15:00:20 <nikhil_k> #startmeeting Glance
15:00:22 <openstack> Meeting started Thu Feb  5 15:00:20 2015 UTC and is due to finish in 60 minutes.  The chair is nikhil_k. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:25 <jokke_> o/
15:00:25 <openstack> The meeting name has been set to 'glance'
15:00:28 <kragniz> o/
15:00:34 <stevelle> o/
15:00:40 <rosmaita> o/
15:00:40 <cpallares> o/
15:00:48 <nikhil_k> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda
15:01:15 <sigmavirus24> o/
15:01:19 <pkoniszewski> o/
15:01:30 <nikhil_k> Ok, Let's get the party started ;)
15:01:36 <nikhil_k> #topic k2
15:01:37 <mfedosin> o/
15:01:44 <nikhil_k> #link https://launchpad.net/glance/+milestone/kilo-2
15:02:09 <ativelkov> o/
15:02:13 <sigmavirus24> Is k-2 finished yet? I saw a lot of activity on cinder
15:02:24 <nikhil_k> A couple of reviews still lingering there. We'd probably see them through in the next 1.5 hours.
15:02:26 <sigmavirus24> Or is glance delaying k-2
15:02:39 <nikhil_k> Glance is being stalled for ativelkov's patch
15:02:59 <nikhil_k> any concerns?
15:03:17 <nikhil_k> Any other patches/bugs/info that we may have missed ?
15:03:39 <nikhil_k> Does everyone think we've made a cleaner release than the previous?
15:04:08 <kragniz> nikhil_k: it's been smoother, I think
15:04:17 <kragniz> (once the last patches get merged)
15:04:24 <nikhil_k> :)
15:04:24 <ativelkov> We'll see in a several days if it's cleaner: the previous one looked good at the milestone, but turned out to be broken in several days )
15:04:35 <jokke_> I think bit earlier shout for attention would not have hurt but other than that I think we're on better shape than at k1
15:04:56 <nikhil_k> jokke_: ah, thanks for bringing that up
15:04:58 <pennerc> o/
15:05:11 <nikhil_k> How can we do a better job at getting people's attn?
15:05:27 <nikhil_k> Does a announcement in the meeting not help at all?
15:05:35 <sigmavirus24> nikhil_k: It helps but focus tends to shift
15:05:36 <jokke_> nikhil_k: that e-mail was good
15:05:43 <sigmavirus24> I think we might need to email the list a bit more periodically
15:06:06 <kragniz> nikhil_k: shouting louder :P
15:06:06 <nikhil_k> jokke_: sigmavirus24: ok, thanks for the pointer.
15:06:26 <sigmavirus24> kragniz: frequency is sometimes more important than volume =P
15:06:45 <nikhil_k> It seems that we've more contributors who visit the ML often than previous set who had a bit more IRC focus
15:06:45 <kragniz> sigmavirus24: true
15:07:27 <nikhil_k> So, the change in dynamics is being felt a bit late. I think we can revisit this topic at the dawn of k3 work.
15:07:42 <nikhil_k> If nothing else..
15:07:56 <nikhil_k> #topic New Glance cores
15:07:59 <jokke_> nikhil_k: I'm not that convinced about the mailing list either, but when you're e-mail pops directly to my inbox, you got my attention :P
15:08:07 <cpallares> sigmavirus24: +1
15:08:08 <nikhil_k> jokke_: :)
15:08:32 <sigmavirus24> nikhil_k: perhaps we need a list of people who review things often and their email addresses
15:08:36 <sigmavirus24> You can ping them directly
15:08:44 * sigmavirus24 is disappointed he wouldn't make that cut =P
15:08:48 <jokke_> perhaps we need glance mailing list :P
15:08:55 <sigmavirus24> jokke_: no that's preposterous
15:08:58 <sigmavirus24> :D
15:08:59 <nikhil_k> sigmavirus24: bingo. I do that already
15:09:09 <nikhil_k> will add you to the list. you called for it
15:10:03 <nikhil_k> #startvote Do we need more (active) Glance cores?
15:10:04 <openstack> Begin voting on: Do we need more (active) Glance cores? Valid vote options are Yes, No.
