09:00:30 <ttx> #startmeeting ptl_sync
09:00:31 <openstack> Meeting started Tue Mar  3 09:00:30 2015 UTC and is due to finish in 60 minutes.  The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:00:33 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:00:35 <openstack> The meeting name has been set to 'ptl_sync'
09:00:50 <ttx> #topic Heat
09:01:10 <ttx> #link https://launchpad.net/heat/+milestone/kilo-3
09:01:19 <asalkeld> ttx: the only bp's in "not started" are convergence blueprints
09:01:39 <asalkeld> there are a bunch of convergence blueprints, this was a very complex blueprint that we
09:01:39 <ttx> yes -- what's your goal with that ?
09:01:39 <asalkeld> broke up into many small blueprints for the purposes of splitting the work amounst the team.
09:01:46 <asalkeld> i'd like to do the convergence blueprints a bit more flexibly for a number of reasons
09:01:46 <asalkeld> 1. they are quite small each
09:01:47 <asalkeld> 2. this feature is disabled by default and we want to do as much as possible so we can enable the feature early in Liberty
09:01:47 <asalkeld> 3. they are very inter-dependant so we can't start many until the initial ones are in
09:02:18 <ttx> asalkeld: would it make sense to develop it in a feature branch ?
09:02:34 <asalkeld> ttx: i prefer merging
09:02:51 <asalkeld> i am not sure we have a good mechanim for that
09:02:52 <ttx> any chance it will be "completed" for kilo (even if disabled) ?
09:03:14 <asalkeld> yeah, but 50/50
09:03:27 <asalkeld> one or two of those might miss
09:03:35 <asalkeld> out of like a dozen
09:03:36 <ttx> ok, I guess it could get exceptions, if it's totally disabled
09:03:57 <asalkeld> yip, it's a big job and we are super focused on it
09:04:13 <asalkeld> shame to just say no more work until branching
09:04:49 <ttx> what about keystone-resources ? That's marked "started"
09:05:05 <asalkeld> the author says he should have a patch up later today
09:05:11 <asalkeld> (so i left it up)
09:05:20 <ttx> I suppose convergence-push-data is part of "convergence" ?
09:05:25 <asalkeld> ttx: yip
09:05:43 <ttx> OK, so basically you have a FPF for everything but convergence (which gets an exception)
09:05:48 <asalkeld> he is working on it today (small item)
09:06:00 <ttx> keystone-resources needs to be up for review by Thursday
09:06:03 <asalkeld> ttx: sounds good
09:06:15 <asalkeld> ok, sounds reasonable
09:06:19 <ttx> yes
09:06:51 <ttx> #info FPF for everything but *convergence* BPs, keystone-resources still needs to put something up
09:07:23 <asalkeld> agree
09:07:47 <asalkeld> anything else ttx ?
09:07:48 <ttx> asalkeld: in other news, I'll be at the ops summit next week, so we'll skip the formal 1:1 (and contact as-needed during that wekk by pining on this channel)
09:07:57 <ttx> week*
09:07:59 <ttx> pinging*
09:08:08 <asalkeld> ok, see you then
09:08:19 <ttx> asalkeld: have a good week!
09:08:23 <asalkeld> you too
09:12:00 <ttx> #topic Nova
09:12:21 <ttx> johnthetubaguy is on holiday this week
09:12:26 <ttx> just a few notes he sent me:
09:12:46 <ttx> #info Some sub groups are busy deciding what goes to liberty, but most seem to be aware of the deadline
09:12:53 <ttx> #info All exceptions are merged now.
09:13:13 <ttx> #link https://launchpad.net/nova/+milestone/kilo-3
13:00:49 <eglynn_> ttx: knock, knock ... ready when you are
13:00:54 <ttx> eglynn_: here you are
13:01:00 <ttx> #topic Ceilometer
13:01:17 <eglynn_> #link https://launchpad.net/ceilometer/+milestone/kilo-3
13:01:18 <ttx> #link https://launchpad.net/ceilometer/+milestone/kilo-3
13:01:21 <ttx> ah!
13:01:34 <ttx> FPF in two days, that leaves 2 outliers
13:01:47 <ttx> configdb-api and conf-datastore-agents
13:01:57 <eglynn_> so good progress on everything except configdb-api & conf-datastore-agents
13:02:02 <eglynn_> exactly
13:02:03 <ttx> Are both expected to post in the coming days, or are you granting exceptions ?
13:02:17 <eglynn_> so FPF is Friday, I;ve made that point repeatedly
13:02:28 <ttx> You seem to be in a good enough shape feature-wise to survive exceptions
13:02:57 <ttx> but then Fabio has 3 specs in progress
13:02:59 <eglynn_> well I'd prefer to have initial patches proposed by Mar 5th, but if needs be we could let it slide into the week after
13:03:09 <eglynn_> it's not just Fabio working on them
13:03:12 <ttx> ok
13:03:20 <eglynn_> he has a HP colleague helping him out
13:03:49 <ttx> ok, you seem to have things under control. happy with what you delivered in kilo overall ?
