14:01:24 <flaper87> #startmeeting Glance Drivers
14:01:25 <openstack> Meeting started Tue Oct 13 14:01:24 2015 UTC and is due to finish in 60 minutes.  The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:26 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:28 <openstack> The meeting name has been set to 'glance_drivers'
14:01:30 <flaper87> Courtesy Glance Drivers' meeting reminder: nikhil_k, flaper87, sigmavirus24, rosmaita, mclaren, dhellmann, jokke_
14:01:41 <flaper87> who's around?
14:02:20 * flaper87 gives some time for ppl to show up
14:02:26 <sigmavirus24> I am
14:02:31 <flaper87> sigmavirus24: w000h00
14:03:05 <flaper87> lets see if we have quorum, otherwise we'll move things to next week
14:03:58 <flaper87> one more minute and I'll end the meeting for lack of participation
14:04:00 * sigmavirus24 nods
14:04:06 <sigmavirus24> nikhil: rosmaita
14:04:15 <sigmavirus24> Is stuart a driver?
14:04:22 <sigmavirus24> (Did that ever happen?)
14:04:22 <flaper87> yeah, I did a courtesy ping
14:04:28 <flaper87> Courtesy Glance Drivers' meeting reminder: nikhil_k, flaper87, sigmavirus24, rosmaita, mclaren, dhellmann, jokke_
14:04:38 <flaper87> sigmavirus24: it happened
14:04:41 <flaper87> :)
14:04:41 <sigmavirus24> I don't see Stuart around
14:04:45 <sigmavirus24> flaper87: I believe it :D
14:04:49 <nikhil> o/
14:04:55 <sigmavirus24> 3
14:05:31 <flaper87> We've enough people to do it
14:05:36 <flaper87> so, lets get this going
14:05:48 <flaper87> #topic Agenda
14:05:51 <flaper87> #link https://etherpad.openstack.org/p/glance-drivers-meeting-agenda
14:05:56 <flaper87> As usual, our agenda is there!
14:06:01 <flaper87> :D
14:06:10 <flaper87> #topic Remove Glance downgrades: https://review.openstack.org/#/c/232364/2 (flaper87)
14:06:22 <flaper87> nikhil: glad you're here since that one requires your attention
14:06:24 <flaper87> :)
14:06:36 <flaper87> I commented on it but it could use a review from you
14:06:41 <nikhil> oh, I think I commented and forgot to come back ?
14:06:58 <flaper87> nikhil: probably, yup :)
14:07:05 <nikhil> oh, right I commeneted on the BP and no the spec
14:07:09 <nikhil> ack
14:07:18 <rosmaita> (sorry, was sitting in #glance by mistake)
14:07:27 <flaper87> rosmaita: lol :D
14:07:30 <flaper87> welcome
14:07:30 <nikhil> any tl'dr', attn, speecifics?
14:07:45 <flaper87> no other than it seems straightforward to me
14:07:56 <nikhil> gotcha
14:08:00 <flaper87> but I remember you having concerns about it so your vote is important
14:08:19 <flaper87> ok, moving on
14:08:21 <nikhil> Thanks flaper87
14:08:25 <flaper87> #topic Glance spec freeze period (flaper87)
14:08:35 <flaper87> I think we should have a spec freeze. One thing that I definitely think should not happen again is to get to the second milestone with specs still waiting to be discussed. Therefore, I think we could have a freeze at mid M-2 (that's between M-1 release and M-2 release) on week from November 28th and December 1st.
14:08:38 <flaper87> We can have a short Spec Freeze Exception (SFE) that closes at the end of the week from the freeze 4-8.
14:08:40 <flaper87> This should give enough time for late specs to be implemented before the feature freeze and it'll allow the rest of the team to focus on reviewing code.
14:08:41 <flaper87> thoughts?
14:09:31 <nikhil> DOes this mean that specs are allowed after the freeze ?
14:09:32 <nikhil> it
14:09:38 <nikhil> it's a soft-freeze?
14:09:53 <flaper87> Specs coming after the freeze should target the next release
14:10:06 <flaper87> I mean, we could go soft with it
14:10:11 <flaper87> As in, evaluate every case
14:10:12 <rosmaita> what about non-approved specs under discussion before the freeze?
14:10:19 <flaper87> but no big features
14:10:25 <flaper87> rosmaita: ideally, we shouldn't get there
14:10:31 <rosmaita> true
14:10:39 <sigmavirus24> flaper87: I kind of like this
14:10:40 <rosmaita> i think it's a good idea
14:10:52 <nikhil> hmm, so feeze at mid M2 means basically specs have about 8weeks after the summit?
