14:00:37 <flaper87> #startmeeting Glance
14:00:38 <openstack> Meeting started Thu Jan  7 14:00:37 2016 UTC and is due to finish in 60 minutes.  The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:39 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:41 <openstack> The meeting name has been set to 'glance'
14:00:44 <flaper87> o/
14:00:53 <jokke_> o/
14:00:59 <avarner_> \o
14:01:00 <sigmavirus24> \o
14:01:07 <jokke_> Happy New Year every1!
14:01:29 <flaper87> Indeed. Happy new year!
14:01:29 <sabari> o/
14:01:34 <flaper87> hope you all had an amazing time off
14:01:43 <flaper87> for those that took time off
14:01:55 <flaper87> and that actually celebrate this thing
14:02:02 <flaper87> otherwise, carry on.
14:02:03 <flaper87> :P
14:02:06 <rosmaita> o/
14:02:18 <flaper87> mmh, there are just 4 of us.
14:02:20 <flaper87> 5
14:02:36 <flaper87> one more and we can start for realz
14:03:04 <flaper87> or I'll just ignore the missing folks and we'll start anyway :P
14:03:17 <flaper87> #topic Agenda
14:03:19 <flaper87> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda
14:03:21 <flaper87> That's our agenda, as usual. Please, add new topics at the bottom if there are things you'd like to talk about.
14:03:39 <flaper87> I don't think we'll go through it, entirely, as there are some topics that I'd rather touch with a broader audience
14:03:49 <kragniz> o/
14:04:02 <flaper87> For instance, we'll skip the artifacts updates as I don't see alex, nikhil nor mfedosin around
14:04:06 <hemanthm> o/
14:04:26 <flaper87> #topic Updates Drivers
14:04:52 <flaper87> Ok, we skipped 2 meetings (I was on holidays on Tuesday and I don't think anyone took over)
14:04:57 <flaper87> But we still have some updates
14:05:01 <flaper87> The image import spec landed
14:05:03 <flaper87> w00000h00000
14:05:08 <sabari> Yey!
14:05:15 <flaper87> (this is the part where rosmaita pastes the link using #link)
14:05:18 <flaper87> :P
14:05:25 <rosmaita> dang
14:05:40 <jokke_> #link https://review.openstack.org/#/c/232371
14:05:44 <flaper87> So, I believe the major changes have been laid down on the spec and that it doesn't require major refactors.
14:05:46 <jokke_> sorry Brian
14:06:03 <flaper87> If you think there are teaks and things that could be improved or just entirely horrible then, please, propose a follow-up patch
14:06:07 <rosmaita> #link http://specs.openstack.org/openstack/glance-specs/specs/mitaka/approved/image-import/image-import-refactor.html
14:06:08 <flaper87> it's a git repo and we have gerrit
14:06:31 <rosmaita> stuart has some reservations that i want to address
14:06:43 <flaper87> right, I still need to get back to those
14:07:00 <flaper87> I'd like to propose to Stuart to discuss those in a follow-up patch w/ his idea
14:07:16 <flaper87> I believe it'd help clearing out what he exactly wants we can comment *just* on that change
14:07:24 <rosmaita> yes, we can enhance the "alternatives" section
14:07:37 <flaper87> That being said, please, to everyone assigned to that spec. GET TO TI!
14:07:40 <flaper87> IT, even
14:07:56 <flaper87> Time is running out even for the implementation so, we should start asap
14:07:59 <flaper87> I've a patch up already
14:08:16 <flaper87> and there are some spec lite bugs that folks can start poking on:
14:08:20 * flaper87 gets the link
14:08:29 <flaper87> #link https://bugs.launchpad.net/glance/+bugs?field.tag=mitaka-new-import-process
14:08:40 <flaper87> I kicked off the work taking the very first one
14:08:55 <flaper87> Do help with the rest if you have some spare cycles
14:09:23 <flaper87> ok, I don't have anything else
14:09:31 <flaper87> rosmaita: ?
