22:01:10 <adrian_otto> #startmeeting containers
22:01:10 <openstack> Meeting started Tue Aug 11 22:01:10 2015 UTC and is due to finish in 60 minutes.  The chair is adrian_otto. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:01:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:01:14 <openstack> The meeting name has been set to 'containers'
22:01:24 <adrian_otto> #link https://wiki.openstack.org/wiki/Meetings/Containers Our Agenda (will revise momentarily)
22:01:30 <adrian_otto> #topic Roll Call
22:01:32 <adrian_otto> Adrian Otto
22:01:38 <apmelton> Andrew Melton
22:01:42 <hongbin> o/
22:01:42 <bradjones> o/
22:01:44 <Tango|2> Ton Ngo
22:01:45 <dane_leblanc> O/
22:01:46 <yuanying-alt> o/
22:01:47 <daneyon_> Daneyon Hansen
22:01:52 <eghobo> o/
22:02:12 <thomasem> o/
22:02:18 <thomasem> Thomas Maddox
22:04:52 <adrian_otto> Revised agenda:
22:04:57 <adrian_otto> #link https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2015-08-11_2200_UTC Our Agenda
22:05:51 <adrian_otto> hello thomasem, eghobo, daneyon_, dane_leblanc, yuanying-alt, Tango|2, bradjones, hongbin, and apmelton
22:05:56 <adrian_otto> #topic Announcements
22:06:14 <adrian_otto> Reminder adrian_otto will be out on 2015-08-25 due to travel to OpenStack Silicon Valley event. An alternate chair is needed.
22:06:14 <vilobhmm> Vilobh Meshram
22:06:25 <adrian_otto> helo vilobhmm
22:06:35 <vilobhmm> hi adrian_otto
22:07:02 <suro-patz> o/
22:07:06 <adrian_otto> hello suro-patz
22:07:13 <hongbin> adrian_otto: is sdake available to chair that meeting?
22:07:20 <sdake> yo
22:07:23 <tcammann1> hello
22:07:32 <sdake> available yes
22:07:36 <adrian_otto> sdake, can you chair on 8/25?
22:07:37 <hongbin> k
22:07:46 <sdake> apologies for being late ;)
22:07:53 <sdake> thanks for pinging me
22:07:55 <adrian_otto> #agreed sdake will chair on 2015-08-25
22:08:09 <adrian_otto> any other announcements form team members?
22:08:21 <adrian_otto> *from
22:08:33 <adrian_otto> #topic Container Networking Subteam Update (daneyon_)
22:08:43 <mfalatic> o/ (network problems today)
22:08:51 <adrian_otto> welcome mfalatic
22:08:53 <daneyon_> not much new from the midcycle
22:09:04 <daneyon_> most cores provided a +2 to the network spec
22:09:21 <adrian_otto> we will review key takeaways from the midcycle further down in today's agenda
22:09:22 <daneyon_> tcammann1 has concerns
22:09:26 <tcammann1> :/
22:09:52 <daneyon_> it owuld be helpful if the core's can come to an agreement on the direction and i will update the spec accordingly
22:10:02 <daneyon_> i can understand where tcammann1
22:10:04 <daneyon_> is coming from
22:10:05 <eghobo> daneyon_: could you share link to your slides
22:10:05 <adrian_otto> ok, should we discuss now, or should we advance the agenda?
22:10:30 <tcammann1> We usually run out of time, now would be better
22:10:37 <daneyon_> maybe in the short-term (as noted in the spec), we focus on exposing params until we figure out the label thing
22:10:39 <tcammann1> Or reschedule a discussion
22:10:49 <adrian_otto> #link http://eavesdrop.openstack.org/meetings/container_networking/2015/container_networking.2015-07-30-18.00.html Previous Meeting
22:11:27 <adrian_otto> tcammann1: okay, please summarize the substance of your concern for discussion
22:11:28 <daneyon_> eghobo here is a link to the midcycle slides http://www.slideshare.net/daneyonhansen/magnum-networking-update
22:11:39 <eghobo> thx
22:11:43 <tcammann1> Read: https://review.openstack.org/#/c/204686/11/specs/container-networking-model.rst
22:11:47 <tcammann1> my comments
22:12:04 <daneyon_> adrian_otto maybe it's best for you to lead a ML discussion to reach a consensus
22:12:10 <daneyon_> and proceed with today's irc meeting
22:12:30 <sdake> ml discussion may be helpful although I think we can sort it out in the review
22:12:34 <tcammann1> the jist of it is that I believe its a bad idea to use a metadata field for real API values which affect Magnum's internals
22:12:51 <sdake> ya we sort of went around on this problem at the midcycle
22:12:59 <sdake> and didn't have any clear consensus on it
22:13:05 <tcammann1> Keep it in the review for now
22:13:09 <sdake> I believe we agreed that some common values should be parameters
22:13:09 <daneyon_> sdake the review will work too.