14:10:53 <rosmaita> besides, if it doesn't work, we can always change
14:10:53 <flaper87> FWIW, nova's freeze is even before the first milestone
14:11:06 <rosmaita> but let's start with announcing a hard deadline and see what happens
14:11:22 <flaper87> I think that's enough time, tbh.
14:11:32 <flaper87> we can be more flexible because we don't have that many specs
14:11:59 <flaper87> but mid M-2 is quite late to have new specs proposed or even other specs approved
14:12:13 <flaper87> ideally, when M-2 is out, we should have patches up for most of them
14:12:20 <flaper87> not saying merged but at least under review
14:12:30 <nikhil> I don;t mind though if we have possibility of some of the more complicated specs then I think we may need more time
14:12:46 <nikhil> like the upload retry one
14:12:55 <flaper87> #vote Mitaka spec freeze on  November 28th and December 1st yes,no,abstain
14:13:02 * flaper87 forgot how to use vote
14:13:05 <flaper87> #help vote
14:13:12 <flaper87> stupid bot
14:13:15 <nikhil> SO, best to raise this at the summit and let people know that since API work is #1 priority it's less likely that their spec/patches are possible
14:13:31 <sigmavirus24> At least for mitaka, right?
14:13:32 <flaper87> #startvote Mitaka spec freeze on  November 28th and December 1st yes,no,abstain
14:13:32 <openstack> Unable to parse vote topic and options.
14:13:37 <flaper87> #startvote Mitaka spec freeze on  November 28th and December 1st
14:13:38 <openstack> Unable to parse vote topic and options.
14:13:49 <sigmavirus24> flaper87: I think you need spaces between options
14:13:53 <sigmavirus24> And maybe a ?
14:13:55 <nikhil> #startvote Mitaka spec freeze on November 28th and December 1st yes,no,abstain
14:13:56 <openstack> Only the meeting chair may start a vote.
14:14:09 <rosmaita> we're getting closer!
14:14:13 <nikhil> ha
14:14:15 <flaper87> I tried w/ commas and it didn't work
14:14:18 <flaper87> :D
14:14:19 <sigmavirus24> '#startvote Should we have a mitaka spec freeze on November 28th and December 1st? Yes, no, abstain'
14:14:24 <flaper87> #startvote Mitaka spec freeze on  November 28th and December 1st? yes,no,abstain
14:14:25 <openstack> Begin voting on: Mitaka spec freeze on  November 28th and December 1st? Valid vote options are yes, no, abstain.
14:14:26 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
14:14:35 <rosmaita> way to go sigmavirus24 !
14:14:35 <flaper87> ah, the `?` is the key
14:14:41 <sigmavirus24> #vote Yes
14:14:47 <flaper87> #vote yes
14:14:50 <rosmaita> #vote yes
14:14:50 <sigmavirus24> flaper87: yeah, thank debian =P
14:15:06 <nikhil> I am fine by either.
14:15:24 <nikhil> So you can take it they way it works and it's majority already!
14:15:35 <flaper87> nikhil: you can abstain :D
14:15:49 <nikhil> oh, so the abstain is me abstaining :P ok!
14:15:55 <nikhil> #vote abstain
14:15:57 <flaper87> I'll announce it in the mailing list anyway
14:16:01 <flaper87> #endvote
14:16:02 <openstack> Voted on "Mitaka spec freeze on  November 28th and December 1st?" Results are
14:16:03 <openstack> yes (3): sigmavirus24, rosmaita, flaper87
14:16:04 <openstack> abstain (1): nikhil
14:16:10 <flaper87> ok, moving on
14:16:19 <flaper87> #topic Spec lite (or no spec) (continued) http://eavesdrop.openstack.org/meetings/glance_drivers/2015/glance_drivers.2015-10-06-14.01.log.html (flaper87)
14:16:27 <flaper87> Last week we discussed a bit this process and I proposed doing something similar to what Nova and Keystone are doing. That is, have an approval process for small features where no spec is required. This would require a launchpad blueprint to be filed and the drivers team to review it.
14:16:30 <flaper87> After looking a bit more into this, I still think it's a good idea but Nikhil raised a good concern about the discoverability of launchpad blueprints that I think is important as well. Therefore, I would like to propose using whishlist bugs for this.