15:10:05 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
15:10:21 * sigmavirus24 wants to vote maybe
15:10:29 <nikhil_k> (all are welcome)
15:10:37 <nikhil_k> #vote Yes
15:10:44 <kragniz> #vote Yes
15:10:47 <pkoniszewski> #vote Yes
15:10:50 <ativelkov> Good reviews > more reviews
15:10:56 <sigmavirus24> == ativelkov
15:10:58 <ativelkov> #vote Yes
15:11:01 <jokke_> ativelkov: ++
15:11:05 <nikhil_k> == ativelkov
15:11:06 <cpallares> #vote yes
15:11:07 <mfedosin> #vote Yes
15:11:09 <sigmavirus24> I'm not sure how to vote
15:11:23 <sigmavirus24> I feel like we have a good deal of mostly active cores already
15:11:25 <nikhil_k> s/how/what/ ?
15:11:29 <jokke_> #vote maybe
15:11:31 <sigmavirus24> nikhil_k: yes
15:11:34 <sigmavirus24> jokke_: that's how I feel
15:11:52 <mclaren> this is for the second proposed batch, right?
15:12:02 * sigmavirus24 waves to mclaren
15:12:04 <nikhil_k> yes, second for kilo
15:12:09 <mclaren> ok, thanks
15:12:13 <mclaren> #vote yes
15:12:30 <stevelle> #vote yes
15:12:47 <nikhil_k> (just one may-be-no so far)
15:12:51 <rosmaita> #vote yes
15:13:32 * jokke_ is bit concerned about nikhil_k's shout to slow down and same time adding people to merge stuff in
15:13:38 <nikhil_k> ok, that'd help. Please speak now if you've serious concerns.
15:14:04 <nikhil_k> jokke_: good point. Was about to bring it up in the next topic in list
15:14:20 <nikhil_k> #endvote
15:14:20 <openstack> Voted on "Do we need more (active) Glance cores?" Results are
15:14:35 <nikhil_k> ok, that'd help. Please speak now if you've serious concerns.
15:14:52 <jokke_> nikhil_k: would you mind to clear which one we need?
15:15:12 <nikhil_k> I think we need a _balance_
15:15:14 <jokke_> do we want to move and get stuff done or do we want to slow down?
15:15:58 <mclaren> I think its about getting the right stuff done
15:16:03 <nikhil_k> In k1 we merged stuff without being prepared of the outcome. In k2 we slowed down and became kinda-blockers for some.
15:16:46 <nikhil_k> Having learned lessons from both the milestones, I feel like we'r prepared for the k3 battle ;)
15:16:49 <jokke_> mclaren: ++
15:16:51 <nikhil_k> learnt*
15:17:18 <sigmavirus24> Yeah, I think there were trivial changes that were good to have but were blocked by the slow-down mentality. Which was net positive regardless
15:17:55 <nikhil_k> #topic oslo related reviews
15:18:00 <nikhil_k> Continuing...
15:18:14 <stevelle> Quality software and slowing down are not contrary to putting more eyes on things either
15:18:59 <nikhil_k> So, I've been asked by the oslo team as to why their sync changes were being nit
15:19:29 <nikhil_k> Although, that being a good thing, we need more bandwidth to babysit such important and cross-project changes
15:20:19 <nikhil_k> And they would be happy if we'd more active cores
15:20:31 <sigmavirus24> I'm responsible for one of the nits and it's mostly out of frustration due to the lack of documentation on some of their projects that I've run into
15:20:44 <jokke_> sigmavirus24: ++
15:21:03 <nikhil_k> sigmavirus24: heh, ok :)
15:21:09 <sigmavirus24> I think asking them to drop links to their docs into the relevant area they're modifying is harmless but perhaps they're not used to being asked for documentation
15:21:25 <jokke_> another thing I'd like to add on top of that is consistency ... those changes rarely fits to the code around them on first rounds
15:21:30 <sigmavirus24> That said, I've been sending patches to improve some of that documentation so they can't call it some random person complaining
15:21:38 <rosmaita> a cross-project effort should put a priority on docs
15:22:19 <nikhil_k> jokke_: sigmavirus24 : yes, I think we need to bring such issues in the Cross project meetings
15:22:33 <nikhil_k> rosmaita: +100
15:22:45 <nikhil_k> It's so hard to keep track
15:22:59 <nikhil_k> anyways, in the interest of time - I'd be available after the meeting on our regular channel for further discussion on this
15:23:12 <sigmavirus24> So the one that I nitted was the oslo.context update. That said, I'm happy to submit a dependent patch (if it hasn't already been merged
15:23:23 <nikhil_k> sigmavirus24: please do
15:23:26 <nikhil_k> #toic Meetup recap
15:23:26 <kragniz> sigmavirus24: I think that's merged
15:23:39 <nikhil_k> #topic Meetup recap
15:23:49 <jokke_> We had lenghty discussion about oslo with Flavio at Saturday and the unfortunate reality is that oslo is heavily killing the readability of the code for laziness and "to reduce the overhead"
15:23:49 <nikhil_k> Not sure who added that?