22:13:26 <adrian_otto> #action adrian_otto to use ML discussion to drive consensus on networking spec
22:13:29 <daneyon_> sdake... correct
22:13:52 <adrian_otto> any more on networking for today?
22:14:00 <daneyon_> it's difficult to ID the exact params until i work through the use case of extracting flannel from TL templates
22:14:20 <daneyon_> i have an idea of some params, i can put some general language about the label/param split
22:14:29 <tcammann1> Works for me :)
22:14:56 <daneyon_> tcammann1 re: works for me. do you mean you will +2 with the general language of params/label split?
22:15:10 <eghobo> +1 daneyon_, tcammann1 I think we should do just label for one release or two
22:15:35 <adrian_otto> I have not abstained from voting, just underwater temporarily.
22:15:53 <eghobo> and cleanup/standardise them after that
22:16:06 <daneyon_> adrian_otto no worries. sounds like the group would prefer feedback in the review instead of a ML thread. I'll let y'all make the call
22:16:09 <sdake> eghobo voice of reason - makes sense - once we know what they are
22:16:17 <tcammann1> If you updated the language around params and labels, yes. I don't expect you to list all the params
22:16:22 <adrian_otto> #undo
22:16:23 <openstack> Removing item from minutes: <ircmeeting.items.Action object at 0xa146f10>
22:16:38 <adrian_otto> #topic Magnum UI Subteam Update (bradjones)
22:16:38 <daneyon_> ok, so it sounds like we have a path fwd
22:16:54 <sdake> i'm certain brad is fast asleep atm
22:17:01 <daneyon_> i will update the spec to include language around the params/label split.
22:17:02 <bradjones> sdake: nope :D
22:17:02 <adrian_otto> brad was awake at roll call
22:17:06 <sdake> oh
22:17:09 * sdake runs :)
22:17:15 <bradjones> I have added a bunch of BPs
22:17:16 <tcammann1> cool, thanks daneyon_ :)
22:17:20 <daneyon_> any other core's have issues with the spec? speak now or forever hold your peace ;-)
22:17:20 <bradjones> #link https://blueprints.launchpad.net/magnum-ui
22:17:26 <adrian_otto> We got a visit from the Horizon team at our Midcycle
22:17:39 <adrian_otto> we expect to have a web UI during Liberty
22:17:52 <adrian_otto> Brad agreed to start giving us weekly updates on that
22:17:58 <bradjones> I have assigned myself to ones relating to bay model and targeted those at l3
22:18:04 <bradjones> If someone else is able to pick up the ones related to bay and work along side me, we could possible get that in too before liberty closes
22:18:27 <adrian_otto> Thai Tran volunteered to putting in the skeleton of a horizon plugin
22:18:39 <bradjones> I met with Thai last week
22:18:43 <bradjones> #link https://review.openstack.org/#/c/211750/ the initial skeleton repo is here
22:18:47 <bradjones> I need to fix a few issues with jenkins and add some skeleton Angular bits then that will be good to merge
22:19:48 <bradjones> I expect to have first revision of my bay model patches out by end of the week
22:19:50 <bradjones> so overall making good progress :)
22:20:03 <tcammann1> nice, bradjones
22:20:26 <daneyon_> bradjones right on!!!
22:20:28 <adrian_otto> from there we expect that putting the panes on that would be an achievable effort measured in days not weeks
22:20:29 <adrian_otto> bradjones: does that track with your expectations?
22:20:29 <adrian_otto> thomasem: you expressed interest
22:20:56 <adrian_otto> !! whoot !!
22:20:57 <openstack> adrian_otto: Error: "!" is not a valid command.
22:21:22 <bradjones> adrian_otto: yeah I expect to move pretty quickly
22:21:39 <adrian_otto> We are sure there will be a Magnum session at Tokyo, so we'd love to be showing that.
22:21:46 <sdake> i think there i ssome leg work to do on that review adrian_otto first
22:22:02 <thomasem> adrian_otto: I am interested in the UI for this. I am hoping to be available for reviews, however, I don't have bandwidth at the moment for the additional work, unfortunately. As is the case that I'm always sorry to say.
22:22:18 <adrian_otto> any more on magnum-ui?
22:22:18 <adrian_otto> #topic Review Action Items
22:22:18 <adrian_otto> 1) sdake/tcammann to sync up on https://blueprints.launchpad.net/magnum/+spec/objects-from-bay at MidCycle
22:22:18 <adrian_otto> Status:
22:22:21 <sdake> bradjones is it couple days per pane type of effort?
22:22:36 <sdake> failed
22:22:39 <sdake> but we got drunk together
22:22:46 <tcammann1> haha, we got close
22:22:54 <tcammann1> I mean to discussing it..
22:22:55 <adrian_otto> should we revise the action?
22:23:14 <bradjones> sdake: I should think 2-3 days to put something together that is in a reviewable state yes
22:23:15 <sdake> needs a main summit track session
22:23:46 <adrian_otto> sdake: ? what topic?
22:23:48 <sdake> adrian_otto so the action is to target it to tokyo
22:23:52 <tcammann1> I think its a sticky issue. Agree with sdake, a summit session would be good
22:24:00 <sdake> in the non-labeled tracks at ODS in tokyo
22:24:11 <adrian_otto> design summit
22:24:17 <adrian_otto> not main summit
22:24:22 <sdake> roger
22:24:26 <adrian_otto> 10/4
22:24:52 <adrian_otto> do we want an action to sync before that, or defer
22:25:28 <adrian_otto> no answer, so I am assigning it to you again
22:25:29 <adrian_otto> #action sdake/tcammann to sync up on https://blueprints.launchpad.net/magnum/+spec/objects-from-bay
22:25:32 <sdake> defer - nothing is going to happen on it - i'm swamped and i think tom is working on the template jinja2 thing
22:25:34 <tcammann1> defer it. Probably need to move the bp out of this release
22:25:37 <adrian_otto> #undo
22:25:38 <openstack> Removing item from minutes: <ircmeeting.items.Action object at 0x9be1bd0>
22:25:43 <adrian_otto> ok, undone.
22:25:49 <sdake> sorry really laggy today for some reason
22:25:50 <adrian_otto> 2) adrian_otto to open tech debt ticket to remove insecure option
22:26:06 <adrian_otto> decided not to do this. Justified by need for debugging/dev setups.
22:26:28 <adrian_otto> discussed at midcycle. Team consensus. If you disagree, see me to debate.
22:26:35 <adrian_otto> 3) apmelton to review snakeoil plugin and assist madhuri with integration
22:26:40 <adrian_otto> Status:
22:27:13 <apmelton> we talked about this at the mid-cycle, snakeoil plugin currently won't support what we need
22:27:37 <adrian_otto> Status: overtaken by events
22:27:46 <adrian_otto> #topic Blueprint/Bug Review
22:27:51 <adrian_otto> New Blueprints for Discussion
22:27:55 <adrian_otto> Team members may nominate new blueprints for discussion so they may be processed.
22:28:02 <adrian_otto> (any nominated for consideration?)
22:28:23 <adrian_otto> this is late in the cycle, so I will stop polling for these soon
22:28:28 <adrian_otto> Essential Blueprint Updates
22:28:47 <adrian_otto> please see: https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2015-08-11_2200_UTC section Essential Blueprint Updates
22:29:11 <adrian_otto> sdake, madhuri, Tango|2, apmelton Any updates?
22:29:27 <adrian_otto> we will touch on TLS in a moment
22:30:03 <Tango|2> I haven't heard back fron Angus Lees.  I have been out in the last few days also after the midcycle
22:30:25 <Tango|2> But I am working on 2 addtional patches to round out the support needed from Magnum
22:30:31 <sdake> adrian_otto pls ping me if you get to open items - my internet is terribl atm and I want to talk about our governance tag
22:30:41 <Tango|2> The rest is debugging the k8s backend
22:30:44 <adrian_otto> ok
22:31:12 <adrian_otto> Blueprints for administrative action
22:31:19 <adrian_otto> #link https://blueprints.launchpad.net/magnum/+spec/magnum-service-list Add service list to magnum (suro-patz)
22:31:25 <adrian_otto> Status: Proposal is to close this as won't fix, as we did not get a strong requirement from the ML-follow-up,
22:31:34 <adrian_otto> #link http://lists.openstack.org/pipermail/openstack-dev/2015-July/070165.html
22:31:44 <adrian_otto> I adjusted the BP
22:32:02 <adrian_otto> if anyone feels strongly, follow up in the whiteboard of that BP
22:32:10 <suro-patz> We have reached consensus on what to be done
22:32:20 <suro-patz> I am about to start implementation
22:32:26 <adrian_otto> thanks suro-patz
22:33:08 <sdake> overloaded - the consensus was what?
22:33:10 <adrian_otto> #topic Midcycle Summary
22:33:29 <sdake> confused by the "close as wont fix" vs "ready to start implementation"
22:33:30 <adrian_otto> #link https://etherpad.openstack.org/p/magnum-liberty-midcycle-topics Midcycle Notes
22:33:58 <suro-patz> consensus was to implement 'coe-service-list' and 'service-list' separately. This has been captured on the BP sdake:
22:34:06 <sdake> cool
22:34:09 <sdake> that wfm
22:34:16 <tcammann1> I think we identified that it would be good to have 'nova like' service list
22:34:18 <adrian_otto> bay-service-list
22:34:28 <adrian_otto> but this is details we can discuss offline
22:34:36 <adrian_otto> conceptually we did agree
22:34:43 <sdake> just got confused by the closed wont fix statement
22:34:50 <tcammann1> same :)
22:34:50 <suro-patz> sdake: The agenda was set before the midcycle.
22:34:54 <sdake> my irc is coming in 10 line chunks
22:35:07 <adrian_otto> ok, so we held the midcycle last week
22:35:18 <adrian_otto> see the etherpad link for notes
22:35:44 <adrian_otto> biggest topics were TLS, Networking, and Web UI
22:36:07 <adrian_otto> I heard from Madhuri yesterday
22:36:50 <adrian_otto> She switched employers from NEC to Intel, and will continue work on Magnum. She was traveling during the Midcycle so she could not participate.
22:37:25 <adrian_otto> Murali has joined the Magnum team and has confirmed we are using the v3 keystone API now, and is plumbing us for trusts to assist with the TLS implementation
22:37:49 <sdake> sweet
22:37:49 <adrian_otto> this will allow for secure scoped communication from bay nodes to the magnum API for the purpose of singing/fetching TLS keypairs
22:38:00 <adrian_otto> apmelton is also helping
22:38:09 <adrian_otto> tcammann1 has been advising.
22:38:30 <apmelton> adrian_otto: do you have madhuri's new email?
22:38:45 <adrian_otto> current plan of record is to implement Anchor as a library to allow a single line of code in magnum to produce sighed TLS cert/key
22:38:45 <apmelton> I need to sync with her on where I can help, but I don't see her online
22:38:57 <adrian_otto> apmelton: not yet. It might take a couple of days for that.
22:38:59 <sdake> i dont have her new work email and also don't hae a personal email for her, but can contact her on linkedin pamelton
22:39:14 <sdake> you want me to relay a message?
22:39:33 <apmelton> sdake: see if you have have her reach out to me at andrew.melton@rackspace.com
22:39:41 <sdake> will do right now
22:39:42 <apmelton> in the mean time I'll start working on updating the template
22:40:03 <apmelton> that should be sufficiently removed that I don't clash with anything she may be working on
22:40:09 <adrian_otto> on the subject of networking daneyon gave an overview talk, and Mohammed from the kuryr project also presented
22:40:10 <apmelton> the template changes*
22:40:33 <adrian_otto> these ideas are surfaced in the container networking spec mentioned earlier in our agenda
22:41:02 <adrian_otto> basically use libnetwork to communicate with existing remote drivers, and devleop remote drivers for openstack integration
22:41:17 <adrian_otto> there are gaps in terms of how to deal with issues like identity trust
22:41:30 <sdake> apmelton message relayed
22:41:35 <apmelton> much appreciated sdake
22:41:41 <adrian_otto> if you are interested, check our kuryr and join the magnum containers networking subteam
22:42:25 <adrian_otto> stauts of magnum Web UI was mentioned earlier in today's agenda
22:42:51 <adrian_otto> good discussions on auto-scaling
22:42:56 <adrian_otto> any questions?
22:43:54 <adrian_otto> (pausing due to IRC latency)
22:44:09 <adrian_otto> #topic Open Discussion
22:44:13 <adrian_otto> sdake: ping
22:44:16 <sdake> yar
22:44:28 <sdake> so we list our governance repo as allowing our project to release whenever we please
22:44:29 <vilobhmm> adrian_otto, sdake, tcammann1 : If its ok with you, I would like to work on https://blueprints.launchpad.net/magnum/+spec/objects-from-bay …I have prior experience of working with object layer and have contributed code in nova, cinder for the same. If its ok with you, depending on the progress is made in upcoming week we can decide what to do with the bp. I can co-ordinate with sdake about what his vision was about this blueprint and star
22:44:45 <sdake> yet we are releasing on l1/l2/l3 cycle and tagging at end of liberty
22:45:03 <sdake> after liberty may make more sense to go to a different release schedule if we desire
22:45:29 <sdake> the good thing about l1/l2/l3 is they are 1 month windows
22:45:35 <sdake> which is a good sprint size for a team our size
22:45:41 <sdake> first size/length
22:46:28 <adrian_otto> vilobhmm: take ownership of the blueprint, and discuss with other stakeholders.
22:46:34 <sdake> vilobhmm wfm but its very complicated ;)
22:47:08 <sdake> so my wall of text was to kick of a discussion
22:47:08 <adrian_otto> sdake: we can conform to the integrated release schedule
22:47:11 <sdake> not seeing that ;0
22:47:18 <adrian_otto> the current designation allows for that
22:47:22 <vilobhmm> adrian_otto : sure…sdake : lets see :)
22:47:41 <sdake> agree its more flexible
22:48:15 <sdake> maybe this is better had on the ml
22:48:17 <adrian_otto> once we deliver consistently on that schedule, we can adjust formally to toggle the goveranance tag for releases
22:48:33 <sdake> i like that course of action adrian_otto
22:49:01 <adrian_otto> but I agree we should aim for conformance, and I can do a better job of following that.
22:49:02 <tcammann1> I think deadlines with our releases are a good thing
22:49:51 <sdake> i have been trying to stay on strct schedue lwith kolla and i can tell you it is very difficult because our gate is imperfect
22:49:51 <hongbin> I have a question
22:50:07 <sdake> magnum has bettr gating but not much :)
22:50:07 <hongbin> what is the plan after l3, are we going to freeze or continue to land features?
22:50:30 <sdake> our governance allows us to land features
22:50:36 <sdake> but i think we should stick with what works
22:50:41 <adrian_otto> we are not required to freeze
22:50:45 <hongbin> good to know
22:51:02 <sdake> a different tag is required to freeze
22:51:06 <adrian_otto> and considering we don't yet have the burden of production deployments of Magnum… should continue landing features
22:51:27 <tcammann1> Freeze would be bad in an early stage of the project
22:51:34 <hongbin> Agree
22:51:46 <adrian_otto> but I will defer to the core reviewer team to pick some work to intentionally defer if it makes sense
22:52:07 <sdake> even the 6mo tag allows feature landing with FFE at PTL's discretion
22:52:15 <sdake> only if the release is handled by the release team is FFE a PITA
22:52:23 <sdake> those are really for super core projects like nova neutron keystone etc
22:52:25 <adrian_otto> so we may compassionately -2 some patches (temporarily)
22:52:43 <adrian_otto> so that's a middle ground between FFE and the wild west.
22:52:59 <sdake> wfm
22:53:07 <adrian_otto> we did that in the last release cycle on a very small number of contributions
22:53:15 <adrian_otto> and later merged them
22:53:18 <sdake> cool well i'm glad we sorted that out now :)
22:53:46 <manjeets_> question here on freeze ?
22:54:25 <adrian_otto> here is the guideline: If you have features that are in scope for prioritized blueprints, we will plan to merge that work
22:54:37 <manjeets_> will the proxy-blueprint patch be merged if its working fine or will be delayed for later merge
22:54:45 <manjeets_> ?
22:54:48 <tcammann1> we aren't having a freeze
22:55:01 <manjeets_> lok
22:55:02 <adrian_otto> if you have work that's not defect remediation, and adds features that are not scoped in prioritized blueprints scoped for Liberty, then we may refrain from merging that
22:55:20 <adrian_otto> and that will be up to the discretion of the core review team
22:56:04 <adrian_otto> basically we should feel there is no risk to including something that's arriving during those final weeks of Liberty
22:56:27 <adrian_otto> does that sound fair to all of you?
22:56:57 <adrian_otto> manjeets_: that's a minor change, and we would agree to merge it
22:57:23 <manjeets_> ok  thanks
22:57:29 <sdake> even mature projects merge after freeze
22:57:32 <sdake> only blueprints require ffe, bugs dont
22:57:49 <adrian_otto> right, if you are fixing a bug, then we will merge that work
22:58:06 <sdake> sounds like we may even merge blueprints ;)
22:58:07 <adrian_otto> we may ask that the bug be properly verified first
22:58:23 <manjeets_> ffe ?
22:58:32 <adrian_otto> only a few minutes remaining
22:58:40 <Tango|2> feature freeze extension
22:58:42 <sdake> feature freeze exception
22:58:58 <manjeets_> thanks
22:58:59 <adrian_otto> our next team meeting will be on 2015-08-18 at 1600 UTC.
22:59:13 <adrian_otto> thanks everyone for attending today.
22:59:14 <tcammann1> feature freeze exuberance
22:59:27 <thomasem> Take it easy!
22:59:29 <sdake> there are all kinds of words that begin with e;)
22:59:32 <adrian_otto> #endmeeting