14:16:32 <flaper87> The assumption is that if a "small feature" requires a blueprint and therefore it requires a priority, impact, more formal approval, then it might as well have a spec. Bugs are easy to track, point to, comment on and the description can be clear and concise, which is all we need.
14:16:34 <flaper87> These bugs can be targeted, tracked, linked on reviews and they can also be marked as `Invalid` if we don't think that request is a good fit for Glance. One other reasoning behind this is that most folks are already familiar with bugs *and* the cases we've had so far have been features that were filed as bugs but then ended up being a feature request.
14:16:36 <flaper87> TL;DR: Lets use whishlist bugs, discuss them as specs and agree if whishlist is enough or if a full spec is needed.
14:16:38 <flaper87> thoughts?
14:17:26 <rosmaita> i like it
14:17:37 <nikhil> That sounds good to me. Pretty straightforward process and easy to deal with.
14:17:57 <nikhil> THough, how do we prevent other wishlist bugs from getting marked as spec-lite?
14:18:01 <nikhil> triage prolly?
14:18:21 <flaper87> yeah, we'll have to do some serious triage! I'm planning to do that next week or after the summit
14:18:32 <nikhil> What I was implying is -- should we have a official tag called "mitaka-lite-specs" for filtering?
14:18:54 <flaper87> ah, sorry! missed it
14:19:08 <flaper87> I think we could use the tag, yes. Although, I don't think we need the release name there
14:19:20 <nikhil> sounds good.
14:19:23 <flaper87> If it needs to be targeted to a release, then it probably should have a spec
14:19:40 <nikhil> I suggested release name for the sake of indexing .
14:19:52 <flaper87> Also, bringing those to meetings (glance/drivers) is totally fine
14:19:59 <nikhil> cool!
14:20:14 <flaper87> but then we'll have to re-tag all of them when new releases start, I'd guess
14:20:18 <nikhil> it should also help with bug triage as a side-effect!
14:20:20 <flaper87> for the sake of formality, lets vote
14:20:26 <nikhil> heh
14:20:42 <flaper87> #startvote Use whishlist bugs as a way to propose lite-specs? Yes, No, Abstain
14:20:43 <openstack> Begin voting on: Use whishlist bugs as a way to propose lite-specs? Valid vote options are Yes, No, Abstain.
14:20:44 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
14:20:46 <flaper87> #vote Yes
14:20:50 <nikhil> #vote yes
14:20:54 <flaper87> sigmavirus24: ^
14:21:06 <flaper87> rosmaita: ^
14:21:11 <sigmavirus24> #vote no
14:21:17 <flaper87> :(
14:21:19 <rosmaita> sorry, saw something shiny
14:21:30 <sigmavirus24> Well
14:21:39 <sigmavirus24> I don't know. I think my vote is maybe
14:21:39 <rosmaita> sigmavirus24: elaborate, please
14:21:49 <flaper87> sigmavirus24: please, elaborate
14:21:52 <flaper87> :)
14:22:07 <sigmavirus24> So, I think we'll need to communicate outwards that wishlist in glance is to only be set by drivers
14:22:19 <sigmavirus24> As in, we'll want to be sure we "approve" something as a wishlist/lite-spec that way
14:22:30 <sigmavirus24> And we currently have bugs that are marked as wishlist bugs
14:22:35 <sigmavirus24> And we'll be overloading the term from general openstack
14:22:50 <sigmavirus24> So I feel like we're going to confuse people and that it might not be a great idea but it *might* work
14:22:51 <flaper87> sigmavirus24: yup, that's the "serious triage session" I mentioned above
14:23:04 <sigmavirus24> Yeah, I was a bit distracted looking into a possible glance bug for a teammate
14:23:05 <sigmavirus24> =P
14:23:18 <sigmavirus24> Yeah, I'm on the fence
14:23:19 <sigmavirus24> #vote abstain
14:23:25 <flaper87> imho, there's only one way to know if it'll work.
14:23:30 <sigmavirus24> Try it
14:23:31 <sigmavirus24> I know
14:23:35 <flaper87> I got some feedback from Keystone
14:23:48 <flaper87> and it seems that they have a mix of LP blueprints and whishlist
14:23:51 <flaper87> I'd rather just have 1
14:23:55 <rosmaita> i like nikhil's idea of a special tag to separate the regular wishlist from the spec-lite wishlist
14:24:08 <flaper87> rosmaita: yes!