14:10:14 <flaper87> I'll take that as a no
14:10:15 <rosmaita> no, just the ovf spec FFE
14:10:16 <flaper87> :D
14:10:32 <flaper87> right, lets talk about it in the next driver's meeting
14:10:35 <rosmaita> (sorry, coffee has not yet got to my fingers this morning)
14:10:38 <rosmaita> ok
14:10:40 <flaper87> we'll have to extend the SFE week a bit
14:10:52 <flaper87> cuz holidays
14:10:57 <flaper87> ok, moving on
14:11:13 <flaper87> (or we can talk about it at the end of the meeting)
14:11:29 <flaper87> #topic Glance virtual mid-cycle in mid M-2
14:11:41 <flaper87> I'll skip the artifact's topic as I need the artifacts folks to be around
14:12:02 <jokke_> flaper87: you skipped the ova as well
14:12:18 <flaper87> Ok, planning to organize a glance virtual mid-cycle. We have the technology, we now need to figure out a time-slot
14:12:22 <flaper87> jokke_: erm, ops? :P
14:12:26 <flaper87> I'll get back to it
14:12:36 <flaper87> man, coffee... moar coffee
14:12:38 <flaper87> MOAR MOAR!
14:13:00 <flaper87> you can tell I just got back from holidays *today*
14:13:02 <flaper87> :P
14:13:12 <flaper87> anyway, I'll send a poll out to the mailing list with time slots
14:13:36 <jokke_> So back to the topic, what's the technology and how long one is in the plans?
14:13:44 <flaper87> Instead of tackling everything during the same day, I've been thinking we could spread 1 day across the week in 1 or 2 hours that could work for most folks
14:14:18 <flaper87> For instance, everyday at ~13UTC for 1 week (this is just an example, not the actual slot)
14:14:30 <jokke_> that's horrible idea ... if we need a day, 2 half days would be ok, but please don't use shotgun to the calendar approach
14:15:22 <rosmaita> hmmm ... i was just going to say, it's worth a try, but maybe jokke_ has a point
14:15:25 <flaper87> right, the problem with that is that some ppl won't be able to attend.
14:15:33 <jokke_> at least for us, it seems that most of the video conferencing stuff does not work from the office so having an hour in middle of day every day for a week means that I won't be participating to most of it
14:15:48 <flaper87> jokke_: mmh
14:16:26 <flaper87> ok, I'll put both options on the email
14:16:30 <rosmaita> maybe 3 2-hour slots across 3 days?
14:16:49 * nikhil sneaks in
14:16:53 <flaper87> rosmaita: yeah, I had something like that in mind.
14:17:33 <flaper87> I do see jokke_'s point. Definitely not going to do a full-day thing
14:17:40 <flaper87> It also depends on the topics
14:17:45 <dhellmann> o/
14:17:45 <nikhil> what are approx dates?
14:17:47 <flaper87> I'm sure artifacts and image import will be there
14:18:08 <flaper87> nikhil: >=Jan16
14:18:17 <nikhil> gotcha
14:19:12 <flaper87> ok, lets see what others have to say in the m-l. TL;DR: I'd like to  increase attendance and ease the choice of days one would attend
14:19:30 <flaper87> if someone is not interested in the import refactor work, that person can simply skip that day
14:20:00 <flaper87> FWIW, I don't think it's such a crazy thing to ask for just 1 week (not the entire week, as rosmaita said, perhaps 3 days)
14:20:10 <flaper87> anyway, anything else ppl would like to add ?
14:20:40 * flaper87 hopes he's still on-line
14:20:48 <rosmaita> sounds good, let's see what the response is
14:20:53 <flaper87> ok
14:21:29 <flaper87> #topic Preparing for next cycle (flaper87)
14:22:00 <flaper87> I'd like to briefly talk about this but I'll likely mention it again next week
14:23:31 <flaper87> Whether or not I'll be running for the PTL position for the N cycle, I'd like to encourage ppl to run for that position. Furthermore, I'd like to help this folks prepare as much as possible for the candidacy and the next cycle.
