21:01:00 <adrian_otto> #startmeeting Solum Team Meeting
21:01:00 <openstack> Meeting started Tue Mar 24 21:01:00 2015 UTC and is due to finish in 60 minutes.  The chair is adrian_otto. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:01:01 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:01:04 <openstack> The meeting name has been set to 'solum_team_meeting'
21:01:21 <adrian_otto> #topic https://wiki.openstack.org/wiki/Meetings/Solum#Agenda_for_2015-03-24_2100_UTC Our Agenda
21:01:39 <adrian_otto> #topic Roll Call
21:01:40 <gpilz> Gilbert Pilz
21:01:40 <datsun180b> ed cranford
21:01:44 <adrian_otto> Adrian Otto
21:01:46 <muralia2> murali allada
21:01:48 <james_li> james li
21:01:51 <devkulkarni> devdatta
21:02:06 <mkam> Melissa Kam
21:03:20 <adrian_otto> hi gpilz, datsun180b, muralia2, james_li, devdatta, mkam
21:03:30 <muralia2> hi adrian
21:03:44 <adrian_otto> #topic Announcements
21:04:09 <adrian_otto> 1) The Solum speaking session was not chosen for the OpenStack Summit in Vancouver….
21:04:20 <dimtruck> Dimitry Ushakov
21:04:34 <gpilz> shoot
21:04:48 <adrian_otto> however, I do expect there is still a place for Solum at the Design Summit in the OpenSource at OpenStack track or "Other Projects" track.
21:06:06 <adrian_otto> 2) PTL Elections
21:06:21 <adrian_otto> community projects are not required to hold PTL elections, but we do anyway.
21:07:08 <adrian_otto> in the past we have tried not to conflict with open elections for OpenStack projects, as there has been suggestions that having dozens of concurrent unrelated elections is confusing for the electorate
21:07:39 <adrian_otto> typically the PTL is selected by the contributors in the prior two releases.
21:08:29 <adrian_otto> how would you all feel about having a call for candidacy that is sent to all of the contributors by direct email instead of using the openstack-dev wiki?
21:08:56 <datsun180b> That sounds fine to me
21:09:05 <devkulkarni> you mean openstack-dev mailing list? sounds fine
21:09:10 <adrian_otto> in the event that there is > 1 candidate, an election would be held.
21:10:05 <adrian_otto> we would have an agenda item in our next team meeting where candidacy could be announced by any challenger.
21:10:05 <adrian_otto> thoughts on this approach, versus an exact repeat of what we did in the last release cycle?
21:10:29 <adrian_otto> note that you can only run for the PTL position if you have contributed code to the prior release
21:10:29 <adrian_otto> I think we have something like 19 qualified individuals
21:10:37 <adrian_otto> devkulkarni: no
21:11:02 <adrian_otto> the specific request was that we not use openstack-dev@lists.openstack.org to request candidacy notices
21:11:10 <devkulkarni> I think no harm in trying this approach
21:11:16 <muralia2> ok
21:11:25 <adrian_otto> but instead send out direct email (by BCC) to individuals who have contributed
21:11:25 <gpilz> ok by me
21:11:31 <devkulkarni> right.. I just meant that in your earlier comment you had said openstack-dev wiki
21:11:39 <adrian_otto> oh, yes
21:11:55 <adrian_otto> the "wiki" meant that we would record the plan on our wiki on an election subpage
21:11:58 <adrian_otto> for posterity
21:12:21 <devkulkarni> ah, got it
21:12:26 <adrian_otto> the aim here is to balance the interests of open community and transparency with the sensitivities of running the OpenStack elections
21:13:02 <devkulkarni> on a similar note — there have been emails on the dev list about not using it to announce core reviewer additions as well
21:13:04 <adrian_otto> I will record this as an #agreed if there are no objections
21:13:10 <devkulkarni> +1
21:13:17 <muralia2> +1
21:13:24 <gpilz> +1
21:13:26 <adrian_otto> devkulkarni: no, core reviewer proposals may still use the list
21:13:32 <james_li> +1
21:13:37 <adrian_otto> this is about selection of project PTLs
21:13:59 <datsun180b> +1
21:14:03 <devkulkarni> right.. I am just saying that I have recently seen emails where folks have argued against using the dev mailing list to announce reviewer additions.
21:14:26 <adrian_otto> yes, there is an alternate idea about that
21:14:30 <devkulkarni> in fact, I haven't seen emails that say that PTL candidacy emails should not be sent to the dev list.. may be I missed those.
