14:00:02 <nikhil_k> #startmeeting Glance
14:00:03 <openstack> Meeting started Thu Jul  9 14:00:02 2015 UTC and is due to finish in 60 minutes.  The chair is nikhil_k. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:07 <openstack> The meeting name has been set to 'glance'
14:00:26 <nikhil_k> #topic agenda and roll call
14:00:33 <nikhil_k> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda
14:00:53 <rosmaita> o/
14:00:55 <abhishekk> o/
14:00:56 <bpoulos> o/
14:00:58 <nikhil_k> Let's see who is around
14:00:58 <mfedosin> o/
14:00:59 <krykowski> o/
14:01:00 <lakshmiS> o/
14:01:04 <TravT_away> o/
14:01:05 <sigmavirus24> o/
14:01:14 <sigmavirus24> kragniz: you around?
14:01:16 <dshakhray> o/
14:01:18 <kragniz> o/
14:01:38 <kragniz> sigmavirus24: aye aye
14:01:44 <sigmavirus24> :D
14:01:52 <nikhil_k> Great turnout!
14:01:57 <nikhil_k> Let's get started
14:02:07 <nikhil_k> #topic Updates
14:02:26 <nikhil_k> #info Artifacts sub-team meeting update
14:02:53 <nikhil_k> We had a decent discussion. ativelkov was working on some oslo-vo changes.
14:03:12 <mfedosin> hi, Alex is not here, but I suppose he is still working on oslo.vo :)
14:03:15 <nikhil_k> API evolution is in progress and needs to be discussed further with the WG.
14:03:39 <krykowski> doesn't he need any help on oslo.vo? :)
14:03:41 <nikhil_k> I guess Alex is working on the summary
14:03:42 <mfedosin> we started to work on stable api yesterday
14:03:54 <nikhil_k> cool
14:04:27 <mfedosin> krykowski, he said that his changes break nova
14:04:51 <mfedosin> and it's not easy to fix because it requires a lot of work on the nova side
14:05:22 <nikhil_k> also, mfedosin is working on the client.
14:05:23 <mfedosin> so Alex finds ways how to workaround that
14:05:33 <jokke_> o/
14:05:36 <sabari> o/
14:05:39 <nikhil_k> The question raised was whether to move ahead with client before the API stabilizes or after
14:05:51 <krykowski> mfedosin, cool, if any help needed I'm in here ;)
14:06:01 <mfedosin> krykowski, thanks
14:06:44 <jokke_> nikhil_k: I'd say this is brilliant opportunity to utilize feature branch on client and go parallel for a while
14:07:12 <jokke_> releasing from master and having the artifact stuff following on feature branch
14:07:13 <mfedosin> I think we can merge the existing code and if api changes we will change the client code too
14:07:13 <nikhil_k> umm, what other need exist for feature branch jokke_ ?
14:07:28 <jokke_> nikhil_k: the branch needs to be created
14:07:54 <mfedosin> jokke_, hmm, good idea, btw
14:07:56 <jokke_> nikhil_k: then obviously if we want CIs running tests they need to be configured
14:07:59 <nikhil_k> mfedosin: there are some other factors like the semver versioning of the client API
14:08:24 <nikhil_k> mfedosin: if it's gunna change we shouldn't add it to master
14:08:34 <nikhil_k> jokke_: makes sense
14:08:51 <nikhil_k> I doubt it we should start the CI before API is fixed
14:08:58 <jokke_> #link http://docs.openstack.org/infra/manual/drivers.html#feature-branches
14:09:03 <nikhil_k> the client question also applies to tempest changes
14:09:27 <nikhil_k> We are running behind
14:09:32 <mfedosin> I strongly agree with jokke_'s suggestion
14:10:06 <nikhil_k> Let's move on for now. We can continue next week on this
14:10:13 <jokke_> that would make the testing bit easier while yest it will have bit of maintenance overhead with rebases to master
14:10:15 <nikhil_k> #info Drivers' team meeting update
14:10:30 <sigmavirus24> +1 jokke_
14:10:32 <nikhil_k> jokke_: yeah
14:11:19 <nikhil_k> For the drivers' meeting, we chatted a bit more on the drivers' team evolution and the specs process
14:11:44 <nikhil_k> I guess people voted on Stuart to be added to the team, given he wants to and has time
14:12:22 <nikhil_k> One more thing was we don't have specs merged. This process needs some change and documentation.
