[Gluster-devel] Release 3.10 feature proposal: Disable creation of trash directory by default

Anoop C S anoopcs at redhat.com
Tue Dec 13 09:30:05 UTC 2016


On Mon, 2016-12-12 at 11:44 -0500, Shyam wrote:
> Anoop,
> 
> I have summarily marked this for 3.10, but have a few requests here,
> 
> 1) Can we open a spec for this?

We do have an accepted design spec under glusterfs-specs namely 'Trash Improvements'[2] among which
this particular proposal is already listed and detailed. Keeping that in mind we thought of not to
extract part of it and bring up another design doc. Would that be enough or do we really need an
exclusive spec file?

[2] https://github.com/gluster/glusterfs-specs/blob/master/accepted/Trash-Improvements.md

> 2) We possibly need to understand backward compatibility issues/concerns 
> if any? I.e existing volumes would already have created .trashcan etc. 
> and how this change impacts those volumes.

It is backward compatible in the sense that it does not break the promise given to consumers of
trash feature before upgrading to newer version. Moreover we are making it easier for those users
who doesn't require this feature on existing volumes to have an option for deleting the trash
directory if not needed. And of course for new volumes we no longer creates trash directory by
default which makes it more convenient for end users.

Being said that we will have to document this change of behavior in release notes for sure.

> 
> Thanks,
> Shyam
> 
> On 12/10/2016 10:02 AM, Anoop C S wrote:
> > Hi all,
> > 
> > As per the current design trash directory, namely .trashcan, will be created at the root when
> > bricks associated with a volume become online and there is a restriction to delete this
> > directory from the volume even when trash feature is disabled.
> > 
> > This proposal is targeted in a such a way that creation and subsequent enforcement on trash
> > directory to happen only when feature is enabled for that volume.
> > 
> > Issue opened at [1].
> > 
> > [1] https://github.com/gluster/glusterfs/issues/65
> > 
> > Thanks,
> > --Anoop C S
> > 


More information about the Gluster-devel mailing list