14:24:30 <flaper87> The rest of the whishlist bugs we have need to be triaged as well
14:24:44 <flaper87> if they require spec then lets mark them as invalid, if they don't then lets triage them
14:25:00 <rosmaita> i think it's worth a try
14:25:06 <flaper87> Leaving the current status aside, lets think about how this could work
14:25:18 <flaper87> I'm happy to clean up whatever mess there may be
14:25:27 <flaper87> rosmaita: you haven't voted
14:25:47 <rosmaita> oops
14:25:51 <rosmaita> #vote yes
14:26:09 <flaper87> coolio, I'll announce this on the mailing list as well
14:26:10 <rosmaita> i am very agreeable today
14:26:10 <nikhil> sigmavirus24: so, do you prefer to use just BP instead?
14:26:18 <flaper87> #endvote
14:26:19 <openstack> Voted on "Use whishlist bugs as a way to propose lite-specs?" Results are
14:26:20 <openstack> Yes (3): nikhil, rosmaita, flaper87
14:26:21 <openstack> Abstain (1): sigmavirus24
14:26:42 <sigmavirus24> nikhil: I think that might be less of an overloading
14:26:49 <sigmavirus24> but I don't have very strong feelings on it
14:27:00 <sigmavirus24> Mostly because I doubt I'll get any more time in Mitaka to work on Glance than I did in Liberty
14:27:00 <flaper87> I'd like to move on real quick onto the next topic and I'll get back to this one
14:27:06 <sigmavirus24> flaper87: fine by me
14:27:09 <flaper87> #topic Image import reloaded https://review.openstack.org/#/c/232371/ (flaper87) (time permitting)
14:27:26 <flaper87> rosmaita: there seems to be quite some discussion going!
14:27:29 <flaper87> that's awesome
14:27:44 <flaper87> I wanted to know what's your plan for the next 2 weeks on it
14:27:52 <flaper87> do you plan to have an update on the spec before the summit?
14:27:56 <rosmaita> yes
14:28:00 <flaper87> or just keep the current, commented, version
14:28:03 <flaper87> ?
14:28:05 <flaper87> aewsome!
14:28:11 <rosmaita> i think we should come close to settling the issue of /tasks or /images
14:28:22 <rosmaita> or at least have a good understanding of the pros/cons
14:28:25 <flaper87> agreed
14:28:36 <rosmaita> so my plan is to give a few more days, then upload a revision
14:28:56 <rosmaita> if there's something i disagree with, i will include it as a "note" in RST
14:28:57 <nikhil> how about re-upload on mon, wed (only)?
14:28:59 <flaper87> What do you think about discussing this on the mailing list? The /tasks vs /images part, I mean
14:29:05 <nikhil> re-upload == new PS
14:29:07 <rosmaita> so it will still be there
14:29:14 <nikhil> I am saying as some comments can get lost :)
14:29:28 <flaper87> do you think discussing the endpoints on the spec would be enough?
14:29:34 <flaper87> I'm fine either way, tbh.
14:29:41 <rosmaita> no, it probably needs a ML discussion
14:29:42 <flaper87> I'd probably prefer to keep the discussion on the spec now
14:29:50 <flaper87> but I know it'll take time
14:29:53 <rosmaita> what i would like to do is revise and see what people thing
14:29:55 <flaper87> and a good email thread might help
14:29:56 <rosmaita> *think
14:30:06 <flaper87> sounds perfect!
14:30:17 <rosmaita> let's talk more on Thurs at the glance meeting
14:30:23 <nikhil> ++
14:30:24 <flaper87> It'd be great to get that thread started before the end of this week, if you have time
14:30:34 <nikhil> small and concise ML "threads"
14:30:34 <flaper87> rosmaita: ++
14:30:41 <rosmaita> i will try
14:30:44 <nikhil> if it works
14:30:54 <flaper87> rosmaita: mind adding it to the glance meeting agenda?
14:31:07 <flaper87> rosmaita: if there's anything we can help with, please, do let us know
14:31:16 <flaper87> even if it's just sending an email out
14:31:24 <flaper87> I know you and Stuart got the spec covered
14:31:26 <rosmaita> ok, will do
14:31:36 <flaper87> sweet
14:31:40 <flaper87> thanks for all your effort there!
14:31:56 <flaper87> we ran out of time!
14:31:56 <rosmaita> np, i really want to see this work out
14:32:02 <flaper87> thank you all
14:32:04 <flaper87> #endmeeting