From: Wols Lists <antlists@youngman.org.uk>
To: "Vojtěch Kletečka" <vojta.kletecka@gmail.com>,
linux-raid@vger.kernel.org
Subject: Re: Data recovery
Date: Thu, 21 Sep 2017 01:37:27 +0100 [thread overview]
Message-ID: <59C309C7.9030300@youngman.org.uk> (raw)
In-Reply-To: <CAEmSg0_=5ni2mA1Q_A1DjZmvjG9u8tKXOQHmnmT33=12yJwYqA@mail.gmail.com>
On 21/09/17 01:08, Vojtěch Kletečka wrote:
> Dear linux-raid subscribers,
>
> may I humbly request your assistance?
>
> I have three disks connected over usb from which I have made raid10
> array with two far coppies. Ocasionally when connecting all three
> disks to another computer one or two of the disks are not visible when
> creating array due to bad contacts on usb port. Sometimes also one of
> the disks suddenly disconnects from the array.
Why are they connected over USB?
> So far I always managed to solve the problems caused by these (very
> poor, I am fully aware of that now) conditions by readding disk and
> letting it synchronize with other disks. But now I apparently caused a
> bigger issue because I am not able to assemble the raid at all.
>
Raid doesn't like USB. Because USB likes dropping disks and then raid
wonders where the disk has gone - sounds exactly like what's happened to
you!
> As far as I remember when I connected the disks today, linux detected
> only one. After few tries I found right usb ports for all three disks
> and they were indeed shown in output of fdisk -l. But device was not
> assembled at first. I have waited a little bit and mdadm finally
> assembled the device (although I am not sure if there were two or
> three devices). Strangely "cat /proc/mdstat" reported that devices are
> 99.something synchronized but supposedly with all blocks synchronized
> (I mean both numbers were exactly the same) and 0kb speed. Weird I
> thought and I issued mdadm --stop /dev/md127 because I wasn't able to
> mount /dev/md127 to filesystem.
>
> That was last thing I was able to do with the array. After that I have
> tried all possible combinations of disks, assembling and incremental
> assembling with/without --force and/or --run but everything without
> any results. All commands just failed, in most cases with these two
> errors:
>
> sudo mdadm --assemble /dev/md127 /dev/sd[cde]1
> mdadm: /dev/md127 assembled from 2 drives - need all 3 to start it
> (use --run to insist).
> (this command caused "kicking non-fresh sdd1 from array!" in dmesg)
>
> sudo mdadm --assemble /dev/md127 /dev/sd[cde]1 --run
> mdadm: failed to RUN_ARRAY /dev/md127: Input/output error
> (this command caused "not enough operational mirrors." in dmesg)
>
> From various examining commands (see archive for all commands
> requested in "asking for help" part of wiki here:
> http://mysharegadget.com/328200519) I have concluded, that one of
> disks is far behind others in synchronization and other two disks
> strangely don't have enough mirrors (although I though there are
> exactly two disks where each block of data is written, at least in
> this case).
>
> sudo mdadm --examine /dev/sd[c-e]1 | egrep 'Event|/dev/sd'
> /dev/sdc1:
> Events : 44607
> /dev/sdd1:
> Events : 36570
> /dev/sde1:
> Events : 44607
>
> Is there any possibility of data recovery? For example by forcing
> mdamd to think that the disks are synchronized?
This looks good for data recovery ...
> As far as I know the disks should be synchronized because last time my
> brother used the array he let it synchronize afterwards. Also I
> haven't used "mdadm --create" yet as it might be dangerous and I am
> not sure if all data saved in array are backed up.
Thank $DEITY for that - do NOT use --create - at least don't use it if
you want to recover your data !!!
>
> Thank you for your time and any suggestions.
>
Get a SATA add in card if you don't have enough SATA ports, and get rid
of USB !!!
I'll let others chime in on how to get your array back - IF you put the
drives on SATA, I hope it's just a simple "--assemble --force" which
will result in everything sorting itself out, but we'll see. But don't
put USB drives into a raid array!
> With kind regards,
>
> Vojtěch Kletečka
Cheers,
Wol
next prev parent reply other threads:[~2017-09-21 0:37 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-09-21 0:08 Data recovery Vojtěch Kletečka
2017-09-21 0:37 ` Wols Lists [this message]
2017-09-21 12:10 ` Phil Turmel
2017-09-21 12:55 ` Vojtěch Kletečka
2017-09-21 13:13 ` Phil Turmel
2017-09-21 13:22 ` Vojtěch Kletečka
2017-09-21 14:37 ` Phil Turmel
[not found] ` <CAEmSg0-AD09KM=S21a-EVGQEEE0=em-PCpb9LAAbd8cP2+ysBA@mail.gmail.com>
2017-09-21 15:15 ` Wols Lists
2017-09-21 15:32 ` Wols Lists
-- strict thread matches above, loose matches on Subject: below --
2012-01-31 6:46 Data Recovery Swapnil Gaikwad
2012-01-31 8:04 ` Mulyadi Santosa
2012-01-31 12:40 ` arshad hussain
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=59C309C7.9030300@youngman.org.uk \
--to=antlists@youngman.org.uk \
--cc=linux-raid@vger.kernel.org \
--cc=vojta.kletecka@gmail.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.