Re: [openstack-dev] Consistent Versioned Endpoints
Sean Dague wrote: > On 01/12/2017 01:35 PM, Scott D'Angelo wrote: >> [...] >> Can we get to this "perfect world"? Let's discuss at the PTG. >> It is my understanding that we do not have the ability to schedule a >> time or room for such a cross-project discussion. Please chime in if >> interested, and/or make your interest known to scottda, mordred, or edleafe. > > Happy to join in on this, it does seem weird there is no time / space > for such things at PTG. We'll have a room available for such inter-team discussions at the PTG. However, since only a fragment of our community will be present at the PTG, we need to be careful to avoid exclusion. Ideally we would only use that room to discuss things that are only relevant to upstream development teams, and use the "Forum" in Boston to hold truly cross-project / community-wide discussions. The typical target for the discussion room at the PTG are therefore ad-hoc discussions between PTG teams, where a separate fishbowl room makes more sense than holding it in a specific team room. As far as scheduling goes, the "discussion" room at the PTG should be available from Monday to Thursday, and scheduled in unconference-style, to give flexibility to have the discussions we need to have. Current plan is to use an ethercalc document to share the schedule. For critical discussions (which don't belong to any given room, fit the cross-section of our community present, and/or can't wait until Boston) we could totally hardcode them in the schedule for the discussion room... but we should probably keep those to a minimum and use team rooms as much as possible. How much time do you think that discussion would need ? If more than a few hours it's probably just simpler to give it a full team room. -- Thierry Carrez (ttx) __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] Consistent Versioned Endpoints
On 1/13/17 7:42 AM, Steve Martinelli wrote: > On Fri, Jan 13, 2017 at 7:39 AM, Sean Dague wrote: > >> On 01/12/2017 01:35 PM, Scott D'Angelo wrote: >>> TL;DR: Let's discuss Version Discovery and Endpoints in the Service >>> Catalog at the PTG in Atlanta. >>> >>> The topic of Versioning and the Endpoints discovered in the Service >>> Catalog was discussed in today's API Working Group Meeting[1]. >>> A previous ML post[2] claimed: >>> >>> In a perfect world, every endpoint would return the same type of >> resource - >>> most likely the versions resource as described in the API WG >> Microversions >>> spec. It would also be nice if version negotiation can happen without >>> requiring authentication, the easiest path to which would be supporting >> the >>> 'max_version' and 'min_version' fields in the root versions resource. >>> >>> One problem is multiple versioned service names in the catalog for a >>> given service[3], as opposed to a single endpoint that would return >>> version info[4]. >>> >>> Can we get to this "perfect world"? Let's discuss at the PTG. >>> It is my understanding that we do not have the ability to schedule a >>> time or room for such a cross-project discussion. Please chime in if >>> interested, and/or make your interest known to scottda, mordred, or >> edleafe. >> >> Happy to join in on this, it does seem weird there is no time / space >> for such things at PTG. >> >> We actually had a rough sketch of a plan last year here which would go a >> slightly different direction, and get that all up into the service >> catalog. That also ensures consistent schema enforcement, as it would >> all be through keystone. >> >> Definitely need keystone folks in the room to be able to make forward >> progress here I think, because part of the tension in the past has been >> understanding domain boundaries, and it would be a shame to do a ton of >> work in all the projects that could be easily done in one project, just >> because of communication gaps. >> > > If theres a room I'll be there, and try to rope in other folks too. > This is possibly the worst suggestion of all time, but I've told the Glance team that we'll finish up by 11:30 a.m. on Friday, which will free up the Glance room for Friday afternoon. If other teams are taking a similar strategy, there wouldn't be too much conflict, and maybe we can hold a few sessions like this, which are important but seem to have slipped through the cracks, on Friday afternoon? __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] Consistent Versioned Endpoints
