linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Austin S. Hemmelgarn" <ahferroin7@gmail.com>
To: Tomasz Chmielewski <mangoo@wpkg.org>,
	linux-btrfs <linux-btrfs@vger.kernel.org>
Subject: Re: yet another "out of space" on a filesystem with >100 GB free space, and strange files which exist but don't exist
Date: Wed, 4 Oct 2017 07:20:19 -0400	[thread overview]
Message-ID: <e2b27980-dabf-540b-c4c1-8628cc1d47c1@gmail.com> (raw)
In-Reply-To: <7c6a0bb4ff9b4f09c7ea6f6c7173a994@wpkg.org>

On 2017-10-04 07:13, Tomasz Chmielewski wrote:
> Kernel: 4.13.4, btrfs RAID-1.
> 
> Disk usage more or less like below (yes, I know about btrfs fi df / show 
> / usage):
> 
> Filesystem      Size  Used Avail Use% Mounted on
> /dev/sda3       424G  262G  161G  62% /var/lib/lxd
> 
> 
> Balance would exit immediately with "out of space", but continues to run 
> after I've removed a few gigabytes from the filesystem.
> 
> 
> Now, I'm seeing some files which exist, but don't. Strange, I know.
> 
> 
> root@lxd02 /var/lib/lxd/containers/mongo-repl04b/rootfs/var/lib/mongodb 
> # ls *set
> ls: cannot access 'WiredTiger.turtle.set': No such file or directory
> 
> root@lxd02 /var/lib/lxd/containers/mongo-repl04b/rootfs/var/lib/mongodb 
> # ls -l|grep set
> ls: cannot access 'WiredTiger.turtle.set': No such file or directory
> -????????? ? ?      ?                ?            ? WiredTiger.turtle.set
> 
> root@lxd02 /var/lib/lxd/containers/mongo-repl04b/rootfs/var/lib/mongodb 
> # mv WiredTiger.turtle.set WiredTiger.turtle.set.Ghost.File
> mv: cannot stat 'WiredTiger.turtle.set': No such file or directory
> 
> root@lxd02 /var/lib/lxd/containers/mongo-repl04b/rootfs/var/lib/mongodb 
> # rm -v WiredTiger.turtle.set
> rm: cannot remove 'WiredTiger.turtle.set': No such file or directory
> 
> 
> 
> What is this file, and why does it exist if it doesn't? How do I remove it?
It's got corrupted metadata, probably the inode itself (IIRC, the dentry 
in BTRFS just matches the inode to the file name, and all the other data 
reported by ls -l is stored in the inode).  If you're running with a 
replicated metadata profile (dup, raid1, or raid10), run a scrub, and it 
may fix things.  If not, you will likely have to run a check in repair 
mode (though I would suggest waiting to hear from one of the developers 
before doing so).  Alternatively, if that's in a subvolume, and you can 
afford to just nuke the subvolume and recreate it, deleting the 
subvolume should get rid of it (though you should still run a check).

Either way, this is likely related to the balance issues you're seeing.

  reply	other threads:[~2017-10-04 11:20 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-04 11:13 yet another "out of space" on a filesystem with >100 GB free space, and strange files which exist but don't exist Tomasz Chmielewski
2017-10-04 11:20 ` Austin S. Hemmelgarn [this message]
2017-10-04 12:46   ` Tomasz Chmielewski
2017-10-04 13:25     ` Dmitrii Tcvetkov

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=e2b27980-dabf-540b-c4c1-8628cc1d47c1@gmail.com \
    --to=ahferroin7@gmail.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=mangoo@wpkg.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).