21:14:37 <adrian_otto> basically to have a governance file in the code repo
21:14:42 <adrian_otto> and you vote in Gerrit
21:14:49 <devkulkarni> I see
21:15:00 <adrian_otto> there has also been talk about having a stackforge elections repo, but that is some time out
21:16:12 <adrian_otto> #agreed the solum PTL shall be selected by the project contributors. Contributors will be invited to announce candidacy at the 2015-03-31 team meeting. If no new candidates are announced during that section of the team meeting IRC agenda, the incumbent PTL will remain in service.
21:16:27 <adrian_otto> I can still undo that if there are any objections
21:17:02 <adrian_otto> ok, any other announcements from team members this week?
21:17:36 <adrian_otto> #topic Review Action Items
21:17:39 <adrian_otto> (none)
21:17:49 <adrian_otto> #topic Blueprint/Task Review
21:17:54 <devkulkarni> adrian_otto: quick question.. do you have a link to the accepted talks for ODS?
21:18:07 <adrian_otto> yes. one moment
21:18:53 <adrian_otto> #link https://openstacksummitmay2015vancouver.sched.org/ OpenStack Summit Agenda
21:19:02 <devkulkarni> thanks adrian_otto
21:19:04 <adrian_otto> devkulkarni: is that what you meant?
21:19:13 <devkulkarni> yep
21:19:18 <adrian_otto> ok, great.
21:19:21 <adrian_otto> 1) (silkey) Pin requests to 2.4.3 https://review.openstack.org/#/c/165603/
21:19:30 <adrian_otto> is silkey present?
21:19:52 <devkulkarni> no.. but I can talk to that task
21:20:11 <adrian_otto> Nick Silkey made his first contribution to the OpenStack ecosystem in Solum
21:20:27 <adrian_otto> Jenkins disliked the proposal
21:20:37 <adrian_otto> devkulkarni: you have the floor
21:20:38 <devkulkarni> basically on the python-solumclient we are getting a SSL cert warning which can be prevented if we use
21:20:49 <devkulkarni> the specific version of requests
21:20:54 <devkulkarni> that is there in the review
21:21:15 <devkulkarni> however, global-requirements does not yet have the 2.4.3 version
21:21:44 <devkulkarni> we could use soft requirements probably
21:21:46 <adrian_otto> could we use "requests>=2.4.3" instead? If not, why not?
21:21:46 <datsun180b> I think in fact global's requests is ahead of 2.4.3, which is our problem
21:22:46 <datsun180b> After 2.4.3 a deprecation message for an SSL issue involving a missing value in a cert is shown with every use
21:22:49 <adrian_otto> so 2.5.0 breaks us
21:23:17 <devkulkarni> datsun180b: what version is global's requests?
21:23:32 <datsun180b> requests>=2.2.0,!=2.4.0
21:24:09 <james_li> soft requirements is for devstack, not sure a devstack is spun up for solum CLIENT tests
21:24:11 <datsun180b> and this isn't so much a problem for solum general as it is for rackspace but i could be wrong about that
21:24:14 <dimtruck> in the name of transparency, let's point out that this is an issue solely with Rackspace's certificates.  HP doesn't have this problem.
21:24:17 <dimtruck> datsun180b: +1
21:24:35 <dimtruck> HP cloud*
21:24:51 <dimtruck> since those are the two main cloud providers we use in upstream
21:24:52 <adrian_otto> there has been some discussion about adjusting all of our certs
21:25:04 <adrian_otto> so this problem may vanish if we take no action
21:25:12 <adrian_otto> are you all aware of this initiative?
21:25:14 <dimtruck> yes sir...but it's dependent on Rackspace's cloud identity
21:25:24 <dimtruck> there's no public timeline available
21:26:36 <adrian_otto> ok, so what I'm not sure about in this review is that all the gate tests are marked with SUCCESS, but Jenkins voted -1. I don't see why.
21:26:53 <datsun180b> strictly because that change to requests doesn't jive with global-reqs
21:27:01 <devkulkarni> adrian_otto: click on 'toggle ci'
21:27:18 <devkulkarni> you will see the requirements don't match issue
21:27:33 <adrian_otto> where is "toggle ci"?
21:27:46 <devkulkarni> at the very bottom of the review page
21:27:54 <devkulkarni> #link https://review.openstack.org/#/c/165603/
21:27:59 <adrian_otto> oh, I found it, thanks!!
21:29:34 <adrian_otto> #link http://logs.openstack.org/03/165603/2/check/gate-python-solumclient-requirements/6c8a53a/console.html#_2015-03-18_22_38_00_003 the gate objection
21:30:11 <adrian_otto> what options do we have to work around this?