On Fri, Jan 13, 2017 at 7:39 AM, Sean Dague wrote: > On 01/12/2017 01:35 PM, Scott D'Angelo wrote: > > TL;DR: Let's discuss Version Discovery and Endpoints in the Service > > Catalog at the PTG in Atlanta. > > > > The topic of Versioning and the Endpoints discovered in the Service > > Catalog was discussed in today's API Working Group Meeting[1]. > > A previous ML post[2] claimed: > > > > In a perfect world, every endpoint would return the same type of > resource - > > most likely the versions resource as described in the API WG > Microversions > > spec. It would also be nice if version negotiation can happen without > > requiring authentication, the easiest path to which would be supporting > the > > 'max_version' and 'min_version' fields in the root versions resource. > > > > One problem is multiple versioned service names in the catalog for a > > given service[3], as opposed to a single endpoint that would return > > version info[4]. > > > > Can we get to this "perfect world"? Let's discuss at the PTG. > > It is my understanding that we do not have the ability to schedule a > > time or room for such a cross-project discussion. Please chime in if > > interested, and/or make your interest known to scottda, mordred, or > edleafe. > > Happy to join in on this, it does seem weird there is no time / space > for such things at PTG. > > We actually had a rough sketch of a plan last year here which would go a > slightly different direction, and get that all up into the service > catalog. That also ensures consistent schema enforcement, as it would > all be through keystone. > > Definitely need keystone folks in the room to be able to make forward > progress here I think, because part of the tension in the past has been > understanding domain boundaries, and it would be a shame to do a ton of > work in all the projects that could be easily done in one project, just > because of communication gaps. > If theres a room I'll be there, and try to rope in other folks too. __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] Consistent Versioned Endpoints
On 01/12/2017 01:35 PM, Scott D'Angelo wrote: > TL;DR: Let's discuss Version Discovery and Endpoints in the Service > Catalog at the PTG in Atlanta. > > The topic of Versioning and the Endpoints discovered in the Service > Catalog was discussed in today's API Working Group Meeting[1]. > A previous ML post[2] claimed: > > In a perfect world, every endpoint would return the same type of resource - > most likely the versions resource as described in the API WG Microversions > spec. It would also be nice if version negotiation can happen without > requiring authentication, the easiest path to which would be supporting the > 'max_version' and 'min_version' fields in the root versions resource. > > One problem is multiple versioned service names in the catalog for a > given service[3], as opposed to a single endpoint that would return > version info[4]. > > Can we get to this "perfect world"? Let's discuss at the PTG. > It is my understanding that we do not have the ability to schedule a > time or room for such a cross-project discussion. Please chime in if > interested, and/or make your interest known to scottda, mordred, or edleafe. Happy to join in on this, it does seem weird there is no time / space for such things at PTG. We actually had a rough sketch of a plan last year here which would go a slightly different direction, and get that all up into the service catalog. That also ensures consistent schema enforcement, as it would all be through keystone. Definitely need keystone folks in the room to be able to make forward progress here I think, because part of the tension in the past has been understanding domain boundaries, and it would be a shame to do a ton of work in all the projects that could be easily done in one project, just because of communication gaps. -Sean -- Sean Dague http://dague.net __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] Consistent Versioned Endpoints
On Thu, Jan 12, 2017 at 12:35 PM, Scott D'Angelo wrote: > Can we get to this "perfect world"? Let's discuss at the PTG. > It is my understanding that we do not have the ability to schedule a time or > room for such a cross-project discussion. Please chime in if interested, > and/or make your interest known to scottda, mordred, or edleafe. IIRC there may be some cross-project rooms available, but I can also offer time/space in the OpenStackClient room for this discussion if needed. dt -- Dean Troyer dtro...@gmail.com __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] Consistent Versioned Endpoints
TL;DR: Let's discuss Version Discovery and Endpoints in the Service Catalog at the PTG in Atlanta. The topic of Versioning and the Endpoints discovered in the Service Catalog was discussed in today's API Working Group Meeting[1]. A previous ML post[2] claimed: In a perfect world, every endpoint would return the same type of resource - most likely the versions resource as described in the API WG Microversions spec. It would also be nice if version negotiation can happen without requiring authentication, the easiest path to which would be supporting the 'max_version' and 'min_version' fields in the root versions resource. One problem is multiple versioned service names in the catalog for a given service[3], as opposed to a single endpoint that would return version info[4]. Can we get to this "perfect world"? Let's discuss at the PTG. It is my understanding that we do not have the ability to schedule a time or room for such a cross-project discussion. Please chime in if interested, and/or make your interest known to scottda, mordred, or edleafe. Scott D'Angelo scott.dang...@ibm.com [1] http://eavesdrop.openstack.org/meetings/api_wg/2017/api_wg.2017-01-12-16.00.log.html [2] http://lists.openstack.org/pipermail/openstack-dev/2016-August/102549.html [3] http://paste.openstack.org/show/594751/ [4] http://paste.openstack.org/show/594754/ __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
