First mistake… *Always* keep a backup (or three)
>
>>
>> Unmount everything:
>
> No need - what you should have done, was just to grow the array by
>
> Partition the new drives exactly like the old ones
> mdadm --add /dev/md0 /dev/sd[cd]1 # note that sd[cd] means sdc and sdd, but can
> be written this way on the commandline
> mdadm --grow --level=5 --raid-devices=4
>
> This would have grown and converted the array to raid5 without any data loss.
>
>> $ sudo mdadm --create /dev/md0 -a yes -l 5 -n 2 /dev/sda /dev/sdd
>
> As earlier mentioned, this is to create a new array, not a conversion.
>
>> So, my question is: Is there a chance to redo the array correctly
>> without losing the information inside? Is it possible to recover the
>> 'lost' partition that existed on RAID 1 to be able to do a convenient
>> backup? Or the only chance is to have a correct disk alignment inside
>> the array to be able to use photorec to recover the files correctly?
>
> As mentioned, it doesn't look promising, but there are a few things that can be
> tried.
>
> Your data may still reside on the sda1 and sdd1, but since it was converted to
> RAID-5, the data would have been distributed among the two drives and not being
> the same on both. Further growing the raid, would move the data around to the
> other disks. I did a small test here on some vdisks to see if this could be
> reversed somehow and see if I could find the original filesystem. I could - but
> it was terribly corrupted, so not a single file remained.
>
> If this was valuable data, there might be a way to rescue them, but I fear a lot
> is overwritten already. Others in here (or other places) may know more about
> how to fix this, though. If you find out how, please tell. It'd be interesting
> to learn :)
>
> PS: I have my personal notebook for technical stuff at
> https://wiki.karlsbakk.net/index.php/Roy's_notes in case you might find that
> interesting. There's quite a bit about storage there. Simply growing a raid is
> apparently forgotten, since I thought that was too simple. I'll add it.
>
> So hope you didn't lose too much valuable data
>
> Vennlig hilsen / Best regards
>
> roy
> --
> Roy Sigurd Karlsbakk
> (+47) 98013356
> --
> I all pedagogikk er det essensielt at pensum presenteres intelligibelt. Det er
> et elementært imperativ for alle pedagoger å unngå eksessiv anvendelse av
> idiomer med xenotyp etymologi. I de fleste tilfeller eksisterer adekvate og
> relevante synonymer på norsk.
next prev parent reply other threads:[~2022-04-05 10:17 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-04-04 15:19 RAID 1 to RAID 5 failure Jorge Nunes
2022-04-04 16:42 ` Wols Lists
2022-04-04 17:17 ` Jorge Nunes
2022-04-05 0:29 ` Roy Sigurd Karlsbakk
2022-04-05 9:17 ` Roy Sigurd Karlsbakk [this message]
2022-04-05 10:50 ` Jorge Nunes
2022-04-05 11:30 ` Roy Sigurd Karlsbakk
2022-04-06 6:57 ` Jorge Nunes
2022-04-06 19:57 ` Wols Lists
2022-04-06 20:46 ` Jorge Nunes
2022-04-06 21:02 ` Roy Sigurd Karlsbakk
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=336816279.3605676.1649150230084.JavaMail.zimbra@karlsbakk.net \
--to=roy@karlsbakk.net \
--cc=jorgebnunes@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