Open Stack

If glare was docker repo api compatible though, I think it would be quite useful. then each tenant doesn't have to set one up themselves.

Thanks,
Kevin

________________________________
From: Hongbin Lu [hongbin.lu at huawei.com]
Sent: Friday, August 05, 2016 1:29 PM
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [Glance][TC][Heat][App-Catalog][Murano][Tacker] Glare as a new Project

Replied inline.

From: Mikhail Fedosin [mailto:mfedosin at mirantis.com]
Sent: August-05-16 2:10 PM
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [Glance][TC][Heat][App-Catalog][Murano][Tacker] Glare as a new Project

Thank you all for your responses!

>From my side I can add that our separation is a deliberate step. We pre-weighed all pros and cons and our final decision was that moving forward as a new project is the lesser of two evils. Undoubtedly, in the short term it will be painful, but I believe that in the long run Glare will win.

Also, I want to say, that Glare was designed as an open project and we want to build a good community with members from different companies. Glare suppose to be a backend for Heat (and therefore TripleO), App-Catalog, Tacker and definitely Nova. In addition we are considering the possibility of storage Docker containers, which may be useful for Magnum.

[Hongbin Lu] Magnum doesn’t have any plan to store docker images at Glare, because COE (i.e. Kubernetes) is simply incompatible with any API other than docker registry. Zun might have use cases to store docker images at Glare if Glare is part of Glance, but I am reluctant to set a dependency on Glare if Glare is a totally branch new service.

Then, I think that comparison between Image API and Artifact API is not correct. Moreover, in my opinion Image API imposes artificial constraints. Just imagine that your file system can only store images in JPG format (more precisely, it could store any data, but it is imperative that all files must have the extension ".jpg"). Likewise Glance - I can put there any data, it can be both packages and templates, as well as video from my holiday. And this interface, though not ideal, may not work for all services. But those artificial limitations that have been created, do Glance uncomfortable even for storing images.

On the other hand Glare provides unified interface for all possible binary data types. If we take the example with filesystem, in Glare's case it supports all file extensions, folders, history of file changes on your disk, data validation and conversion, import/export files from different computers and so on. These features are not presented in Glance and I think they never will, because of deficiencies in the architecture.

For this reason I think Glare's adoption is important and it will be a huge step forward for OpenStack and the whole community.

Thanks again! If you want to support us, please vote for our talk on Barcelona summit - https://www.openstack.org/summit/barcelona-2016/vote-for-speakers/ Search "Glare" and there will be our presentation.

Best,
Mike

On Fri, Aug 5, 2016 at 5:22 PM, Jonathan D. Proulx <jon at csail.mit.edu<mailto:jon at csail.mit.edu>> wrote:

I don't have a strong opinion on the split vs stay discussion. It
does seem there's been sustained if ineffective attempts to keep this
together so I lean toward supporting the divorce.

But let's not pretend there are no costs for this.

On Thu, Aug 04, 2016 at 07:02:48PM -0400, Jay Pipes wrote:
:On 08/04/2016 06:40 PM, Clint Byrum wrote:

:>But, if I look at this from a user perspective, if I do want to use
:>anything other than images as cloud artifacts, the story is pretty
:>confusing.
:
:Actually, I beg to differ. A unified OpenStack Artifacts API,
:long-term, will be more user-friendly and less confusing since a
:single API can be used for various kinds of similar artifacts --
:images, Heat templates, Tosca flows, Murano app manifests, maybe
:Solum things, maybe eventually Nova flavor-like things, etc.

The confusion is the current state of two API's, not having a future
integrated API.

Remember how well that served us with nova-network and neutron (né
quantum).

I also agree with Tim's point.  Yes if a new project is fully
documented and integrated well into packaging and config management
implementing it is trivial, but history again teaches this is a long
road.

It also means extra dev overhead to create and mange these
supporting structures to hide the complexity from end users. Now if
the two project are sufficiently different this may not be a
significant delta as the new docs and config management code would be
need in the old project if the new service stayed stayed there.

-Jon

__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe<http://OpenStack-dev-request@lists.openstack.org?subject:unsubscribe>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20160805/1ad8fb5c/attachment.html>