[Gluster-users] Disaster scenarios

[Niels de Vos]

On Fri, Sep 19, 2014 at 02:04:21PM +0100, Mark s2c wrote:
> I have 24 petabytes under management for example and a twice daily geo
> replication. A hapless junior SA executes a gluster volume delete. Is
> my only option to get to the georeplicated copy before it next runs
> and duplicates the deletes to the remote backup?

The 'gluster volume delete' command does not delete the contents of the
volume. The bricks that were used to store your contents, will still
have the data unmodified. It is possible to re-create the volume with
the same bricks to 'restore' the volume.

If a user of the volume deleted files, and you are interested in
recovering that before the delete operation is carried over to the
geo-replication slave, you will be interested in the Trash xlator [1].
There is a feature proposed for the upcoming 3.6 release that makes it
possible to find deleted files in a special 'trash' directory. It seems
that the feature was not expected to be completed within 3.6 timeline
[2], but the schedule got extended a little. I am not sure when the
Trash xlator will be considered stable (adding the engineers from the
wiki page on CC for their input and wiki update).

HTH,
Niels


1. http://www.gluster.org/community/documentation/index.php/Features/Trash
2. http://www.gluster.org/community/documentation/index.php/Planning36