13:04:13 <eglynn_> well, the ceilo LP doesn't capture the progress made on gnocchi in parallel
13:04:32 <eglynn_> so taken as a whole, a fair bit has been acheived
13:04:38 <ttx> eglynn_: if you had to single out a couple key features, what would that be ?
13:04:56 <ttx> (top of your head, no need to over-think it)
13:05:05 <eglynn_> probably the events/notification pipelines that gordc has been working on
13:05:23 <eglynn_> (bringing us closer to what stacktach is capable of on that side)
13:05:34 <eglynn_> also the gabbi API testing work that cdent has done
13:05:41 <ttx> ok, I'll look into that!
13:05:48 <eglynn_> (potential for application to a bunch of other projects)
13:06:03 <eglynn_> and of course the gnocchi work that jd__ has been spearheading
13:06:13 <eglynn_> (though that's at arm's length from ceilo for now)
13:06:51 <ttx> eglynn_: in a big-tent model, would you keep gnocchi in the "ceilometer team" stuff ? Or would it make sense as a separate team ?
13:07:20 <ttx> iow, how disjoint is gnocchi dev from ceilometer dev ?
13:07:35 <eglynn_> I think I mentioned to you before that jd__ is interested in "big tenting" gnocchi
13:07:55 <eglynn_> he probably hasn't made a direct approach yet to the TC as he was on PTO the last two weeks
13:08:03 <ttx> ok
13:08:20 <eglynn_> the dev teams are a bit disjoint in the sense that gnoochi-core is a subset of ceilo-core
13:08:24 <ttx> I'll be at the Ops summit and skipping formal 1:1s next week
13:08:40 <eglynn_> but I've been encouraging jd to extent that gnocchi-core team
13:08:51 <ttx> ack
13:09:04 <eglynn_> are the formal mechanics/process for making a big tent application all set now?
13:09:05 <ttx> so ping me on IRC or by email if you need me for anything
13:09:10 <eglynn_> coolness
13:09:30 <ttx> eglynn_: yes. We just don't advertise the process that much since we'd like to start slow
13:09:35 <ttx> https://wiki.openstack.org/wiki/Governance/NewProjectTeams
13:09:49 <eglynn_> cool, I'll pass that on Julien
13:09:54 <ttx> but right now it's so slow nobody applied
13:09:58 <ttx> :)
13:10:17 <eglynn_> cool, gnocchi might be your first customer so :)
13:10:41 <ttx> he should anticipate the question of the relationship with ceilo, since I expect quite a few TC members are not up to speed on that
13:10:53 <ttx> (I for one would benefit from a refreshed view)
13:11:03 <eglynn_> yep, that makes sense
13:11:14 <ttx> if you don't have anything else, talk to you later!
13:11:26 <eglynn_> cool, thanks for your time :)
13:11:30 <ttx> SergeyLukjanov: yt?
13:11:38 <SergeyLukjanov> ttx, hi!
13:11:43 <ttx> #topic Sahara
13:11:47 <SergeyLukjanov> #link https://launchpad.net/sahara/+milestone/kilo-3
13:12:12 <SergeyLukjanov> so, we have a bunch of new approved specs and lots of the merged features
13:12:40 <ttx> Still 10 missing final code up, so FPF might be a bit short for you?
13:12:45 <SergeyLukjanov> few of items are now candidates to move to the Liberty release (not started high :( )
13:13:09 <ttx> SergeyLukjanov: last week you were unsure of observing FPF in Sahara, what's your take today ?
13:13:27 <ttx> granting a few exceptions maybe ?
13:13:45 <ttx> (but not started stuff should definitely be deferred, unless it's really small
13:13:47 <ttx> )
13:14:31 * SergeyLukjanov counting nuber of potential FPFE
13:14:47 <ttx> You fully control FPF fwiw
13:14:51 <ttx> so you decide :)
13:15:02 <ttx> (I only enforce FF and FFEs)
13:15:31 <SergeyLukjanov> okay, so, I think that we'll try to make FPF this week with a few exceptions
13:15:47 <ttx> sounds good
13:16:01 <SergeyLukjanov> it looks like it's possible to do, mostly all of the important is stuff is on review (that was started)
13:16:19 <SergeyLukjanov> I'm going to defer non-started specs to Liberty soon
13:16:23 <ttx> Are you happy with what you got done in this cycle ?
13:16:45 <ttx> if you had to single out a couple key kilo features, what would that be ?
13:16:55 <ttx> (I'd like to look into them before we release)
13:17:12 <SergeyLukjanov> ttx, yeah, it's going extremely great, we got a lot of new contributors and new feautres
13:17:45 <SergeyLukjanov> ttx, it's about maturity, operability plus brand new CDH support (Cloudera disto of Hadoop)
13:18:09 <SergeyLukjanov> plus a lot of improvements in Heat integration, security
13:18:25 <ttx> ok
13:18:34 <ttx> Will skip next week 1:1s as I'm at the Ops summit
13:18:41 <ttx> ping me directly in case of need
13:18:53 <SergeyLukjanov> in a few words - several new plugins and tons of tech, UX and securityimprovements
13:18:58 <SergeyLukjanov> ttx, okay, thx
13:19:10 <ttx> Have a great week!
13:19:12 <SergeyLukjanov> I think there is no question from my side
13:19:16 <SergeyLukjanov> ttx, thx, you too!
13:19:27 <ttx> dhellmann: ready when you are
13:19:33 <dhellmann> ttx: hi, there!
13:19:35 <ttx> #topic Oslo
13:19:45 <ttx> #link https://launchpad.net/oslo/+milestone/kilo-3
13:19:56 <ttx> #link https://launchpad.net/oslo/+milestone/next-kilo
13:20:21 <ttx> dhellmann: hwo are those graduations coming up ?
13:20:39 <dhellmann> versionedobjects has some infra-related changes for dan to make
13:21:04 <dhellmann> the reports one was kickstarted when I suggested we drop it last week, so it's just starting but there is a repository to be reviewed and  imported
13:21:50 <ttx> ok, we'll see how it goes
13:22:16 <dhellmann> yeah, it's unlikely to be done by k-3, so I suppose we could remove it, but I want to let solly keep working on it one way or the other
13:22:47 <ttx> Any particular kilo achievement you'd like to highlight ? My impression are that the key ones are procedural
13:23:06 <ttx> like a streamlined release process
13:23:27 <dhellmann> yes, although the policy and versionedobjects work has brought us new contributors as well
13:23:27 <ttx> but I may have missed a critical lib graduation :)
13:24:42 <dhellmann> hmm, is there a way to have lp show all of kilo, or just milestones?
13:25:45 <ttx> apparently the magic behind finsing comon milestones across projects in a projectgroup doesn't extend to series
13:25:52 <ttx> finding*
13:26:20 <dhellmann> the oslo.log graduation wrapped up this cycle will be important for the work we need to do based on sdague's cross-project blueprint next cycle
13:26:34 <ttx> https://blueprints.launchpad.net/oslo-incubator/kilo only has the incubator
13:26:34 <dhellmann> also the common request context library
13:27:40 <dhellmann> yeah, I'm looking at the separate milestone pages one at a time instead
13:28:04 <dhellmann> I don't think there were any library features that were really big
13:28:08 <ttx> Like I told the others, we'll skip 1:1s next week while I'm at the Ops summit
13:28:16 <dhellmann> ok
13:28:27 <ttx> I added "Add library stable release procedures/policy" for the cross-project meeting today
13:28:46 <ttx> although I'm not sure that will result in a very useful discussion at the current stage
13:28:55 <ttx> but we need to move on
13:29:06 <ttx> the fact that not enough people care shouldn't block it
13:29:54 <dhellmann> disappointing number of votes on that one
13:30:00 <ttx> would be good to see if Brant can remove his -1 based on your reply
13:30:39 <dhellmann> I'll ask him to look again
13:30:46 <ttx> dhellmann: anything else you wanted to discuss ?
13:31:15 <dhellmann> we have a few releases to make today, but none of the changes are very big so I don't expect trouble
13:31:37 <ttx> ok, talk to you later then!
13:31:42 <dhellmann> have a good day!
14:33:09 <ttx> dhellmann: added a couple comments on the library stable spec
14:33:21 <dhellmann> ttx: ack
14:33:25 <ttx> (to make sure I got it right initially :)
15:00:15 <mestery> ttx: Here and waiting sir.
15:00:23 * mestery knows he's early ... again ;)
15:00:24 <dhellmann> ttx: I responded with a new draft
15:00:35 <dhellmann> mestery: stop showing off ;-)
15:00:41 <mestery> dhellmann: lol :)
15:01:06 <mestery> dhellmann: My arrival time depends on whether or not we have a neutron meeting preceding this 1:1, and the status of my kids on some mornings :)
15:01:42 <dhellmann> mestery: mine really only depends on whether I've been up in time to have breakfast first
15:01:53 <mestery> dhellmann: rofl :)
15:02:38 <ttx> #topic Neutron
15:02:49 <ttx> And I thought I couold have a break, sigh
15:02:53 <ttx> mestery: o/
15:03:21 <ttx> #link https://launchpad.net/neutron/+milestone/kilo-3
15:05:41 <ttx> dhellmann: +2ed
15:05:53 <mestery> ttx: I've been going through the list, there are many BPs which are close to merging and need a final push, I expect a lot to land later this week.
15:06:29 <ttx> I count 11 that still need code proposed to match FPF
15:06:52 <mestery> ttx: Yes, in yesterday's meeting I reiterated that to everyone.
15:06:53 <ttx> Feeling ready to defer those ? Or granting exceptions ?
15:06:58 <mestery> ttx: I expect some of those to not have code proposed by FPF
15:07:07 <mestery> I'll grant some exceptions I expect.
15:07:10 <ttx> rigth, but then there is plenty enough to review already
15:07:22 <mestery> y
15:07:49 <ttx> mestery: overall, how would you say that cycle went -- It felt a bit more under control to me, but then I'm not really in it
15:08:14 <mestery> ttx: I agree! It's gone super from my perspective! We've had some slips with thigns as always ,but overall, I'm very happy with the way things are going.
15:08:33 <ttx> If you had to single out a couple of key features, what would that be ? I guess the decomposition of -aaS and plugin stuff is one
15:09:00 <mestery> Yes, decomposition and the split of the advanced services. Those, along with all the work refactoring and stabilizing all the agents have been huge.
15:09:18 <mestery> I expect Neutron Kilo to be of high quality in the stability department
15:09:22 <mestery> Famous last words I guess :)
15:09:50 <ttx> I'll be at the ops sumit next week, so not around for our 1:1s
15:10:09 <mestery> Ack, sounds good and safe travels, it's cold and snowy in the eastern US :)
15:10:22 * ttx looks up weather
15:10:26 <mestery> lol
15:10:57 <ttx> I hope that's Celsius and not Fahrenheit
15:11:13 <mestery> rofl
15:11:44 <ttx> mestery: ok, have a great week, and get ready for the hammer at the end of the week :)
15:12:33 * mestery straps on the kevlar :)
15:12:38 <mestery> thanks ttx, safe travels, talk in 2 weeks!
15:12:38 <ttx> nikhil_k: ready when you are
15:15:25 <nikhil_k> ttx: o/
15:16:31 <ttx> #topic Glance
15:16:38 <ttx> #link https://launchpad.net/glance/+milestone/kilo-3
15:16:58 <ttx> nikhil_k: looks like you're pretty close to target so far
15:17:10 <ttx> Could you add a priority to https://blueprints.launchpad.net/glance/+spec/metadefs-upgrade-by-json-file ?
15:17:25 * nikhil_k thought he did
15:17:40 <ttx> reloading still returns undefined
15:17:49 <nikhil_k> done
15:17:56 <nikhil_k> (sorry, that was new)
15:18:08 <ttx> So, only https://blueprints.launchpad.net/glance/+spec/catalog-index-service is missing code up
15:18:18 <ttx> which means you're technically very close to a FPF
15:18:39 <nikhil_k> ttx: that has code up, let me check if everything is in place though
15:18:52 <nikhil_k> and I will change the status accordingly
15:19:03 <ttx> At this point I'd spend cycles reviewing and landing those features, an exception there wouldn't really hurt
15:19:52 <ttx> nikhil_k: do you feel you accomplished what you wanted in this cycle ?
15:20:04 <ttx> (it's not done yet, but since we'll soon freze...)
15:20:09 <ttx> +e
15:20:42 <nikhil_k> ttx: :) almost.. code is being partially reviewed
15:20:50 <ttx> nikhil_k: if you had to single out a couple of new kilo glance stuff, what would that be ?
15:20:53 <nikhil_k> so some features are half complete
15:21:26 <nikhil_k> ttx: catalog index service and artifacts
15:21:30 <nikhil_k> ttx: and this one https://blueprints.launchpad.net/glance/+spec/deactivate-image
15:21:56 <nikhil_k> that one has code in relatively better shape and needs a little bit of work
15:22:12 <ttx> ack, thx, will try to have a look into that
15:22:31 <ttx> I'll miss the 1:1s next week due to Ops Summit, but feel free to ping me as needed
15:22:44 <ttx> That is all I had.
15:23:04 <nikhil_k> cool, will keep that in mind
15:23:32 <nikhil_k> ttx: just fyi, I proposed zhiyan for stable core (email subject [openstack-dev] [stable] [Glance] Nomination for glance-stable-maint )
15:23:51 <ttx> nikhil_k: yes, saw that -- you don't really need other cores to +1 that btw
15:24:09 <ttx> I just need to make sure additions have read the stable branch policy before I add them
15:24:28 <ttx> so let me know when I should reach out
15:24:45 <ttx> at this point I'm waiting for your green flag
15:24:48 <nikhil_k> ttx: whenever convenient from now
15:24:54 <ttx> oh ok.
15:24:56 <ttx> WIll send
15:24:59 <nikhil_k> ttx: should I say that on the email or we'r good?
15:25:07 <ttx> I'll post something there
15:25:15 <nikhil_k> thanks ttx !
15:25:20 <ttx> doing that now
15:25:36 <nikhil_k> cool
15:27:12 <ttx> thingee: howdy
15:27:19 <thingee> ttx: hi!
15:27:20 <ttx> #topic Cinder
15:27:31 <ttx> #link https://launchpad.net/cinder/+milestone/kilo-3
15:27:54 <ttx> We are past your FPF (was on March 1st) and your k3 status matches that
15:28:05 <ttx> I'd say you're looking pretty good
15:28:32 <thingee> thanks!
15:28:44 * ttx should create liberty series and l1 milestone in LP
15:28:51 * ttx adds to todo list
15:29:34 <ttx> How would you say this cycle went, now that we have started descent to release ?
15:30:19 <thingee> This milestone makes up for all of it :)
15:30:37 <ttx> thingee: if you had to single out a couple features, what would that be ?
15:31:34 <thingee> well I was going to say that all the driver updates didn't really help in the last release. Haven't quite figured out what to do there.
15:32:02 <thingee> Kind of hard when you implement new features in k-1 or k-2 and then there's a small window for drivers to do implement them.
15:32:36 <ttx> thingee: looks like you need... a longer cycle ! (just kidding)
15:32:44 <thingee> heh
15:33:11 <thingee> I'm not really sure what the answer there is yet. People get scared when you say last milestone will just been bug fixes/core related things
15:33:22 <ttx> so you have new features but drivers have not necessarily caught up to them ?
15:34:06 <thingee> that's correct. I have plans on better coordinating with drivers in the future on new features. Mostly because only the active vendors keep up. I just think communication there has always not been that great.
15:34:33 <thingee> But I also want some features to eventually be considered features that I can say will work across all backends.
15:34:44 <ttx> thingee: I think the real solution is to accept that they won't be up to date and ocument that properly
15:34:48 <ttx> +D
15:35:00 <thingee> So they remain as optional features and not part of our minimum features list. The minimum feature list is what makes our driver api so great in my opinion.
15:35:12 <ttx> something like an API microversion that they implement up to
15:35:25 <ttx> and "driver updates" would increment that microversion support
15:36:23 <ttx> so you'd have version x.y.z which adds feature a and b, and then some drivers submit a x.y.z support update and some others don't
15:36:44 <ttx> or you can go the neutron road and decouple them
15:36:59 <ttx> or you can do a longer cycle :)
15:37:42 <thingee> That's an interesting idea. Well we have the ABC stuff done in an abstract class for the driver now. The minimum features are represented in this base minimum abstract class. All drivers must support that, or cinder won't even start up properly. unit tests would fail. As we graduate new features, we move them to that.
15:37:59 <thingee> you can tell if all drivers implement it a feature you move to that.
15:38:28 <ttx> right, I think that's sane -- I was more wondering how to solve the "new feature adopeted in few drivers" issue
15:38:39 <thingee> communication
15:38:45 <thingee> to start
15:38:46 <thingee> :)
15:38:49 <ttx> you could have features as levels and then ask drivers to submit level updates
15:39:23 <ttx> and then document what each driver level is. But that may require a discipline people don't have
15:39:36 <ttx> (i.e. what happens if they impelment all of level 45 except feature C
15:39:43 <thingee> it has shown to be more successful in the third party ci then previous attempts anyways. As much as I want vendors to be around all the time to listen to updates.
15:40:01 <ttx> ok, anyway, longer discussion ahead
15:40:13 <ttx> I'll be at the ops sumit next week, so missing our 1:1
15:40:21 <thingee> I'll be there
15:40:26 <ttx> great!
15:40:28 <thingee> good time to get feedback
15:40:29 <ttx> see you there then
15:40:56 <ttx> david-lyle: ready when you are
15:41:10 <david-lyle> ttx: readu
15:41:17 <david-lyle> close enough
15:41:22 <ttx> #topic Horizon
15:41:32 <thingee> ttx: I'll talk to you more about this later. thanks for the help
15:41:43 <ttx> #link https://launchpad.net/horizon/+milestone/kilo-3
15:42:14 <ttx> david-lyle: IIRC you said you would try to enforce FPF at the end of the week
15:42:22 <ttx> you seem to be a bit far away from there
15:42:50 <ttx> 16 blueprints are not yet in "needs code review"
15:42:59 <david-lyle> I think there will be a couple of exceptions, I'm repruning list
15:43:09 <david-lyle> haven't made it to the bottom yet
15:43:15 <ttx> ok
15:43:42 <david-lyle> high priority item around launch instance rework will likely have a couple of exceptions
15:43:43 <ttx> I thihnk to maiximize the quantity of stuff getting done, having a clean (and lean) list of stuff to review in the lmast two weeks helps
15:43:56 <ttx> I can't type today
15:44:15 <david-lyle> yes, we've been sprinting/review daying on the launch instance work
15:44:20 <ttx> alright. So you will reprune the list and grant a couple exceptions
15:44:24 <david-lyle> yes
15:44:39 <ttx> That sounds good to me
15:45:24 <ttx> So what are the likely key things that will land in kilo, now that we are sufficiently close to the end ?
15:45:36 <ttx> I bet the launch instance rework is one
15:45:42 <david-lyle> still looking to move to specs for next release and avoid the open assignment to milestones
15:46:40 <david-lyle> launch instance, table rework (search)
15:46:48 <david-lyle> are the biggest items
15:47:04 <david-lyle> big UX improvements that set us up well for future progress
15:47:40 <david-lyle> I will end up picking up the django item
15:47:44 <ttx> david-lyle: as far as big-tent goes, I think Horizon will have to decide what they directly support and what they externally support
15:47:53 <ttx> that may translate into two tags
15:48:09 <ttx> something like horizon-built-in and horizon-plugin
15:48:17 <david-lyle> yes, we're continuing to discuss that
15:48:27 <ttx> projects you feel comfortable supporting would be horizon-built-in
15:48:40 <ttx> projects providing a horizon-plugin of reasonable quality could be horizon-plugin
15:48:42 <david-lyle> I think a lot of the overhead right now is from things that could be plugin
15:49:06 <ttx> right, like I wrote elsewhere, I expect big-tent to clarify things and increase focus rather than decrease it
15:49:26 <david-lyle> I think that's wise, trying to cover too much scope, we just drown
15:49:31 <ttx> and be honest about what we (horizontal stuff) can support and not
15:50:04 <ttx> I prefer we say "we support A B C and we help the others by giving them advice" rather than "we support anything they throw at us"
15:50:20 <david-lyle> the UX team will become more integral once things move to plugins
15:50:22 <ttx> and then not deliver consistent quality
15:50:45 <ttx> indeed
15:50:57 <david-lyle> ttx, I agree, I think horizon provides a solid, extensible base on some core services and their interrelation
15:51:15 <david-lyle> the rest builds on that outside in plugins
15:51:28 <david-lyle> I think that's not just true for horizon
15:51:41 <david-lyle> so I think we're in violent agreement
15:51:58 <ttx> right, that's the base -- doesn't mean you can't support directly /some/ stuff, but that should entirely be Horizon team decision, not the TC or someone else's
15:52:14 <ttx> exactly same with release management
15:52:29 <david-lyle> yes
15:52:43 <ttx> as long as you provide base framework / tools / advice for those other teams to do it themselves, that's sane (and more scalable)
15:52:55 <ttx> Anyway, I'll be at Ops Summit next week, so skipping this 1:1
15:53:00 <david-lyle> what would be the process for extricating code already included?
15:53:09 <david-lyle> has anyone thought about that?
15:53:36 <ttx> david-lyle: that would be a repo split -- not worse that what neutron did this cycle
15:53:46 <david-lyle> with horizon we support plugins, so rolling code out into plugins wouldn't be terrible, but maybe not an easy conversation
15:53:51 <ttx> we can conserve history and all
15:54:14 <ttx> technically not that difficult. Socially...
15:54:21 <david-lyle> yes
15:54:32 <ttx> so.. got to run; have a great week, and ping me if needed!
15:54:40 <david-lyle> ok thanks
16:51:41 <ttx> morganfainberg: ready when you are
16:51:57 <morganfainberg> O/
16:52:04 <ttx> #topic Keystone
16:52:17 <ttx> #link https://launchpad.net/keystone/+milestone/kilo-3
16:52:43 <ttx> That leaves 2 blueprints still needing to drop code before eow
16:52:45 <morganfainberg> So today going to push out a couple of those.
16:53:00 <ttx> reseller & keystone-tokenless-authz-with-x509-ssl-client-cert
16:53:09 <ttx> shorter title and longer title evar
16:53:09 <morganfainberg> Reseller is going to push out and x509 likely is as well
16:53:39 <ttx> Feels like you have things under control
16:53:48 <morganfainberg> Code that will likely merge today will be the domain sql one and the last of the ae/kelt/fernet token ones.
16:54:13 <ttx> As we begin our descent to final release, what are your first thoughts on how this cycle went for Keystone ? Did you end up doing what you wanted ?
16:54:16 <morganfainberg> Domain sql might request a ffe, but that is just to gate not because it's still being written.
16:54:58 <morganfainberg> I feel that we crammed too much into k3. I want to roll back fpf to m2 for next release of I'm ptl
16:55:09 <morganfainberg> If*
16:55:14 <ttx> If you had to highlight a couple of kilo features/successes, what would those be ?
16:55:25 <morganfainberg> the fernet/ae token is a big win
16:55:46 <ttx> I still need to dig in that. Was hoping one of those blogposts your team is so fond of about it
16:55:48 <morganfainberg> It is addressing the biggest scaling issue of keystone. The persistence requirement for tokens.
16:56:16 <morganfainberg> Yeah I haven't had the time to write one up:). Dolph wrote up a good one as did lance.
16:56:47 <ttx> "Keystone to Keystone Service Providers" is the next step in k2k federation right
16:56:47 <morganfainberg> Also the move away from "extensions" is a big win I'm happy about.
16:57:05 <morganfainberg> Yes k2k is the next step in federation.
16:57:28 <morganfainberg> We also have added code to let deployers map users from an external Idp to a local user.
16:58:00 <morganfainberg> Instead of federated users always being ephemeral.
16:58:15 <ttx> ack, will try to dig in all of those for my personal culture
16:58:28 <ttx> I'll skip 1:1s next week due to Ops Summit
16:58:30 <morganfainberg> K2k needed that for some bigger use cases (cern iirc really could use that)
16:58:34 <morganfainberg> ++
16:58:34 <ttx> ping me in case of need
16:58:44 <ttx> That is all I had
16:58:52 <ttx> have a great week!
16:58:55 <morganfainberg> after ops summit I'll un -1 the sql downgrade spec
16:59:01 <ttx> yap
16:59:02 <morganfainberg> Same to you, cheers.
17:00:02 <ttx> notmyname: ready when you are
17:01:12 <notmyname> good morning. I'm ready
17:01:20 * notmyname just responded to the release tags patch
17:01:23 <ttx> #topic Swift
17:01:38 * ttx looks
17:01:59 <notmyname> we can carry that on in gerrit :-)
17:02:20 <notmyname> ok. /me has 8 minutes until a phone call
17:02:20 <ttx> heh, yes. If teher was a simple solution we'd all agree already
17:02:27 <notmyname> :-)
17:02:29 <ttx> ok, what news ?
17:02:36 <notmyname> two things off the top of my head
17:02:44 <ttx> I see work landing on the encryption branch alright
17:02:49 <notmyname> #info Swift's OPW intern has completed her project
17:02:50 <notmyname> yay
17:02:55 <ttx> yay
17:03:00 <notmyname> this week is the last OPW scheduled week
17:03:06 <notmyname> so perfect timing
17:03:15 <notmyname> also, status update on EC...
17:03:29 <notmyname> we've made good progress in the last week. things are looking good
17:03:45 * notmyname is continually impressed by and happy to work with swift contributors
17:03:48 <ttx> notmyname: any window for the big merge to master ?
17:04:15 <notmyname> this week (ie today->tomorrow) we'll be making the call, based on current momentum and outstanding work, if we will commit to an EC beta in Kilo
17:04:23 <ttx> ok
17:04:41 <ttx> That sounds good
17:04:45 <notmyname> if so (IF), then we will schedule the merge to master in about 4 weeks. allowing for 2 weeks for that to land before the RC
17:04:57 <ttx> I'll be at Ops Summit next week, so skipping this 1:1 -- keep me posted on the decision though
17:05:03 <notmyname> scheduling the RC around the 10th. and the merge window starting mar 27
17:05:08 <notmyname> I too will be at the ops meetup
17:05:21 <ttx> great! I'll see you there
17:05:30 <notmyname> those dates above are dependent on the EC decision
17:05:39 <ttx> understood
17:05:47 <notmyname> in other news, yes the crypto work in progressing
17:06:05 <notmyname> and on the project management side, I made something
17:06:07 <ttx> #info call to be made on including EC beta in kilo this week
17:06:20 <ttx> #info if it is a go, will schedule the merge to master in about 4 weeks. allowing for 2 weeks for that to land before the RC
17:06:39 <notmyname> http://goo.gl/9EI0Sz and http://goo.gl/uRzLBX are 2 new/updated review dashboards for swift. on is for reviewers. one is an overview
17:07:09 <ttx> Starred Patches, interesting
17:07:09 <notmyname> the reason I want to share those is because I'm doing something there I haven't seen done before: a dashboard listing stuff starred by other people
17:07:14 <notmyname> ya, I think so
17:07:35 <notmyname> so on the review one, it includes the PTL's starred patches. those are the same as the priority reviews
17:07:40 <notmyname> from https://wiki.openstack.org/wiki/Swift/PriorityReviews
17:08:05 <ttx> you should definitely share those on some of the angry threads
17:08:06 <notmyname> on the project overview dash, there is a section for "Starred by any core". that gives me an idea of what people are interested in and looking at
17:08:43 <notmyname> I think it will be helpful to us. Even yesterday I found it useful for review queue hygiene
17:09:03 <notmyname> that's all I've got. anything from you? anything I can do?
17:09:17 <ttx> nothing comes to mind right now
17:09:23 <ttx> Jump on next meeting!
17:09:26 <ttx> and see you next week
17:09:28 <notmyname> ok. see you next week
17:09:39 <ttx> devananda: ready when you are
17:10:17 <devananda> ttx: hi!
17:10:21 <ttx> #topic Ironic
17:10:28 <ttx> #link https://launchpad.net/ironic/+milestone/kilo-3
17:10:57 <ttx> So I see 5 blueprints needing code land before FPF eow
17:11:01 <ttx> I mean cope up
17:11:05 <ttx> code up*
17:11:27 <devananda> yea, we're going through that this week to make sure everything's in shape
17:11:51 <ttx> You already have a lot to review -- planning to strictly observe FPF, or are exceptions considered ?*
17:12:01 <devananda> pretty strict
17:12:08 <devananda> there's far more than we'll actually be able to do already accepted
17:12:12 <devananda> and we know that
17:12:21 <ttx> good :)
17:12:34 <devananda> we have not had to bump things in the past, because last cycle we didn't allow in things that we didn't have confidence in landing
17:12:40 <devananda> so this will be new for our community
17:12:51 <ttx> So, now that you ahve a pretty good idea of what will actually likely be in kilo, how would you say that cycle went for Ironic ?
17:12:58 <devananda> we've put a lot of time into vetting designs, so there's confidence in that
17:13:16 <devananda> but there is probably not enough time left to review all the code for each of those
17:13:44 <devananda> i think many core reviewers do not like this structure
17:13:49 <devananda> *release structure
17:14:06 <devananda> and would rather have continual development & stabilization with much smaller iterations
17:14:22 <devananda> but we'll see how the rest of the cycle goes, and i'm sure we'll talk about it in vancouver
17:14:31 <ttx> we definitely will
17:14:38 <devananda> I expect in a few weeks, there will be a minor revolt
17:14:47 <devananda> when I kick half of these BPs to Liberty
17:15:01 <ttx> It's weird, because the recent push was rather away from significant milestones and considering them just a tag
17:15:10 <devananda> and development essentially stalls for a month
17:15:28 <devananda> it's sort of like we just built up momentup for the cycle
17:15:29 <ttx> and it feels like people want to tag more often
17:15:42 <devananda> last two milestones were mostly spec work, with some foundational code merged
17:15:54 <devananda> now that that's done, there's a BUNCH of work that folks are / have done, that we want to land
17:15:58 <devananda> but the window to land it is too small
17:16:12 <devananda> because of (seemingly) artificial milestones
17:16:24 <ttx> right, so there was a mismatch with the time-based segment and your internal wave
17:16:28 <devananda> yup
17:16:56 <ttx> in THEORY, the various points in the cycle shoudl help in aligning the internal momentum with the cycle
17:17:13 <ttx> but as we made milestones less significant, I think we lost that
17:17:26 <devananda> indeed
17:17:42 <devananda> also, now that ironic has microversion support in the API
17:17:49 <devananda> folks are more inclined to allow faster iteration on things
17:18:02 <ttx> right
17:18:17 <devananda> I'm not sure that's wise -- just because we CAN do microversions does not mean we should start being less cautious about API changes
17:18:22 <devananda> but it is a trend
17:20:05 <ttx> OK, so as I said to the others, skipping 1:1s next week due to Ops Summit
17:20:33 <devananda> *nod*
17:20:41 <ttx> last question would be -- what are the key achievements for the Ironic team this cycle ? I want to lok into those before we go deeper in release mode
17:20:51 <devananda> say hi to folks for me - i wish I was going, but need a break from travel
17:21:34 <devananda> key achievements ... state machine is a massive one
17:21:36 <devananda> #link https://blueprints.launchpad.net/ironic/+spec/new-ironic-state-machine
17:21:48 <devananda> we might not finish implementing all of it this cyucle, though we're going to try
17:21:51 <devananda> and we're close
17:22:45 <devananda> configdrive support
17:22:47 <devananda> #link https://blueprints.launchpad.net/ironic/+spec/driver-periodic-tasks
17:22:53 <devananda> woops, wrong link
17:22:53 <devananda> #undo
17:23:01 <devananda> #link https://blueprints.launchpad.net/ironic/+spec/expose-configdrive
17:23:31 <devananda> adding zapping and cleaning stages, which in turn allow RAID and firmware config to be implemented by drivers
17:23:42 <devananda> also, node introspection through ironic-discoverd
17:23:53 <devananda> which is still on stackforge, pending integration testing
17:24:07 <devananda> once it has real CI, i'd like to move it to openstack/ namespace
17:24:22 <ttx> devananda: is that likely to be happening in kilo ?
17:24:26 <devananda> no
17:24:31 <ttx> ok
17:24:47 <ttx> devananda: need to switch to Trove. Have a great week!
17:24:52 <devananda> ok! yo utoo
17:24:58 <ttx> SlickNik: ready?
17:25:47 <SlickNik> ttx: here
17:26:02 <SlickNik> #link https://launchpad.net/trove/+milestone/kilo-3
17:26:17 <ttx> #topic Trove
17:26:32 <SlickNik> whoops, got a bit too eager there :)
17:26:35 <ttx> Those are interesting colors
17:26:59 <ttx> So still 3 in jeopardy, likely to be dropped at end of week
17:27:06 <ttx> And a whole pile to review
17:27:13 <ttx> Is that a fair summary ?
17:27:42 <SlickNik> Yes — been driving to see which ones of the 3 can make it before FPF, and getting folks to review the ones that are already out there.
17:27:51 <SlickNik> Good summary.
17:27:57 <ttx> sounds all good to me
17:28:30 <ttx> Asking the same question I asked others... what are the key features you think are in Kilo Trove ?
17:28:36 <ttx> or will be, rather
17:28:46 <ttx> features or key team achievements
17:29:03 <ttx> I'd like to look into them before I'm stuck in release mode
17:29:47 <SlickNik> Support for a two new datastores is top of the list — CouchDB and Vertica.
17:30:26 <ttx> ack
17:30:29 <SlickNik> Support for Vertica clusters.
17:30:52 * ttx doesn't even know what Vertica is, so I'll start by that one :)
17:31:08 <ttx> SlickNik: I'll be away at the Ops SUmmit next week, so skipping 1:1
17:31:23 <SlickNik> Vertica = column based datastore
17:31:24 <ttx> don't hesitate to ping me on IRC or email in case of need though
17:31:34 <SlickNik> Will do
17:32:01 <ttx> SlickNik: have a great day and week
17:32:05 <ttx> That concludes today's series of 1:1 syncs. Thanks for listening...
17:32:08 <ttx> #endmeeting