16:00:05 <adrian_otto> #startmeeting containers
16:00:08 <openstack> Meeting started Tue Jul 12 16:00:05 2016 UTC and is due to finish in 60 minutes.  The chair is adrian_otto. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:09 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:12 <openstack> The meeting name has been set to 'containers'
16:00:15 <adrian_otto> #link https://wiki.openstack.org/wiki/Meetings/Containers#Agenda_for_2016-07-12_1600_UTC Our Agenda
16:00:21 <adrian_otto> #topic Roll Call
16:00:24 <strigazi> Spyros Trigazis
16:00:25 <Drago> o/
16:00:26 <adrian_otto> Adrian Otto
16:00:26 <muralia> Murali Allada
16:00:31 <vijendar1> o/
16:00:40 <dane_leblanc> o/
16:00:42 <jvgrant_> Jaycen Grant
16:00:43 <tango> Ton Ngo
16:00:59 <mtanino> o/
16:01:22 <adrian_otto> hello strigazi, Drago, muralia, vijendar1, dane_leblanc, jvgrant_, tango, and mtanino
16:01:48 <adrian_otto> I'll wait a moment for more participants to join us
16:01:59 <rpothier> Rob Pothier
16:02:00 <adrian_otto> hello rpothier
16:03:10 <adrian_otto> #topic Announcements
16:03:13 <adrian_otto> I am serving as chair today while hongbin is on vacation.
16:03:43 <adrian_otto> 1) If you are submitting talks for Barcelona, the deadline is tomorrow.
16:03:57 <adrian_otto> any other announcements form team members?
16:04:07 <adrian_otto> *from
16:04:23 <adrian_otto> #topic Review Action Items
16:04:26 <adrian_otto> (none)
16:04:32 <adrian_otto> #topic Essential Blueprints Review
16:04:38 <adrian_otto> 1) Support baremetal container clusters (strigazi)
16:04:45 <adrian_otto> #link https://blueprints.launchpad.net/magnum/+spec/magnum-baremetal-full-support
16:04:49 <strigazi> I'm fixing the gate job for k8s-ironic #link https://review.openstack.org/#/c/336433/
16:04:55 <strigazi> I rebased the fixed templates on k8s drivers patch #link https://review.openstack.org/#/c/320968/
16:05:02 <strigazi> Madhuri is working on swarm but couldn't attend the meeting today.
16:05:21 <strigazi> That's it for this week
16:05:43 <adrian_otto> thanks strigazi. I'll review that today.
16:05:52 <adrian_otto> any questions about this topic?
16:06:17 <adrian_otto> ok, advancing to the next...
16:06:25 <adrian_otto> 2) Magnum User Guide for Cloud Operator (tango)
16:06:27 <adrian_otto> #link https://blueprints.launchpad.net/magnum/+spec/user-guide
16:06:40 <tango> I uploaded the section for bay:  https://review.openstack.org/#/c/340670/
16:07:17 <tango> This should cover all the key new sections.  Next I will start consolidating existing docs into the user guide
16:07:41 <tango> like TLS, Mesos
16:07:48 <tango> Horizon plugin
16:07:57 <tango> So, steady progress
16:08:03 <adrian_otto> tango: thanks. When you consolidate, lets take sections that belong both in the user guide and the quickstart, and brea them into common files that each includes
16:08:15 <adrian_otto> so we follow the DRY principle
16:08:31 <adrian_otto> otherwise maintenance may get sloppy
16:08:34 <tango> DRY?
16:08:47 <adrian_otto> Don't Repeat Yourself (code/doc duplication)
16:08:52 <tango> ok
16:08:57 <adrian_otto> tx!
16:09:03 <tango> That's all I have for now
16:09:07 <adrian_otto> Any discussion on this BP?
16:09:14 <strigazi> yes
16:09:34 <strigazi> about lbaas, are we going to add it to the user guide?
16:09:47 <strigazi> We are close to decouple
16:09:56 <tango> Which aspect of lbaas?
16:10:07 <strigazi> and I could remove it from the install-guide
16:10:10 <adrian_otto> youmean guidance for how to add it, or integrate with it?