14:13:11 <nikhil_k> Also, Liberty is unlikely to have many specs merged as our focus is stability, support of v2 in Nova as possible given we match their constraints
14:13:48 <nikhil_k> And we need to have Artifacts in a state that's acceptable to the community - that may mean moving the API out of the EXPERIMENTAL mode
14:15:18 <nikhil_k> I guess the first thing to cover would be finding the specs that are likely to be merged in Liberty and are not in priority.
14:15:52 <nikhil_k> Fox example: I mentioned to the person who was interested in impl quotas that the feature seems rather difficult even for L3
14:16:31 <sigmavirus24> nikhil_k: agreed on that
14:17:10 <sigmavirus24> personally I hope to have some time soon to review bpoulos's work on image signing and encryption. we said at the summit we wanted to pull that in
14:17:40 <nikhil_k> yeah, I would like to see that one through. bpoulos , sigmavirus24
14:17:45 <bpoulos> sigmavirus24: I would appreciate that
14:18:09 <sigmavirus24> bpoulos: I hope to rope in reaperhulk too
14:18:21 <sigmavirus24> (who was at your session standing over to your right against the wall)
14:18:31 <bpoulos> sigmavirus24: great, yeah, I've met him before
14:19:12 * flaper87 kinda late
14:19:27 <nikhil_k> Let's move on for now. We can chat more if needed in open discussion
14:19:44 <nikhil_k> #topic M-Release name
14:20:07 <nikhil_k> I see kragniz updated the etherpad?
14:20:27 <nikhil_k> what's more likely than Meiji (明治) ?
14:20:41 <kragniz> from the ML thread, it looks like another name will be picked
14:21:04 <nikhil_k> thanks
14:21:22 <nikhil_k> #topic Return request ID to caller
14:21:27 <jokke_> Meiji had apparently some minority target issues ... normal open OpenStack communication principles followed and none of the reasoning was told
14:21:43 <nikhil_k> :)
14:21:56 <flaper87> That name changed
14:22:03 <sigmavirus24> jokke_: the reasoning was explained in the thread that asked we not use it
14:22:09 <flaper87> sorry, lag
14:22:14 <flaper87> I think it's fine now
14:22:16 <sigmavirus24> flaper87: you should be sorry =P
14:22:24 <flaper87> sigmavirus24: :D
14:23:20 <nikhil_k> We had some discussion on the Return request ID proposal in the CPL meeting.
14:23:49 <nikhil_k> The proposal is being reworked to impl return of obj (it's WIP)
14:23:54 <abhishekk> nikhil_k: yes, I am working on POC for solution provided by lifeless
14:24:07 <nikhil_k> #link http://eavesdrop.openstack.org/meetings/crossproject/2015/crossproject.2015-07-07-21.01.log.html#l-58
14:24:30 <abhishekk> #link https://review.openstack.org/156508
14:24:37 <nikhil_k> That one seems interesting and was hoping to get more feedback based on the use case lifeless described (from folks here)
14:25:00 <nikhil_k> may be there are more return val expected than req IDs
14:25:11 <nikhil_k> input at this point would be nice
14:25:28 <abhishekk> nikhil_k: most probably in next week CPL meeting i will come up with code
14:25:41 <nikhil_k> abhishekk: thanks
14:25:47 <nikhil_k> #topic Travel Support program
14:25:51 <sigmavirus24> I'd like to see the POC before weighing in. Lifeless' suggestion makes sense
14:25:58 <nikhil_k> #info for awareness https://wiki.openstack.org/wiki/Travel_Support_Program
14:26:38 <nikhil_k> Those who want to travel to Tokyo, this is a nice opportunity to get funding outside of the regular corporate channels
14:27:13 <nikhil_k> sigmavirus24: thanks
14:27:20 <nikhil_k> #topic Rolling upgrade  (krykowski)
14:27:58 <krykowski> the first question is if we are aware of any incoming issues related to upgrading Glance with no downtime between releases, e.g. Kilo -> Liberty?