15:24:24 <rosmaita> jokke_: tell us more
15:24:30 <jokke_> I did ... would just like to have short recap as I was unfortunately bouncing back and forth on the phone conf (after I finally found out about it)
15:25:05 <sigmavirus24> stevelle: sent a recap to our (Rackspace Private Cloud's) team internal list. Would it be helpful if something similar was sent to the ML jokke_ ?
15:25:15 <nikhil_k> If there are no takers for this topic, we can move one..
15:25:16 <kragniz> sigmavirus24: that would be great
15:25:34 * sigmavirus24 seems to have volunteered stevelle
15:25:42 <rosmaita> sigmavirus24: would be good to all be on the same page about what went on
15:25:50 <sigmavirus24> == rosamaita
15:26:08 <jokke_> rosmaita: one thing is the lack of documentation and the way for example moving all named parameters to **kwargs just to avoid maintenance of the function calls
15:26:25 <nikhil_k> ok, moving on..
15:26:26 <sigmavirus24> I'll ask stevelle to send a similar message to the ML
15:26:28 <sigmavirus24> Let's move on
15:26:29 <rosmaita> jokke_: that looks like an antipattern
15:26:41 <nikhil_k> jokke_: rosmaita stevelle sigmavirus24 : can we please continue in the open discussion?
15:26:52 <stevelle> agreed
15:26:54 <rosmaita> sorry
15:26:59 <jokke_> rosmaita: also the unfortunate reality oslo being not the place to everyone contribute but the excuse everyone using not to
15:27:05 <nikhil_k> #topic Metadef extension
15:27:07 <jokke_> nikhil_k: sure, sry
15:27:22 <nikhil_k> pkoniszewski: was that you or Travis ?
15:27:28 <pkoniszewski> it's me
15:27:46 * nikhil_k hands the mic to pkoniszewski
15:27:59 <pkoniszewski> currently api consumer has no idea how to combine multiple values with a single key (im talking about metadefs)
15:28:18 <pkoniszewski> this is small change to expose which scheduler operators are suitable with given property/object
15:28:40 <pkoniszewski> I kindly ask for some reviews, there is also dependency in Horizon and probably i'll beg for exception there
15:28:50 <pkoniszewski> (bp + code are ready)
15:28:58 <sigmavirus24> pkoniszewski: links?
15:29:13 <pkoniszewski> https://blueprints.launchpad.net/glance/+spec/metadata-multivalue-operators-support
15:29:35 <pkoniszewski> and current speec: https://review.openstack.org/#/c/147264/5/specs/kilo/metadata-multivalue-operators-support.rst
15:29:36 <sigmavirus24> ty
15:29:48 <nikhil_k> pkoniszewski: ok, just started this https://etherpad.openstack.org/p/kilo-glance-k3
15:29:56 <sigmavirus24> oh this
15:29:58 <sigmavirus24> I remember this
15:29:58 <nikhil_k> please add some context and links to that
15:30:05 <pkoniszewski> sure!
15:30:19 * sigmavirus24 wonders if he should make the obvious Jurassic Park reference
15:30:35 <pkoniszewski> that's all for me :)
15:30:41 <rosmaita> sigmavirus24: please do
15:30:50 <sigmavirus24> "I know this! This is Unix!"
15:30:55 <sigmavirus24> (for rosmaita)
15:31:02 <rosmaita> ty
15:31:21 <nikhil_k> pkoniszewski: thanks!
15:31:23 <nikhil_k> #topic python-glanceclient and glance_store releases
15:31:38 * ativelkov thinks he should rewatch the JP movie, as he doesn't remember that
15:31:45 <nikhil_k> We'd plan to do these sometime next week. May be Mon or Tues
15:32:10 <kragniz> we still have about four patches to merge in the client
15:32:16 <nikhil_k> Are there any objections?
15:32:21 <nikhil_k> ok
15:32:38 <nikhil_k> kragniz: Let's get some reviews on them then...
15:32:47 <nikhil_k> another reason for more cores
15:33:32 <nikhil_k> kragniz: Do you have the list by chance?
15:33:59 <jokke_> #link https://bugs.launchpad.net/python-glanceclient/+milestone/v0.16.0
15:34:08 <jokke_> that's the one I assume?