21:30:17 <devkulkarni> we checked with rackspace's cloud orchestration team.. they have similar issue, but they are ignoring it for now
21:30:52 <adrian_otto> what is the community impact if we ignore this too?
21:31:08 <adrian_otto> probably none, right?
21:31:09 <devkulkarni> frankly, not much
21:31:12 <datsun180b> epsilon squared
21:31:13 <devkulkarni> yeah
21:31:42 <adrian_otto> ok, so I suggest we fix that in an internal patch for now if it's custome rimpacting
21:31:56 <adrian_otto> and leave upstream alone
21:32:06 <adrian_otto> fair enough?
21:32:12 <devkulkarni> sounds good to me
21:32:21 <adrian_otto> any other viewpoints to consider?
21:32:37 <datsun180b> just need to close silkey's patch then, and explain why
21:32:49 <adrian_otto> I will take care of that right now
21:33:05 <adrian_otto> while I do that, do team members have other work items to discuss today?
21:33:21 <datsun180b> thank you kindly
21:33:27 <devkulkarni> adrian_otto: I have one
21:33:40 <adrian_otto> proceed!
21:33:50 <devkulkarni> I remember we had discussed sometime back about release of python-solumclient on a periodic cadence
21:34:11 <devkulkarni> what do you think about doing a release on a specific day each week — at your convenience
21:34:49 <adrian_otto> yeah, I'm open to that
21:34:55 <devkulkarni> cool
21:35:00 <adrian_otto> in fact I am willing to cut a release on demand any time you all want.
21:35:36 <devkulkarni> that will be awesome as well.. whichever is the most convenient for you
21:35:39 <adrian_otto> I should be able to act on those sort of requests on the same day.
21:35:42 <datsun180b> is the process any more complicated than you tagging a release?
21:35:53 <adrian_otto> it's really easy
21:36:03 <adrian_otto> the only hard part is signing the release with GPG
21:36:03 <datsun180b> is it so easy it's documented somewhere
21:36:15 <adrian_otto> basides that it's a couple of git commands
21:36:20 <adrian_otto> git -t tag
21:36:23 <adrian_otto> git push gerrit
21:36:24 <devkulkarni> also, on that note, I remember you had also mentioned about making it such that the process is automated triggered on every merge
21:36:25 <adrian_otto> that's it.
21:36:50 <adrian_otto> devkulkarni: I think that may be possibe
21:37:16 <adrian_otto> we can discuss to find the right effort/benefit ratio
21:37:21 <devkulkarni> cool.. but probably we could do that later..
21:37:22 <devkulkarni> right
21:37:40 <adrian_otto> if I'm not fast enough doing it by request, then we should see about an automation approach
21:38:02 <datsun180b> would you need to deputize someone else with a permission you have that we lack?
21:38:03 <devkulkarni> :) sounds good.
21:38:40 <datsun180b> and, should it happen, would those permissions transfer to any future PTL, or would you still be the one to cut those releases?
21:39:14 <datsun180b> all i'm trying to do is reduce the number of "adrian plz cut a release kthx" emails in the future
21:39:30 <adrian_otto> the individual with the PTL role can tag releases
21:39:53 <adrian_otto> the only gotcha is that that individual is required to GPG sign the commit
21:40:56 <adrian_otto> it may also be possible to have multiple individuals to whom that access level is delegated.
21:41:05 <adrian_otto> we could explore that option as well
21:41:42 <devkulkarni> +1 to datsun180b.. +1 to adrian_otto
21:41:48 <adrian_otto> so ping me on IRC, and if no answer then email me. You all have my phone number too, you can SMS.
21:41:58 <adrian_otto> the only time I don't answer an SMS is if I'm in flight.
21:42:18 <datsun180b> i foresee a couple situations where you might be called upon more than once a day, just trying to test the structure before we have to put any real weight on it
21:42:37 <adrian_otto> that's fine with me
21:42:44 <devkulkarni> adrian_otto: the last four digits — are they 5417 or 6417
21:42:50 <adrian_otto> honestly, this process takes just a couple of minutes
21:43:11 <adrian_otto> devkulkarni: actually neither. I will PRIVMSG it to you.
21:43:19 <devkulkarni> ok.. sounds good
21:44:41 <devkulkarni> thanks adrian_otto
21:46:51 <adrian_otto> ok
21:46:59 <adrian_otto> #topic OpenDiscussion
21:47:08 <adrian_otto> #topic Open Discussion
21:47:40 <datsun180b> so take a look at this: http://status.openstack.org/zuul/ and filter for solum
21:47:51 <devkulkarni> oh right!! F20 gate
21:48:01 <devkulkarni> james_li: want to take this one?
