[Gluster-Maintainers] Gluster 4.0 release: Update from maintainers meeting

Vijay Bellur vbellur at redhat.com
Wed Aug 23 19:16:33 UTC 2017

On Sun, Aug 20, 2017 at 4:13 AM, Amar Tumballi <atumball at redhat.com> wrote:

> Hey Ants,
> Below is the message we want to get out from ant colony to users and
> developer community this week. (Was pending AI on me since 2 weeks). I want
> each of you to review this and give a Ack/No Ack on this.
> Also if you think major edits are required for message, we can collaborate
> on hackmd page :
> https://hackmd.io/GwIwnGBmCsCM0FoDMsAcBTBAWADLLCYATLJiOmBVgMYAmAhvbLUA#
> ----------------
> Hello,
> 2 weeks back we (most of the maintainers of Gluster projects) had a
> meeting, and we discussed about features required for Gluster 4.0 and also
> the possible dates.
> Summary:
>    - It is agreed unanimously that the Gluster 4.0 should be feature
>    based release, and not just time based.
>    - The discussions were on what are the blocking features for 4.0
>    release, and below were some features which we are considering as blocker
>    for release.
>       -
> *glusterd2: *
> The scalable management interface, which will come with REST api and other
> admin friendly options. Will be the ONLY major blocker for release.
>    - *Thin clients (code: gfproxy):*
> Not a blocker as it is slated to get into master branch before next
> release cut on 3.x series too. But we are calling it a major feature of 4.0
> as it brings many benefits and also change some of the assumptions and
> current design principles of glusterfs project.
>    - *Protocol/XDR changes:*
> Technically, Gluster's RPC is implemented in a way to support changing of
> protocols inbetween even the minor releases, but the major changes like
> adding another fop (like statx(), fadvise() or similar) will need some
> protocol changes.
> Also in discussion is the changes to dictionary structure on wire. We will
> continue to work on each of this individually, and plan to work towards
> meeting the release timeline.

Looks good to me. We could add Halo Replication to other features that
users can anticipate if  there are volunteers to work / test it out. We
have not talked about this feature in any of our release messaging.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/maintainers/attachments/20170823/0d20d696/attachment.html>

More information about the maintainers mailing list