15:34:08 <sigmavirus24> thanks jokke_
15:34:13 <sigmavirus24> jokke_: I believe so
15:34:31 <sigmavirus24> Most of those have +2s on them iirc
15:34:32 <nikhil_k> jokke_: I don't think reviewers are going and finding links from the milestone pages
15:34:35 <jokke_> there's quite a bunch of stuff ... we need to hit it if we want to get the client out there
15:34:40 <kragniz> jokke_: yes, I was going off the list on trello
15:34:51 <nikhil_k> otherwise, my 3 emails and multiple announcements would have worked for k2
15:35:01 <kragniz> which is a smaller list
15:35:15 <nikhil_k> kragniz: yes please
15:35:24 <jokke_> then we probably need to clean up that milestone as well?
15:36:07 <nikhil_k> jokke_: yeah, may be. good point :)
15:36:08 * TravT is scrambling to check out of hotel.  apologizes for mostly being absent today
15:37:00 <nikhil_k> Anyways, let's try to have at least one release early next week
15:37:09 <nikhil_k> #topic Glance upgradeability - versioned objects
15:37:29 <nikhil_k> pkoniszewski: please take over the mic
15:37:51 <pkoniszewski> so once I mentioned versioned objects to support smooth upgrades in Glance
15:38:13 <pkoniszewski> spec to show the idea is ready
15:38:27 <pkoniszewski> if anyone is interested how I see this in Glance you can find it there - https://review.openstack.org/#/c/151194/1/specs/kilo/glance-upgrades.rst
15:39:08 <nikhil_k> pkoniszewski: Let's sync up offline late next week to set a priority on this
15:39:13 <pkoniszewski> sure
15:39:22 <nikhil_k> our channel is logged now, anyways
15:39:28 <pkoniszewski> just want to add that Heat will test it in K-release
15:39:33 <mclaren> random thought: I'm not sure what we support in terms of api -> registry upgrade? eg juno api -> kilo registry
15:40:21 <jokke_> mclaren: good point, would not be bad idea to ensure that grenade tests that?
15:40:46 <pkoniszewski> it will, Heat is working on this
15:41:05 <mclaren> sounds good
15:41:08 <pkoniszewski> last weekend they made new project in oslo, so it will accelerate now
15:41:08 <nikhil_k> .awesome.
15:41:40 <nikhil_k> #topic k-3
15:42:08 <nikhil_k> So, let's club this and mini-summit topic (proposed earlier)
15:42:21 <nikhil_k> We'd some productive discussions at the mini-summit
15:42:56 <nikhil_k> The sad thing is that we will not have enough bandwidth to support all the proposal given our current (core) velocity.
15:43:40 <nikhil_k> So, the plan for k3 should be more focused and prioritized from the get go.
15:44:09 <sigmavirus24> nikhil_k: preferrably on smaller and achievable goals
15:44:15 <nikhil_k> If you do not have core reviewers assigned to the spec now (don't add me to every one), it will be hard to get them in soon
15:44:30 <nikhil_k> sigmavirus24: +1
15:45:40 <nikhil_k> Given two major features proposed, we need enough eyes on them. Besides, there need to be bunch of stability/perf upgrades like the next two in the agenda.
15:46:29 <nikhil_k> I recommend all big features to have more than one representative not in their them with whom they sync up weekly.
15:47:01 <ativelkov> not in their..?
15:47:19 <nikhil_k> Since we get to sync up only once a week here, it would be great to organize a small sync up on #openstack-glance periodically
15:47:29 <nikhil_k> err,
15:47:37 <sigmavirus24> nikhil_k: I kind of like this idea
15:47:42 <nikhil_k> not in their - group/team/org
15:48:00 <sigmavirus24> I think some people will find it challenging because of timezones, but those who can manage it would probably benefit from it
15:48:22 <kragniz> one thing we could do is have a whiteboard with the status of certain key reviews
15:48:34 <nikhil_k> sigmavirus24: right, tz is the major hurdle
15:48:48 <nikhil_k> kragniz: ++
15:48:55 <sigmavirus24> nikhil_k: idea: migrate all glance devs to timezones within +/- 2 of each other.
15:49:01 <nikhil_k> How about we start with https://etherpad.openstack.org/p/kilo-glance-k3
15:49:04 <nikhil_k> ?
15:49:09 <kragniz> (taking inspiration from ironic's whiteboard https://etherpad.openstack.org/p/IronicWhiteBoard)
15:49:14 <nikhil_k> and them evolve into something scalable
15:49:17 <nikhil_k> then*
15:49:19 <jokke_> sigmavirus24: you're all welcome to GMT+-2
15:49:27 <jokke_> ;)
15:49:29 <sigmavirus24> :D jokke_
15:49:39 <ativelkov> +1 on whiteboard
15:49:59 <nikhil_k> kragniz: nice link!
15:50:03 <nikhil_k> #link https://etherpad.openstack.org/p/IronicWhiteBoard
15:50:21 <kragniz> having something like that would be pretty handy
15:50:37 <kragniz> they have a link to it in the topic of their IRC channel
15:50:39 <nikhil_k> ok, we seem to have something to start with..
15:50:48 <nikhil_k> s/with/from/g
15:51:07 <nikhil_k> #topic log request id mappings from cinder
15:51:17 <nikhil_k> rajesht: o/
15:51:28 <rajesht> We have submitted the patch (https://review.openstack.org/#/c/152820/) for nova project.
15:51:36 <rajesht> Similar changes we have to do in glance project as well.
15:51:44 <rajesht> For that I have already submitted the blueprint (https://blueprints.launchpad.net/glance/+spec/log-request-id-mappings-from-cinder)
15:51:44 <rajesht> and in two days time, will submit spec as well.
15:51:45 <jokke_> Based on recent discussion on the mailing list, I have strong feeling that this will not happen in Kilo
15:51:49 <rajesht> Need your opinion about the blueprint ?
15:52:45 <nikhil_k> rajesht: agree with jokke_  on this
15:52:47 <kragniz> rajesht: this is happening in L on nova, right?
15:52:57 <rajesht> yes
15:53:02 <rajesht> but not sure
15:53:10 <rajesht> we are trying to make it in kilo
15:53:18 <sigmavirus24> I thought nova was frozen for K
15:53:39 <nikhil_k> rajesht: ok, let's see. I've marked it in https://etherpad.openstack.org/p/kilo-glance-k3 . Please ping me if you do not get any response in a week.
15:53:40 <rajesht> yes
15:53:54 <rajesht> sure
15:54:06 <abhishekk> nikhil_k: We will be up wiht the spec in day or two for this
15:54:18 <nikhil_k> sg. thanks abhishekk
15:54:20 <nikhil_k> #topic Review for glance-specs
15:54:20 <jokke_> #link http://www.mail-archive.com/openstack-dev@lists.openstack.org/msg45149.html
15:54:39 <nikhil_k> abhishekk: so, mclaren had a nice demo during the minisummit
15:54:55 <nikhil_k> I think we just need to sync up there
15:55:04 <abhishekk> nikhil_k: yes
15:55:22 <abhishekk> I think code is in good shape
15:55:32 <abhishekk> and specs needs review
15:56:14 <abhishekk> It will be great if I get some early feedback on specs and patch
15:56:26 <sigmavirus24> abhishekk: is that the SIGHUP work?
15:56:33 <nikhil_k> ok, sure. I guess drivers team will review it soon as you pointed it out here.
15:56:38 <nikhil_k> sigmavirus24: yes
15:56:47 <abhishekk> sigmavirus24: yes
15:57:04 <nikhil_k> #topic Open Discussion
15:57:07 <TravT> is there anybody in europe (I'm looking at you kragniz) that I could refer to Julie Pichon on the Horizon team to help with v2 horizon / glance work.
15:57:29 <TravT> I'm working with her on it, but I've been traveling the last two weeks
15:57:30 <nikhil_k> sigmavirus24: rosmaita stevelle jokke_ : the oslo conversation again..
15:57:50 <sigmavirus24> so jokke_'s link just now seems relevant to rajesht's spe
15:57:51 <TravT> she just has a couple questions that I have had time to dig into
15:57:52 <sigmavirus24> *spec
15:57:53 <kragniz> TravT: probably
15:58:01 <nikhil_k> Do we have some action items or agena to be proposed to the Cross project meeting?
15:58:09 <jokke_> TravT: yeah we had quick chat about the topic while driving from the airport with kragniz and Julie at Sun ;)
15:58:26 <TravT> ahh, you all went to FOSDEM
15:58:34 <kragniz> indeed
15:58:36 <TravT> great!
15:58:55 <jokke_> nikhil_k: we're quite short on time, should we sync up on this offline rather than rushing through?
15:59:02 <nikhil_k> jokke_: agreed
15:59:21 <sigmavirus24> jokke_++
16:00:00 <sigmavirus24> Thanks everyone
16:00:01 <nikhil_k> #endmeeting