21:48:08 <adrian_otto> ok, so I might also mention that Magnum was approved to join the OpenStack project list about an hour ago by unanimous vote of the TC. This might be a good time for Solum to think about any future plans for Magnum integration.
21:48:14 <datsun180b> 59 hours, 32 hours, 22 hours, 2 and a half hours
21:48:22 <james_li> just talked with infra guys
21:48:31 <james_li> the F20 node won't boot
21:48:32 <james_li> again
21:48:36 <datsun180b> any longer and our tasks are going to have to be portrayed by James Franco
21:48:46 <devkulkarni> adrian_otto: congratulations
21:48:52 <adrian_otto> thanks devkulkarni
21:48:56 <datsun180b> oh congratulations magnum
21:49:14 <muralia2> woohoo! congrats
21:49:17 <james_li> so they suggested we switch to F21 or centos, the latter is better in their opinion
21:49:19 <devkulkarni> will be nice to start discussing where/how magnum can fit in with solum.
21:49:31 <devkulkarni> james_li: oh is that so?
21:49:33 <dimtruck> nice!
21:49:36 <devkulkarni> so what are the next steps?
21:50:00 <adrian_otto> devkulkarni: for Magnum?
21:50:14 <james_li> since we use F20 for some historical reasons and it is now non-voting
21:50:25 <devkulkarni> adrian_otto: could one of the magnum folks write a spec of possible integration points .. will it replace heat within solum? will it sit next to heat?
21:50:26 <james_li> I would say switch to centos
21:50:39 <devkulkarni> adrian_otto: sorry.. I was asking for next steps to james_li
21:50:42 <adrian_otto> james_li: +1
21:51:08 <gpilz> i'm finding the parallel discussion of Magnum and the f20 gate confusing
21:51:12 <adrian_otto> devkulkarni: there are multiple ways it might fit
21:51:30 <adrian_otto> but let's revisit that later
21:51:34 <adrian_otto> there is no hurry
21:51:58 <devkulkarni> adrian_otto: sure
21:52:24 <devkulkarni> going back to f20
21:52:35 <devkulkarni> james_li: what is required to switch over to centos?
21:52:44 <devkulkarni> could we disable f20 first quickly?
21:53:00 <adrian_otto> I think that's in the project-config repo now
21:53:07 <datsun180b> it might help to explain the why before we worry about the if and the how
21:53:07 <devkulkarni> for centos gate — we will have to put in some efforts to see that everything works
21:53:41 <james_li> we need infra guys approval for any changes
21:53:50 <james_li> to gate
21:54:39 <devkulkarni> james_li: yes.. but from process point-of-view, what are the steps? 1) disable f20 2) test with centos 3) propose addition of centos
21:54:40 <devkulkarni> ?
21:55:13 <james_li> for centos we definitely needs infra guys' help for amending the patch to change gate config, thats some additional work
21:55:40 <james_li> for disabling F20 we can handle the patch by ourselves, just wait for their approval
21:56:03 <devkulkarni> got it.. so could we make the decision right now to disable f20? what do others think?
21:56:15 <datsun180b> i _think_ it's just a matter of removing a line or two from infra's project-config
21:56:25 <james_li> +1 for disabling F20 for now
21:56:36 <devkulkarni> datsun180b: yep.. the change should be straight forward.
21:56:42 <datsun180b> of course, testing that kind of change is super tough without feedback from the gate!
21:56:51 <devkulkarni> I am more concerned about we being a single flavor (Ubuntu) system in the time being
21:57:12 <james_li> later we can add a centos gate
21:57:13 <datsun180b> devkulkarni: counterarguments: git and docker
21:57:25 <adrian_otto> #link https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/solum.yaml Solum Gate Jobs
21:57:31 <devkulkarni> having said that — I do care about patches getting merged — so I am a +1 to turn off f20
21:57:32 <adrian_otto> I don't see the fedora job in there
21:58:05 <james_li> adrian_otto: gate-solum-devstack-dsvm-f20
21:58:15 <devkulkarni> adrian_otto: line 49
21:58:24 <adrian_otto> yep, that's what we need to patch
21:58:43 <adrian_otto> ok, time to wrap up here
21:58:47 <datsun180b> just removing that line? i'll pull the trigger on this review then
21:59:05 <adrian_otto> you remove that job entry, yes
21:59:20 <datsun180b> oh this diff doesn't say anything near a line 49
21:59:40 <adrian_otto> Our nextmeeting is 2015-03-31 at 2100 UTC. See you all then!
21:59:45 <adrian_otto> #endmeeting