public inbox for linux-xfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Eric Sandeen <sandeen@sandeen.net>
To: xfs@oss.sgi.com
Subject: Re: Error in `xfs_repair': double free or corruption
Date: Fri, 16 Oct 2015 11:29:19 -0500	[thread overview]
Message-ID: <562125DF.3050006@sandeen.net> (raw)
In-Reply-To: <trinity-74aaa980-3667-4c6d-9e1e-60f3f70dc888-1445005584585@3capp-gmx-bs54>



On 10/16/15 9:26 AM, Dragon wrote:
> Hello Emmanuel,
> 
> kernel version is 3.16-0.bpo.2-amd64 and xfsprogs in version 3.2.1
> from stable. the question is if xfs_repair can repair a raid which
> has missed a disk with over 240gb bad blocks.

Well, we shouldn't hit that "double free or corruption" bug.  But you
have an old xfsprogs, so the suggestion of using something up to date
is a good one.

As for "can xfs_repair repair a raid which is missing a disk with 240g
of bad blocks?" -

xfs_repair cannot make something out of nothing; it should always end
up with a filesystem which has consistent metadata, but that might be
the result of throwing away a lot of un-fixable things.

All too often I see people try to use xfs_repair on a raid which is
in bad shape; not properly reassembled, etc, and that's only going to
make things worse.  You must get your storage into the best shape you
can, first, and then xfs_repair will do its best to create consistency,
but it cannot recreate lost (meta)data, in general.

Running xfs_repair -n, or creating a metadump image, and running a full
repair on that as a test, is often a good idea before you commit to
changes on your original filesystem.

-Eric

> gretings
> 
> 
>     Hello,
>     i have a problem after repaired a degraded Softare Raid6. System is
>     Debian Jessi and Root and Boot are on Raid1, and Data on Raid6 on md2
>     which miss one Disk. The mainproblem was a missing Disk because of
>     badblocks which i cloned and exchanged agains a new one. After that i
>     could successfully rebuild the raid (md2), while now only one disk is
>     midding. But i recognised that xfs reports inode problems. Therefor i
>     run xfs_repair on the raid device md2, but get message "Error in
>     &#96;xfs_repair': double free or corruption" What can i do the fix
>     this? thx 
> 
> 
> You didn't provide enough information (kernel and xfs-progs version at
> the very least) but you should use the latest version of xfs_repair, the
> one included in Debian is probably quite old. Here's one new:
> 
> http://update.intellique.com/pub/xfs_repair-4.2.0.gz
> 
> Or if you like it better in debian form:
> http://update.intellique.com/repository/pool/storiq3/unstable/amd64/xfsprogs_4.2.0+1_amd64.deb
> 
> 

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

  reply	other threads:[~2015-10-16 16:29 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-10-16 14:26 Error in `xfs_repair': double free or corruption Dragon
2015-10-16 16:29 ` Eric Sandeen [this message]
  -- strict thread matches above, loose matches on Subject: below --
2015-10-19 11:16 Dragon
     [not found] <trinity-16122364-b1e6-4dc9-9057-f2df3aa193bb-1445001446455@3capp-gmx-bs54>
2015-10-16 13:18 ` Dragon
2015-10-16 13:26   ` Emmanuel Florac

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=562125DF.3050006@sandeen.net \
    --to=sandeen@sandeen.net \
    --cc=xfs@oss.sgi.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox