#openstack-meeting log

09:59:51 <bauzas> #startmeeting climate
09:59:52 <openstack> Meeting started Mon Nov 25 09:59:51 2013 UTC and is due to finish in 60 minutes.  The chair is bauzas. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:59:53 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:59:56 <openstack> The meeting name has been set to 'climate'
10:00:05 <bauzas> morning all
10:00:14 <bauzas> who's there ?
10:00:24 <scroiset> \o/
10:00:36 <DinaBelova> o/
10:00:43 <DinaBelova> hello, guys
10:00:48 <bauzas> Hi
10:00:53 <scroiset> Hi Dina
10:01:12 <bauzas> I had no news about Francois, he's probably still ooo
10:01:28 <bauzas> as he was at the SC 2013 expo
10:01:46 <bauzas> hi Nikolay_St
10:01:48 <DinaBelova> ok, see it
10:02:01 <bauzas> #link https://wiki.openstack.org/wiki/Meetings/Climate#Agenda_for_November.2C_25
10:02:17 <bauzas> today I'll host the chair
10:02:29 <DinaBelova> and we have no special agenda, I suppose
10:02:35 <DinaBelova> today, I mean
10:02:36 <bauzas> nope, reviewed it
10:02:56 <DinaBelova> ok, see it in our chanel
10:03:00 <bauzas> as said, any other concerns should be raised during the "open" topic
10:03:02 <bauzas> so
10:03:15 <bauzas> #topic action items from last meeting
10:03:25 <DinaBelova> http://eavesdrop.openstack.org/meetings/climate/2013/climate.2013-11-18-10.04.html
10:03:25 <bauzas> #link http://eavesdrop.openstack.org/meetings/climate/2013/climate.2013-11-18-10.04.html
10:03:34 <bauzas> was faster than you :)
10:03:50 <DinaBelova> mmm, I see otherwise :D
10:03:54 <bauzas> let's go by action item/person
10:04:04 <DinaBelova> ok
10:04:20 <bauzas> so, I made a few comments on https://review.openstack.org/#/c/52296
10:04:35 <bauzas> Nikolay_St also had action for rebasing it
10:04:39 <bauzas> he did
10:04:41 <DinaBelova> I see Nikolay has only rebased it
10:04:43 <DinaBelova> yes
10:04:55 <bauzas> yup, Nikolay_St, had you time to review my comments ?
10:04:58 <DinaBelova> Nikolay, will you please fix comments?
10:05:30 <SergeyLukjanov> o/
10:05:37 <DinaBelova> Ok, I suppose we may create action item for Nikolay
10:05:42 <DinaBelova> SergeyLukjanov, hello
10:06:08 <bauzas> I'm sorry, we have terrible Internet connection at our office in Grenoble
10:06:15 <bauzas> sure
10:06:30 <bauzas> so I won't be able to quickly glance at the reviews
10:06:31 <DinaBelova> bauzas, yes, that's always a problem
10:06:36 <Nikolay_St> bauzas: yeap
10:06:57 <Nikolay_St> I'll do it in few days
10:06:59 <bauzas> #action Nikolay_St Review comments from bauzas on https://review.openstack.org/#/c/52296 and amend patch
10:07:12 <Nikolay_St> I'm close to end with tests
10:07:24 <bauzas> cool
10:07:37 <bauzas> DinaBelova: your actions ?
10:07:44 <bauzas> POC using shelved instances ?
10:07:45 <DinaBelova> yes, let's go to them
10:08:00 <bauzas> had you time for looking at it ?
10:08:09 <DinaBelova> there was no time to it...
10:08:23 <DinaBelova> so other two action items are done/in progress
10:08:28 <bauzas> ok
10:08:51 <DinaBelova> as for POC, I will do it after oslo.messaging intergation
10:09:08 <bauzas> ok
10:09:17 <DinaBelova> or Nikolay will do it himself in case he will end up with his current tasks quicker
10:09:23 <bauzas> ok
10:09:38 <Nikolay_St> yeap
10:09:51 <Nikolay_St> I think I'll end soon
10:09:57 <bauzas> #agreed POC on shelved instances to be done after oslo.messaging BP
10:10:04 <DinaBelova> Nikolay_St, as for your action atimes
10:10:13 <DinaBelova> items***
10:10:26 <DinaBelova> wow, keyboard wants to kill me
10:10:45 <bauzas> well, Nikolay_St only had action to rebase previous patch
10:10:54 <DinaBelova> yes, and that's it
10:10:58 <bauzas> yup
10:11:11 <Nikolay_St> yeap
10:11:13 <bauzas> let's talk about proc/cons
10:11:23 <bauzas> of triggering Openstack releases for Climate
10:11:38 <bauzas> I think that's a bit early for discussing that
10:11:49 <bauzas> we need to see our velocity
10:12:16 <bauzas> ie. we can agree to aim to deliver for icehouse-2
10:12:33 <DinaBelova> I suppose we have already started using them (while BPs assignment and time management) - at least in terms of bugs and BPs
10:12:41 <bauzas> and target blueprints accordingly
10:12:42 <SergeyLukjanov> i1 will be 1.5 weeks, i2 is in 2 months
10:13:03 <DinaBelova> i1 is impossible for something really working
10:13:13 <SergeyLukjanov> bauzas, agreed, i2 looks ok to have first release
10:13:16 <bauzas> yup, but all, please keep in mind that should be a temptative attempt
10:13:19 <DinaBelova> i2 is better
10:13:23 <bauzas> yup
10:13:33 <Nikolay_St> +1 to Dina
10:13:37 <bauzas> we have a few iteams for i-1
10:13:51 <DinaBelova> we have several i1 dates
10:13:54 <bauzas> yup,
10:13:55 <DinaBelova> in launchpad
10:14:05 <bauzas> do we agree to postpone them to i2 ?
10:14:14 <bauzas> just for clarification ?
10:14:20 <DinaBelova> but all of them are about critical bugs/bps for critical discussion
10:14:38 <bauzas> DinaBelova: as said, that doesn't mean we can't deliver for i1
10:14:59 <bauzas> I would say let's see what happens with the deliverables
10:15:15 <bauzas> target i2 for Climate v0.1
10:15:25 <SergeyLukjanov> bauzas, what's the profit to keep i1 w/o releasing it?
10:15:41 <bauzas> that's a good question
10:15:51 <DinaBelova> I think first we won't  have release on i1
10:16:09 <bauzas> I target i1 for a few BPs just saying these ones should be delivered to trunk during i1
10:16:21 <DinaBelova> but there are several things that should be merged, etc. for i1 date
10:16:22 <DinaBelova> yes
10:16:25 <DinaBelova> the same thing
10:16:37 <bauzas> to me, there are 2 things to consider :
10:16:48 <bauzas> 1/ we have a rolling delivery about BPs and bugfixes to trunk
10:16:57 <bauzas> 2/ we tag a special release for Climate
10:17:04 <DinaBelova> It seems to me we may leave these already created i1 things as is
10:17:12 <DinaBelova> Cannot catch 2/
10:17:17 <DinaBelova> May you explain?
10:17:25 <bauzas> that's what I would call Climate v0.1
10:17:29 <SergeyLukjanov> not i1, i2, but 0.1 for example
10:17:33 <SergeyLukjanov> in the middle of iX
10:17:55 <Nikolay_St> SergeyLukjanov: you mean between i1 and i2?
10:18:03 <SergeyLukjanov> yep
10:18:05 <DinaBelova> bauzas, do you want v0.1 for i2? Or just separated line of releases?
10:18:12 <DinaBelova> Like SergeyLukjanov said?
10:18:14 <bauzas> 0.1 for i2 yes
10:18:26 <bauzas> ie. having a special tag release
10:18:31 <Nikolay_St> and so 0.x for i3?
10:18:45 <bauzas> possibly yes, possible not
10:18:58 <DinaBelova> I suppose, Sylvain, that's ok
10:19:04 <DinaBelova> but Sergey meant something else
10:19:07 <DinaBelova> i think
10:19:09 <SergeyLukjanov> bauzas, I've proposed it on prev meeting and you was on the side of following OpenStack release process ;)
10:19:24 <SergeyLukjanov> there are several options I see now
10:19:31 <SergeyLukjanov> and several input params for it
10:19:32 <bauzas> SergeyLukjanov: that's exactly why we're discussing now :D
10:19:44 <SergeyLukjanov> first of all the scope for "0.1"
10:20:01 <SergeyLukjanov> it should be at least one working "reference" plugin excluding fake one
10:20:20 <bauzas> if we consider 0.1, we would need to tag BPs accordingly
10:20:36 <SergeyLukjanov> one cons for using iX as some first release - tarballs will not be uploaded to pypi
10:20:37 <bauzas> ie. saying for 0.1, you will find all of these features
10:20:58 <bauzas> SergeyLukjanov: interesting
10:21:20 <DinaBelova> One quick moment, I'm ready to manage releases for every decided point
10:21:33 <DinaBelova> whatever we'll decide here now
10:21:39 <SergeyLukjanov> IMO 0.1, 0.1.X, [0.2,] Icehouse is the best option for climate
10:21:51 <SergeyLukjanov> to have some tarballs published to pypi
10:21:51 <DinaBelova> SergeyLukjanov, agree
10:22:02 <bauzas> agree
10:22:05 <scroiset> agree
10:22:06 <SergeyLukjanov> and make releases when something will be ready
10:22:09 <Nikolay_St> +1
10:22:16 <bauzas> scroiset: ?
10:22:22 <scroiset> +1
10:22:26 <DinaBelova> I've had a look on how Savanna is working with releases and this way seems to be nicest one
10:22:44 <SergeyLukjanov> we used it before the incubation
10:22:57 <SergeyLukjanov> to be much more flexible to release new features
10:23:00 <bauzas> #agreed Climate releases 0.1, .., 0.X to be considered during Icehouse release
10:23:00 <SergeyLukjanov> we can discuss dates and probably adjust them with iX
10:23:02 <SergeyLukjanov> release
10:23:04 <DinaBelova> and I think that's comfortable for un-incubated projects, yes
10:23:11 <DinaBelova> ok, great
10:23:11 <bauzas> yup
10:23:27 <DinaBelova> so now we have only open discussions left
10:23:30 <bauzas> we just need to scope Climate 0.1
10:23:37 <SergeyLukjanov> yep
10:23:38 <bauzas> DinaBelova: nope
10:24:00 <bauzas> #action Scope Climate 0.1
10:24:35 <DinaBelova> Is that for now or for further discussions?
10:24:45 <SergeyLukjanov> the best option is to have one virt and one hardware reservation plugins
10:24:45 <bauzas> let's just review 2nd topic
10:25:00 <bauzas> #topic high priority issues
10:25:20 <bauzas> https://review.openstack.org/#/c/57650/ is just rebased
10:25:24 <bauzas> oops
10:25:36 <bauzas> s/rebased/patchsett'd
10:25:40 <DinaBelova> and tests added, yes
10:25:56 <bauzas> #action bauzas Review https://review.openstack.org/#/c/57650/ by this week
10:26:02 <DinaBelova> ok, I think this context staff we may merge soon
10:26:15 <DinaBelova> in case there will be nothing to add
10:26:24 <bauzas> sure
10:26:31 <bauzas> there was only a typo
10:26:33 <bauzas> to me
10:26:42 <bauzas> should be merged by today if no objections
10:26:47 <bauzas> https://review.openstack.org/#/c/57200/
10:26:53 <bauzas> policy management
10:27:00 <bauzas> I'm nearly done with it
10:27:08 <DinaBelova> I'm going to take a look on it today
10:27:12 <bauzas> there is still a typo in the policy.json I delivered
10:27:24 <DinaBelova> and leave comments if find  something
10:27:27 <bauzas> and I also need to add an extra method get_admin_roles()
10:27:44 <bauzas> for implementing the elevated() method in the context
10:28:10 <DinaBelova> Ok, I see that in the commit message
10:28:14 <bauzas> that's not hard work
10:28:22 <DinaBelova> btw, masterofuniverse is nice role
10:28:25 <bauzas> so we don't need an is_admin flag
10:28:34 <DinaBelova> I had much fun reading it
10:28:49 <DinaBelova> ok, i think
10:29:06 <DinaBelova> we should also review/merge that by the end of this week
10:29:18 <DinaBelova> because policies are quite important
10:29:44 <bauzas> #action bauzas Deliver last patchset for https://review.openstack.org/#/c/57200/ by eob today
10:30:11 <DinaBelova> never say last for patchsets :D
10:30:19 <bauzas> any other high priority issues to discuss ?
10:30:26 <DinaBelova> I suppose no
10:30:29 <DinaBelova> that's it
10:30:32 <bauzas> if not, let's move to open discussion
10:30:38 <bauzas> #topic open discussion
10:30:59 <DinaBelova> I have nothing to add
10:31:08 <DinaBelova> if we speak about extra topics
10:31:09 <bauzas> well, I'll open a bug for the exceptions management
10:31:21 <DinaBelova> to discuss
10:31:27 <DinaBelova> bauzas, ok
10:31:41 <YorikSar> bauzas: Unfortunately we don't have some final words from Duncan on that ML thread. So I still don't think it's wise to purge is_admin just because.
10:31:46 <bauzas> #action bauzas Open a bug for exception handling with code
10:31:55 <SergeyLukjanov> #info https://review.openstack.org/#/c/58152/ auto updates for requirements
10:32:04 <YorikSar> bauzas: Oh, not Duncan, but Dolph.
10:32:08 <SergeyLukjanov> #info https://review.openstack.org/#/c/57675/ climate channel logging
10:32:23 <bauzas> YorikSar: what I can propose is to add a get_admin_roles() from policies management
10:33:02 <bauzas> so we could put all the admin roles in the roles[] list if elevated()
10:33:23 <bauzas> the policy.json file I will propose won't make use of the is_admin flag
10:33:25 <YorikSar> bauzas: The point is that we don't have a way to get all roles that can be treated as admin roles because it can even be one separate admin role for every call.
10:33:35 <DinaBelova> but still they have different meanings
10:33:36 <DinaBelova> yes
10:33:44 <bauzas> yup
10:34:10 <bauzas> what I can say is that the policy handler will provide if necessary a list of roles given by policy.json
10:34:12 <bauzas> that's it
10:34:41 <bauzas> based on the context_is_admin rule in policy.json
10:35:00 <bauzas> SergeyLukjanov: good initiative for logging :)
10:35:07 <YorikSar> bauzas: Oh... That sounds hacky...
10:35:08 <DinaBelova> YorikSar, what do you think about it?
10:35:10 <bauzas> SergeyLukjanov: I gave a +1
10:35:27 <YorikSar> bauzas: Ok, let's move this discussion out of this meeting.
10:35:43 <bauzas> well, as Dolph said, we need to define a rule in policy.json saying which roles do have admin rights
10:35:44 <DinaBelova> we all did it i think :D for Sergey's commits :)
10:36:07 <DinaBelova> ok, do we something else to discuss?
10:36:12 <bauzas> YorikSar: ok, let's discuss that on the regular channel
10:36:17 <bauzas> DinaBelova: I don't think so
10:36:22 <bauzas> thanks all
10:36:30 <DinaBelova> ok, bye!
10:36:35 <bauzas> DinaBelova: I'll send the email about minutes
10:36:46 <bauzas> you ok ?
10:36:53 <DinaBelova> ok, sure
10:36:57 <bauzas> cool thanks
10:37:00 <bauzas> thanks buddies
10:37:05 <bauzas> #endmeeting