16:10:19 <strigazi> How to add it
16:10:49 <adrian_otto> ideally we should put a stub of the documentation with high level guidance as part of the commit(s) to decouple
16:11:05 <adrian_otto> it does not need to be perfect, but at least something directional
16:11:25 <strigazi> ok
16:11:38 <adrian_otto> we can iterate on it as needed to make it nice and tight
16:11:44 <tango> Let me take a closer look and we can figure out the best approach
16:11:46 <strigazi> On my next revision I'll put it with the optional services
16:12:06 <adrian_otto> tango, that much makes sense to me, Agreed?
16:12:23 <tango> yes
16:12:30 <adrian_otto> ok, great
16:12:49 <adrian_otto> any more on this topic?
16:13:15 <adrian_otto> 3) COE Bay Drivers (jamie_h, muralia)
16:13:21 <adrian_otto> #link https://blueprints.launchpad.net/magnum/+spec/bay-drivers
16:13:31 <muralia> We made good progress on this blue print last week
16:13:41 <muralia> most driver patches got merged.
16:13:46 <muralia> there's one remaining. https://review.openstack.org/#/c/339196/
16:13:59 <muralia> it needs to be reviewed and merged.
16:14:21 <muralia> once this is merged, the next step is to focus on using stevedore to dynamically load the driver
16:14:42 <muralia> so cores, please review the patch.
16:14:45 <muralia> thats all for now
16:14:53 <adrian_otto> thanks muralia.
16:14:57 <adrian_otto> Discussion on this BP?
16:15:25 <adrian_otto> ok, next one
16:15:27 <adrian_otto> 4) Create a magnum installation guide (strigazi)
16:15:33 <adrian_otto> #link https://blueprints.launchpad.net/magnum/+spec/magnum-installation-guide
16:15:36 <strigazi> This is the list of Installation tutorials and guides. Magnum is there. http://docs.openstack.org/project-install-guide/draft/index.html
16:15:48 <adrian_otto> should this BP be marked as implemented?
16:16:01 <strigazi> No, I'm testing debian packages that I have built from master and I'll push this week.
16:16:17 <strigazi> after that we mark it as implemented
16:16:24 <adrian_otto> ok, thanks.
16:16:25 <strigazi> I'll push three guides. debian with debconf, debian without debconf and ubuntu without debconf.
16:16:43 <adrian_otto> any questions on this BP?
16:17:02 <adrian_otto> The next topic is one I'd like to ask everyone about
16:17:03 <adrian_otto> Magnum UI Subteam Update (bradjones)
16:17:22 <adrian_otto> each week hongbin calls for this, and there is near universal silence afterward.
16:17:28 <adrian_otto> is anyone working on magnum-ui?
16:17:45 <Drago> I haven't seen a patch from bradjones for months
16:17:46 <adrian_otto> SHould this be dropped from the team meeting agenda?
16:18:24 <Drago> Or look for someone else to pick it up?
16:18:26 <adrian_otto> perhaps it should be dropped from the standing agenda, and only added by request as there are topics to share/discuss.
16:18:36 <adrian_otto> that's why I'm asking now.
16:18:46 <adrian_otto> I have a feeling the magnum team and the magnum-ui teams actually don't overlap
16:19:08 <tango> We can add to the agenda as needed
16:19:21 <adrian_otto> and if there is no activity, should we move to remove magnum-ui as a project?
16:19:30 <adrian_otto> or move it to an archived state?
16:20:19 <tango> Can we ping the ui team to get some feedback on their plan?
16:20:23 <adrian_otto> ok, I'll drop it from the agenda template for next week, and hongbin can decide what he'd like to do about it.
16:20:56 <adrian_otto> I can leave a placeholder there about inactivity concern
16:21:18 <adrian_otto> ok, next topic
16:21:27 <adrian_otto> #topic Kuryr Integration Update (tango)
16:22:27 <tango> I was on the IRC last night but apparently the Kuryr team did not meet.
16:23:02 <adrian_otto> Are there magnum team members working on this currently? I have not noticed reviews for this yet.
16:23:18 <tango> Now that I am back, I am resuming the integration work with our Swarm bay.
16:23:43 <adrian_otto> ok, be sure to base that on the swarm bay driver
16:24:06 <adrian_otto> no need to wait on that, but find the related patches currently in review
16:24:20 <tango> Yes, I noticed.  I am still in development mode.  Last thing I tried was to bring up the Kolla image for running Openvswitch agent
16:24:50 <adrian_otto> ok
16:25:03 <tango> Once I get the prototype working, I will start submitting patches
16:25:18 <tango> That's all for now
16:25:19 <adrian_otto> any more questions/discussions on kuryr integration?
16:25:34 <adrian_otto> #topic Other blueprints/Bugs/Reviews/Ideas
16:25:37 <adrian_otto> I have one
16:26:02 <adrian_otto> any others to register interest in before I start?
16:26:26 <tango> I would like to bring up Swarm 2.0
16:26:49 <adrian_otto> tango: ok. Thinking of doing that as a new bay driver?
16:27:07 <tango> Yes, or at least start the discussion
16:27:19 <tango> It's still early
16:27:41 <strigazi> tango: you mean 1.2?
16:27:48 <muralia> do we know the release date yet?
16:28:01 <tango> No, actually 2.0, separate from 1.2
16:28:05 <mtanino> Are there a plan to support Docker swarm mode at docker 1.12?
16:28:05 <strigazi> ok
16:28:31 <strigazi> 1.12 have a lot of functionality of swarm
16:28:37 <adrian_otto> yes, we will definitely use 1.12
16:28:41 <mtanino> strigazi:  yeah
16:29:09 <strigazi> we could start working on a bay driver for docker 1.12
16:29:14 <adrian_otto> so tango, I might suggest skipping swarm 2.0 and doing docker 1.12 instead
16:29:29 <muralia> should that just be version 2.0 of the same swarm bay driver we have today?
16:29:44 <strigazi> I would say a new driver
16:30:16 <strigazi> There are a lot new thing introduced in 1.12
16:30:22 <strigazi> *things
16:30:27 <adrian_otto> my rationale is because 1.12 bakes swarm right into docker.
16:30:28 <tango> 2.0 is significantly different, so it would make sense to have separate driver for 1.12 and 2.0
16:30:40 <muralia> drivers were based on just image-coe combination
16:30:57 <adrian_otto> oh, I see what you are thinking
16:30:58 <tango> We can do 1.12 for now, and start thinking about 2.0
16:31:07 <adrian_otto> ok, yes, that should be the order.
16:31:25 <mtanino> tango: looks good idea :)
16:31:57 <adrian_otto> any other reviews or BP's needing team discussion today?
16:33:14 <adrian_otto> ok, I have a question
16:33:35 <adrian_otto> I'm working with https://github.com/openstack/openstack-ansible-os_magnum
16:33:55 <adrian_otto> when we use that to set up Magnum, we get a working API, but...
16:34:23 <adrian_otto> we are stuck when we create a bay because the trustee user (bay user) is unable to download the certificate from the magnum API
16:34:33 <adrian_otto> the trust token is rejected by keystone
16:34:51 <adrian_otto> so we *think* that something about the roles for the trust are goofed up
16:35:26 <tango> Have you tried just using Heat alone?
16:35:29 <adrian_otto> so the symptom is that you can create a bay, but it remains CREATE_IN_PROGRESS forever
16:35:47 <adrian_otto> the cloiud-init script part-006 fails because of the condition I mentioned above
16:35:54 <adrian_otto> so cloud-init never finishes.
16:36:10 <adrian_otto> devstack seems to work fine
16:36:30 <adrian_otto> when I look at both devstack and osad deployment of magnum, both appear to be configured the same
16:36:47 <adrian_otto> each with the "admin" user and and admin role as the trustor and the bay user as the trustee.
16:36:54 <adrian_otto> so I'm at a loss to explain it.
16:37:48 <adrian_otto> One idea is that with osad, the trustee (bay user) has *two* roles assigned (heat_stack_owner and admin)
16:38:07 <adrian_otto> and in devstack, the trustee (bay user) only has *one* role assigned (admin).
16:38:25 <adrian_otto> is it possible that our policy check does not properly respect multiple roles?
16:39:13 <adrian_otto> strigazi or tango, any ideas?
16:40:03 <tango> I am setting up Magnum manually in a standard OpenStack environment.  I am seeing a problem that seems to be similar.
16:40:18 <tango> Keystone rejects the trust token
16:40:29 <strigazi> I'll try to regenerate it. Both swarm bay and k8s are not able to finish?
16:40:38 <tango> When I tried just running a Heat template, it fails also.
16:40:43 <chris__hultin> strigazi: We haven't tried Swarm yet.
16:40:55 <adrian_otto> we see the problem with k8s bays
16:41:08 <muralia> are the keystone versions the same as in devstack?
16:41:52 <chris_hultin> muralia: I'm not sure about DevStack, but we're running stable/mitaka versions of everything besides Magnum
16:42:03 <muralia> ok
16:42:05 <chris_hultin> adrian_otto:  Do you know what keystone version we are running?
16:43:33 <strigazi> We have keystone in liberty and we haven't come across this problem AFAIK
16:43:45 <strigazi> I'll ask again tomorrow
16:43:48 <adrian_otto> tango: in your malfunctioning setup, if you do a "openstack trust show <trust_id>" on the trust listed in the /etc/sysconfig/heat-params file, so you see multiple roles? Or do you see only "admin"
16:43:59 <adrian_otto> chris_hultin: I can look
16:44:45 <strigazi> chris_hultin adrian_otto for magnum we use the same domain as heat
16:45:03 <adrian_otto> devstack uses the domain "Default"
16:45:17 <adrian_otto> for the magnum trusts at least
16:45:25 <adrian_otto> I mean...
16:45:32 <adrian_otto> the admin users are in the Default domain
16:45:55 <tango> I will have to check later, but I think it would be multiple roles
16:46:11 <adrian_otto> The only user I see in the magnum domain in devstack is "trustee_domain_admin"
16:46:43 <strigazi> instead of this user we use heat_domain_admin
16:46:56 <adrian_otto> oh, ok
16:46:57 <strigazi> we, at CERN
16:47:03 <adrian_otto> we can give that a try
16:47:52 <strigazi> We did so because of CERN restrictions on users
16:48:19 <strigazi> We haven't try with magnum_domain_admin at all
16:49:06 <adrian_otto> okay, so I'll ping you guys in #openstack-containers with what we learn. I'd love to have ansible as another deployment mechanism to add to our list.
16:49:39 <strigazi> fyi, we use puppet
16:49:50 <adrian_otto> thanks to you both!
16:49:51 <adrian_otto> #topic Open Discussion
16:50:34 <Drago> I would like reviews on https://review.openstack.org/#/c/338535/ and https://review.openstack.org/#/c/338537/
16:50:48 <Drago> It will complete the decouple-lbaas blueprint
16:50:55 <strigazi> :)
16:51:21 <adrian_otto> whoot!
16:51:48 <Drago> Also, please give your input on whether you think this should be backported to mitaka in https://review.openstack.org/#/c/337390/ and https://review.openstack.org/#/c/336702/
16:51:48 <tango> will do
16:53:48 <muralia> Also, please start adding topics to the mid-cycle etherpad
16:53:54 <adrian_otto> Drago: 337390 is not passing gate
16:54:01 <Drago> https://etherpad.openstack.org/p/magnum-newton-midcycle-topics
16:54:18 <adrian_otto> looks like Abishek's revisions may have caused that?
16:55:11 <Drago> adrian_otto: 337390 is against the stable/mitaka branch
16:55:26 <adrian_otto> yes, I see that
16:55:37 <Drago> I don't really know where it is code wise. I'm of the mind that it shouldn't be backported
16:55:52 <strigazi> +1
16:55:58 <adrian_otto> then why did you submit it for review?
16:56:03 <Drago> I did not
16:56:38 <Drago> It lists me as the author, but I do not own that patch (Owner: Abishek Chanda)
16:56:46 <Drago> *Abhishek
16:56:49 <adrian_otto> oh, that's confusing
16:57:10 <adrian_otto> is it becuase he copied the commit from the master branch as a backport?
16:57:25 <Drago> I think so
16:57:30 <adrian_otto> I see that he uploaded patch set 1. Got it.
16:59:14 <adrian_otto> ok, thanks everyone for attending. Our next meeting will be Tuesday 2016-07-19 at 1600 UTC. I will be on vacation.
16:59:32 <adrian_otto> #endmeeting