14:23:59 <flaper87> In addition to that, I'd also like to work with these folks on scheduling long term plans and working out ideas for N that might be impacted by our work in M
14:24:21 <flaper87> So, this is a call for ppl willing/hoping/planning to run for Glance PTL in N to reach out
14:24:34 <flaper87> I don't think there's anything wrong in "revealing" these plans in advance
14:24:41 <flaper87> if anything, I think it just helps the project
14:24:56 <dhellmann> ++, it would be great to have a strong group of folks working together on project leadership
14:25:09 <flaper87> And again, I'd like to encourage ppl to do this and run for Glance PTL
14:25:16 <jokke_> flaper87: any examples in mind for that second (items in N affected by M)?
14:25:35 <nikhil> So, if this is just a call for candidacy I will wait. If the topic is broader and suggests bringing up pain points and potential specs for N I do have a few things.
14:25:39 <flaper87> jokke_: There are follow-up works to do on the import refactor, for instance.
14:25:49 <rosmaita> jokke_: what flaper87 said
14:26:09 <rosmaita> (it was -7 C this morning, my fingers have not yet warmed up)
14:26:29 <flaper87> nikhil: it's not a formal call for candidacy. It's a "if you are planning to run for it, whether you end up doing so or not, I'd love to know that and work with you"
14:27:14 <flaper87> Especially for folks that have never run for a PTL position and would like to give it a try
14:27:23 <nikhil> flaper87: gotcha, and I have a few things brought up to me in my company meetings that people want to see in glance in 2016
14:27:28 <sigmavirus24> ++ for a larger group of candidates in N
14:27:40 <flaper87> One thing I've always said (and I stand by it) is that one of PTL's job is to create new PTLs
14:27:52 <sigmavirus24> It's hard enough choosing between 2 excellent candidates. Having to choose between more than 2 will be excruciating :P
14:27:57 <flaper87> and I'd like to do that, I'd love to see more leadership in Glance
14:28:05 <flaper87> sigmavirus24: :D
14:28:36 <flaper87> ok, I'll bring it again next week when there'll be more ppl around
14:28:43 <flaper87> bring it up*
14:28:51 <jokke_> sigmavirus24: not only selecting between them, but putting them in your priority order :P
14:29:07 <flaper87> nikhil: I'm all ears for those things. Want to add them to next week's agenda?
14:29:19 <nikhil> I like the general idea of anticipant the potential code changes in next cycle and helping with candidacy accordingly
14:29:20 <flaper87> also, happy to discuss them in -glance or the m-l
14:29:37 <flaper87> coolio
14:29:44 <nikhil> meaning helping people vote on candidates etc
14:30:38 <flaper87> I'd also like to help people become candidates. There's lost of potential in this team
14:30:47 <nikhil> lots*
14:30:53 <flaper87> ok, I don't have anything else to say in this topic
14:30:57 <flaper87> nikhil: lol, yeah!
14:31:01 <flaper87> (facepalm)
14:31:36 * flaper87 obviously meant lots, hope that was clear
14:31:40 <flaper87> okidoki
14:31:53 <flaper87> back to the drivers and SFE
14:31:55 <flaper87> #topic Preparing for next cycle (flaper87)
14:31:58 <flaper87> ops
14:32:05 <flaper87> #topic OVF SFE
14:32:43 <flaper87> #link https://review.openstack.org/194868
14:33:09 <flaper87> ok, there's a SFE request for that spec. I know jokke_ has some thoughts about it as he has mentioned them in previous meetings
14:33:36 <mclaren> how does this play with the image import work?
14:33:40 <flaper87> As I mentioned on the spec and the mailing list thread, I believe it's fine to help this spec move forward as it won't impact our priorities and the code is already on gerrit
14:34:13 <mclaren> do we plan to allow importing ovfs using the new import method?
14:34:34 <flaper87> It does not impact the image import work as the implementation is in a task. My comments suggested to have *just* the task implemented
14:34:42 <jokke_> so for some reason I didn't get to vote on the spec, but if that one question of mine gets reasoned/answers I'd be +1 for letting that in at this stage still considering the amount of efforts and patience the OVA folks have demonstrated and that proposal looks really reasonable to me
14:35:00 <mclaren> so there won't be an api to import an ovf until after the new import work is completed?
14:35:12 <nikhil> so, this was basically dragged into SFE as we waiting on import work approval, is that so?
