[Gluster-users] gluster verbosity

Niels de Vos ndevos at redhat.com
Tue Jan 20 09:09:23 UTC 2015


On Mon, Jan 19, 2015 at 02:33:55PM +0100, Nathanaël Blanchet wrote:
> Hello,
> 
> Most of my gluster mounting points (in replica mode) are  full of sequential
> directories (a0, a1, etc...) in .glusterfs. Those data usually reach several
> GB and finally block the volume.
> Are those data usefull and is there a way to reduce them, maybe with less
> verbosity?

These directories are used to find files by their GFID (volume wide
inode). The actual files are hardlinks to the files on the volume/brick
and do not take any extra space. The symbolic links you find under
.glusterfs/ point to directories on the volume/brick.

Some more details have been posted here:

    http://joejulian.name/blog/what-is-this-new-glusterfs-directory-in-33/

Does this clear things up?

Niels
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20150120/72a12f0f/attachment.sig>


More information about the Gluster-users mailing list