14:28:19 <krykowski> as we are api-oriented (which is stateless), we do not provide backwards incompatible changes and as I know we do not use internal rpc communication between services we should be free of issues when upgrading Glance to newer version. is that correct, or have missed something?
14:28:51 <nikhil_k> krykowski: I guess the questions here is does oslo.vo support the CRUD style dynamic prop protections?
14:29:04 <jokke_> have anyone paid any attention lately to make that happen? :P
14:29:27 <jokke_> I think we need to do API version bump on this cycle again
14:29:39 <nikhil_k> jokke_: oops, any link?
14:29:42 <flaper87> jokke_: what changes require it ?
14:30:18 <krykowski> nikhil_k: what do you mean by CRUD style dynamic prop protections?
14:30:19 * jokke_ tries to find ... just reviewed it like 2 days ago
14:30:45 <nikhil_k> krykowski: we have a possible rpc channel from g-api to g-reg
14:30:49 <flaper87> jokke_: fwiw, I did mention in a change that approving it would require a version bump, I just don't recall which one
14:31:26 <nikhil_k> krykowski: Data consistency when you are allowed to read a subset of metadata and write another subset of the metadata units
14:31:30 <flaper87> I wonder if we need v.o at all for rolling upgrades in glance
14:31:55 <flaper87> v.o, afaiu, were originally created to communicate with services like nova-conductor
14:32:09 <flaper87> which would act as the *only* interface on the database
14:32:20 <krykowski> nikhil_k: will need to get familiar with the g-reg
14:32:20 <flaper87> in glance's case, it could either be glance-api and/or glance-registry
14:32:25 <nikhil_k> well for offline migrations and static stlye data passing between services
14:32:38 <jokke_> flaper87: I think we're talking about the same change
14:32:56 <flaper87> but I don't think we've had backwards incompatible changes in the schema lately
14:33:02 <flaper87> but I might be wrong there
14:33:12 <jokke_> there has been also quite a few proposals with APIImacts flying lately, not sure how many of them merged
14:33:17 <flaper87> mentioning the schema since that's what v.o would help with
14:33:30 <nikhil_k> It's not just DB schema that matters afaict
14:33:47 <flaper87> what I'm trying to say is that, I'd love to see identified what problems we have for rolling upgrades first
14:33:57 <jokke_> ++
14:33:59 <flaper87> nikhil_k: I know, but that's what v.o would help with
14:34:09 <nikhil_k> it's the image locations migration as first example
14:34:31 * flaper87 was talking about v.o
14:34:35 <flaper87> sorry for any confusion
14:34:52 <nikhil_k> offline migration
14:35:07 <nikhil_k> as an example: image loc that has creds
14:35:28 <nikhil_k> and we want to move it to using configs (swift store)
14:35:49 <flaper87> Can we, as we normally do, identify the issues we have? it'd be really useful to work on a better plan to solve those issues
14:35:57 <nikhil_k> yep
14:36:11 <jokke_> so yeah sorry if I was too aggressive on that, but IMO we don't have rolling upgrade per default before someone points out that there is no issues, due to the fact that it's not priority (unfortunately) in our dev process
14:36:14 <flaper87> nikhil_k: the locations one is a good (and perhaps tough) one
14:36:18 <nikhil_k> the spec should ideally have that in the problem it tries to solve
14:36:45 <krykowski> sure is, will dig a bit and update the spec with specific issue
14:36:59 <flaper87> krykowski: thank you. FWIW, I think this is an important effort
14:37:00 <nikhil_k> ++
14:37:09 <flaper87> I'm happy you are willing to spend time on it
14:37:14 <jokke_> ++
14:37:37 <krykowski> flaper87: i'm here to serve you ;)
14:38:09 <nikhil_k> thanks krykowski . I guess we can revisit this in the later week.
