Open Stack

Hi Rob,

Thanks, initial thoughts regarding your questions are inline below.

John

On Wed, Sep 12, 2012 at 10:43 AM, Esker, Robert <Rob.Esker at netapp.com>wrote:

>  Hi John,
>
>  Good to see you running and thanks for your efforts thus far...  I do
> have a few questions about your thoughts on Cinder's future:
>
>  • We (NetApp) have heard from folks looking to potentially use Cinder
> independently of Nova, et cetera (as a generalized storage backend
> abstraction layer).  Where do you stand on retaining the option to do so?
>
> I would strongly agree with this mindset.  This was in fact part of a
number of discussions at the Folsom summit last spring.  Ideally I would
like to see Cinder mature to a point where it could be used exactly in this
sort of manner.  Plugged in to different architectures and used as it's own
entity with or without Nova, but obviously working with Nova is the main
charter and first priority.

 • We've also been asked by several in the operator (or soon to be)
> community to look at expanding the role of Cinder beyond block storage
> only.  This was touched upon a bit at the last summit.  Given the
> significant effort associated with breaking out Cinder it wasn't a focus at
> the time though.  We're planning talks on this at the Grizzly design summit
> next month, but more significantly have prototyped Cinder support for
> accommodating shared filesystems.  We've specifically avoided submitting
> most of it while Folsom is in play, but will do so as soon as Grizzly revs
> up.  Could you please comment on your philosophy for the future of Cinder
> as it applies to extending the API for such?
>

I don't think my opinion regarding this has ever been much of a secret
(regardless of how unpopular it may be) :)  During the initial proposals to
start the Cinder project I shared my opinion that the idea was to provide
an independent block storage service that would be consumed by Nova and
possibly others.  I've never been a big fan of the idea of shoe horning in
all sorts of extras (network shares, file systems etc.).  My experience has
been that if you don't have a very clear and focused definition regarding
what your purpose is you tend to not be very good at anything.  I'd like to
keep Cinder focused on being a Block Storage service, not an NFS service or
other shared file system service etc.

That being said, I also understand the needs/desires from folks in the
community.  As you know NetApp submitted changes to Cinder to enable NFS
support and they were accepted and merged.  I would never object to
something that the majority of team members and especially end users ask
for. I would just like to avoid trying to be everything to everyone and in
turn not being very good at any of it.  In my opinion if changes are
somewhat isolated and the only real impact is things like new drivers (such
as the NFS addition) without impacting the API etc then I'm all for it.

I also think that this is a philosophy that can evolve over time as the
project matures.  Right now I see a significant need for improvements to
block storage and that's where I would like to focus efforts.  Down the
road when Cinder is the *best* block storage service that it can be, that
opens up the door to a change in philosophies and focus.


>
Thanks!
>
>  Rob Esker
> NetApp, Inc.
>
>
>  On Aug 31, 2012, at 3:27 PM, John Griffith <john.griffith at solidfire.com>
>  wrote:
>
>  Hello,
>
> For the many folks that don't know me, my name is John Griffith and I'd
> like to announce my candidacy for Cinder PTL.
>
> Qualifications:
>
> I've been an active contributor to Nova since November of 2011, starting
> out as many by adding a driver, but then began helping out anywhere and
> everywhere I could. IRC questions/conversations, assigning bugs to myself
> etc.  In addition I've been involved in multiple conversations with
> customers interested in OpenStack and have been acting as an OpenStack
> evangelist to a number of both small and large companies looking to deploy
> OpenStack in public and private clouds.
>
>
> Contribution over the last six months:
>
> During the last summit I lead conversations around the idea of Cinder and
> organized the effort to kick off the project. Since then I've been acting
> as interim PTL for the Cinder project and have successfully driven the
> effort to promote Cinder to core status as an OpenStack project. I'm
> continually learning, and absolutely love doing the work!
>
> While it's been a learning experience I have been fulfilling the
> responsibilities or PTL, and dedicated every day to the success of the
> project. I'm passionate about the success of Cinder and believe it has vast
> potential and is critically important to OpenStack as a whole.
>
>
> Most critical aspects for Cinder in the next 6 months:
>
> The next six months are going to be defining for Cinder. The most
> important thing is going to be showing a smooth transition from Nova-Volume
> to Cinder is possible. There's a lot of concern regarding this subject and
> it's something that I keep in mind with any change/fix or addition to the
> code.
> Aside from the initial release, the critical aspects for Grizzly are going
> to be gain momentum in terms of visibility and participation, focus on
> quality and provide new features. Quality is my number one priority, and I
> think there is a lot of potential to harden the core code in the project
> and greatly improve the user experience as well as capabilities.
>
>
> Philosophical ideas regarding being a PTL:
>
> There are of course all of the items listed in the being a PTL guide. In
> addition however I think it is very important for a PTL to be accessible in
> a multitude of ways. This includes questions, helping people that want to
> get involved with OpenStack and most of all listening to ideas/input. One
> of the things that makes Open Source work is the vast pool of ideas and
> suggestions from various views and backgrounds, keeping an open mind and
> constantly challenging existing paradigms leads to a better project.
> I believe it's crucial to build a strong core team, listen to input from
> users, storage vendors and other developers and make something that is
> truly better than anything else that's out there. I don't believe we should
> strive to offer the *same* functionality as other cloud platforms, but
> offer something that is MUCH better!
>
>
>
> If you have any questions or there are any further details I can provide
> don't hesitate to ask.
>
> Thanks,
> John
>
>  email:    john.griffith at solidfire.com
> irc:        jgriffith
>
> _______________________________________________
> Mailing list: https://launchpad.net/~openstack
> Post to     : openstack at lists.launchpad.net
> Unsubscribe : https://launchpad.net/~openstack
> More help   : https://help.launchpad.net/ListHelp
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20120912/01b3323c/attachment.html>