14:35:21 <flaper87> mclaren: yes
14:35:24 <flaper87> nikhil: right
14:35:35 <flaper87> we were waiting to see if it would have impacted the image import work
14:35:38 <mclaren> yes there won't or yes there will?
14:35:45 <rosmaita> jokke_: you were commenting on patch 20 when patch 21 was the latest ... the new interface won't allow you to vote unless you are commenting on the current PS
14:36:01 <jokke_> rosmaita: oh darn ... makes sense
14:36:02 <nikhil> I am okay with it in that case, my stance has been +2 on them limiting this to admin only and task only (non-standard import of ovf)
14:36:27 <jokke_> nikhil: yes I did like that as well
14:36:36 <mclaren> so an admin will be able to import an ovf before the other import spec is completed?
14:36:43 <flaper87> mclaren: The task API, which will be moved under admin only (and the task admin only) there won't be any public API allowing this to be used
14:36:45 <flaper87> The plugging between this and the new image import can be done in N
14:37:22 <mclaren> ok, but it can be exposed to users if the deployer wants
14:37:30 <nikhil> flaper87: oh wait, has this been publicized? this == task becoming admin only
14:37:50 <flaper87> It has but it needs more noise
14:38:00 <nikhil> I am not sure how my team is gunna play with that
14:38:06 <nikhil> adds an item to check
14:38:15 <flaper87> mclaren: well, yeah. TBH, deployers could do anything
14:39:00 <flaper87> The default is that it's not public, it's not consumable publicly and it won't be until the new image import is completed
14:39:07 <mclaren> I'm not talking about 'anything' I'm talking about using APIs
14:39:38 <flaper87> mclaren: I understand what you're saying. What I'm saying is that the default and recommended configs don't suggest that
14:39:50 <mclaren> it's just not clear why we're working on two separate apis to import ovfs at the same time
14:40:03 <flaper87> we are not working on 2 separate APIs
14:40:14 <flaper87> There's a task API that exists already that we're working on deprecating
14:40:15 <mclaren> I thought there was an admin api?
14:40:23 <flaper87> jeeez
14:40:25 <flaper87> So
14:40:29 <flaper87> there's the task API
14:40:32 <flaper87> it exists already
14:40:35 <flaper87> it's in glance
14:40:43 <flaper87> and we want to move it to admin only
14:40:44 <flaper87> it's usable
14:40:48 <sabari> import ovf's would just be alongside conversion/introspection tasks.
14:40:52 <flaper87> and tasks could be plugged to it already
14:40:52 <sabari> and we need to refactor all of them when the new import process lands.
14:40:54 <sabari> right /
14:41:09 <flaper87> The spec does not suggest using the OVF in any of those APIs
14:41:31 <flaper87> The spec proposes implementing the *task* that could be, eventually, run by such API that exists already
14:41:48 <mclaren> there will be a functional test which tests an api which imports an ovf?
14:41:56 <flaper87> What we're doing (and this is in the image import refactor spec) is moving the current, not wanted, task API to admin only
14:42:13 <flaper87> there will be a functional tests that tests the task engine and the OVF import
14:42:15 <flaper87> not the API calling it
14:43:02 <flaper87> The functional tests will test the task *engine*
14:43:05 <flaper87> not the task API
14:43:08 <flaper87> is that clearer ?
14:43:41 <flaper87> At least, that's what I expect to see
14:43:54 * flaper87 re-reads the spec and makes sure it's correctly explained there
14:43:54 <sabari> that's my understanding too
14:44:12 <rosmaita> me three
14:44:51 <jokke_> me four
14:45:04 <flaper87> ok, please, folks. Give the spec another read, ping maliniB on IRC if there's something urgent to address and then vote on the spec
14:45:04 <mclaren> how does that task get exercised / used by an admin who wants to import an ovf?
14:45:21 <flaper87> I'll review it tomorrow again
14:45:44 <flaper87> mclaren: I don't mean to come out harsh ubt I'd really recommend you to take a look at the current task API and how it works.
14:45:57 <sabari> the way this works is no different from conversion/introspection tasks we have currently.