14:38:26 <krykowski> nikhil_k: yup, thanks for the direction!
14:38:34 <nikhil_k> #topic Reviews, Bugs and Releases
14:39:01 <kragniz> we need a new glance_store release for the py34 work
14:39:13 <flaper87> kragniz: wooooohhhhoooooooooo
14:39:22 <nikhil_k> krykowski: lifeless is the POC for that
14:39:38 <nikhil_k> oops, I meant kragniz
14:39:52 <kragniz> nikhil_k: yeah, just bringing it up here
14:39:58 <nikhil_k> jokke_: was that you on python-glanceclient stable/kilo
14:39:59 <jokke_> also queued up bunch of stable stuff ... glance_store gotta wait until Cinder gets their client requirements fixed, but glanceclient backports would need some love
14:40:00 <nikhil_k> ?
14:40:03 <nikhil_k> cool
14:41:05 <jokke_> if we merge the few in review atm. I think it would be good cut off to kick a release out of door as well
14:41:08 <sigmavirus24> kragniz: thanks for bringing that up
14:42:01 <wko> I could really use some core reviews on https://review.openstack.org/#/c/195820/
14:42:45 <flaper87> wko: hey thanks for raising that but please, wait until the open discussion
14:43:06 <wko> ok, sorry
14:43:10 <nikhil_k> jokke_: any more updates for us?
14:43:54 <jokke_> nikhil_k: not really ... seems that tests are bringing up .eggs/ so proposed gitignore for them now for all the repos, that's all from me now
14:44:09 <nikhil_k> cool, thanks
14:44:25 <nikhil_k> #topic Open Discussion
14:44:36 <nikhil_k> wko: flaper87 ^
14:44:37 <jokke_> and obviously, please people when fixing bugs that does not change APIs remeber to mark them backport potential and/or propose backports so we get the stables fixed ;)
14:45:08 <wko> I could really use some core reviews on https://review.openstack.org/#/c/195820/ :)
14:45:35 <jokke_> wko: sorry I was eyeing that but it was too databasey for me ;)
14:45:44 <flaper87> I'll be out starting tomorrow until the 20th. If folks need something from me, this is the right moment
14:45:45 <flaper87> :D
14:45:59 <wko> and a +workflow on https://review.openstack.org/#/c/138612/
14:46:06 <jokke_> flaper87: the stable reviews ;)
14:46:32 <flaper87> jokke_: I was doing them on gertty
14:46:32 <mfedosin> btw, 200k was today
14:46:34 <nikhil_k> wko: flaper87 wait, thanks waiting on thespec
14:46:36 <flaper87> I guess it hasn't synced
14:46:43 <mfedosin> https://review.openstack.org/#/c/200000/
14:47:06 <jokke_> flaper87: <3 after than you can enjoy your holidays ;)
14:47:29 <flaper87> jokke_: w0000h0000 :D
14:47:47 <jokke_> mfedosin: gratz!!!
14:48:11 <mfedosin> danke :) Darja will finish it today
14:48:12 <flaper87> nikhil_k: https://review.openstack.org/#/c/154695/ that one could use your +2A (or other vote) when you get a chance
14:48:18 <jokke_> mfedosin: I'll just -2 that so we can freeze it forever, feel free to repropose ;)
14:48:49 <nikhil_k> thanks flaper87 . will take a look
14:49:45 <mfedosin> jokke_, ;) then we will have unhandled exceptions from store forever
14:50:18 <jokke_> mfedosin: that's a small prize to pay :P
14:51:29 <nikhil_k> If nothing more, we can conclude the meeting a few mins early.
14:51:40 <flaper87> that's what I call success
14:51:45 <jokke_> sounds good ... thanks everyone!
14:51:47 <nikhil_k> ;)
14:51:50 <flaper87> cheers
14:51:53 <nikhil_k> Thanks all!
14:51:59 <nikhil_k> have a nice vacay flaper87
14:51:59 <kragniz> flaper87: have a good holiday!
14:52:03 <nikhil_k> #endmeeting