14:45:59 <flaper87> It's the only way ppl can trigger tasks right now until the new image import is done
14:46:19 <nikhil> IIUC, it checks the context.admin value
14:46:23 <flaper87> but again, we're moving away from it and the spec does not recommend using that API, which is what I care about as far as the OVF spec goes
14:46:41 <nikhil> so, I think what mclaren is asking is
14:46:45 <flaper87> Also, the code does this already
14:46:51 <flaper87> IIRC
14:47:16 <nikhil> this task API would be exposed publicly to let the admins use it, but it's a bad experience for the users as it's visible (and may be confusing)
14:48:13 <nikhil> or are we recommending the operators to run another set of glance-api nodes that having admin only access and enable this ovf import?
14:49:08 <flaper87> I don't think we're recommending anything as far as the API goes. The recommendation is (and will be) to simply don't use the task API and hoefully forget it ever existed
14:49:34 <flaper87> If people use it (and they likely will) they'll need to figure out the best way to do so without shooting themselves on the foot
14:50:28 <flaper87> Not sure if I'm explaining myself, tbh.
14:50:34 <flaper87> Anyway, 10mins left
14:50:41 <flaper87> lets discuss this on the spec
14:50:50 <flaper87> #topic Open Discussion
14:51:21 <nikhil> so, there's a plan for "openstack hackathon"
14:51:37 <nikhil> for 3 days in the week of Feb 29
14:52:13 <nikhil> and I have been asked to look into the interest as it's gunna be hosted at IBM
14:52:37 <flaper87> nnniiice!
14:52:39 <flaper87> Where exactly?
14:52:49 <jokke_> is that US thing or all around?
14:52:51 <nikhil> the location is yet TBD and I am hoping to anticipate interest and give preference accordingly
14:53:00 <nikhil> it's a world wide stuff
14:53:10 <nikhil> so not necessarily focused groups
14:53:32 <flaper87> mmh, I think depending on the location I might be able to attend
14:53:33 <nikhil> rather openstack wide devs meeting in a various places across globe
14:53:45 <sabari> nikhil: that's awesome - any links ?
14:54:08 <flaper87> I think the idea is great, just not sure how it works with ppl's travel budget considering there are mid-cycles and summit
14:54:10 <nikhil> I am currently told about potential  China and US locations
14:54:11 <flaper87> still worth a try
14:54:52 <nikhil> no link yet, this is in planning and I am organizing one group
14:55:20 <nikhil> how likely are people to travel to NYC?
14:55:35 <flaper87> that's definitely easier for me and I like NYC :P
14:55:40 <sabari> ah, I just did like last week :D but won't mind :P
14:55:41 <flaper87> not JFK, though
14:55:52 <nikhil> I just need some initial head count to even say that we come up with a formal proposal for ML
14:56:38 <nikhil> btw, there's a UK location too
14:56:47 <sabari> nikhil I can anytime travel to NYC :)
14:56:56 <kragniz> whereabouts in the UK?
14:57:12 * jokke_ woke up on the UK spot as well
14:57:14 <nikhil> kragniz: not sure, edinburg most likely
14:57:27 <kragniz> hmm, okay
14:57:29 <flaper87> nikhil: is that OpenStack hackathon or Glance ?
14:57:40 <nikhil> Openstack hackathon
14:57:51 <flaper87> kk
14:57:53 <nikhil> so you can choose to be at the place and group of your choice
14:58:22 <nikhil> and we should be able to connect virtually as well
14:58:47 <nikhil> that week looks like a GTD week so I like the idea
14:59:23 <nikhil> so I will take a positive note back and come up with more info later
14:59:35 <nikhil> thanks for the initial feedback
14:59:39 <flaper87> ++
14:59:43 <flaper87> thanks for bringing it up
14:59:47 <flaper87> anything else? anyone ?
15:01:06 <flaper87> take that as a no
15:01:09 <flaper87> thanks everyone!
15:01:10 <jokke_> thanks all
15:01:18 <flaper87> Happy to be back to work and hacking with you all!
15:01:22 <sabari> thanks
15:01:23 <flaper87> #endmeeting