15:00:10 <jruzicka> #startmeeting RDO meeting - 2017-08-23
15:00:11 <openstack> Meeting started Wed Aug 23 15:00:10 2017 UTC and is due to finish in 60 minutes.  The chair is jruzicka. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:12 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:14 <openstack> The meeting name has been set to 'rdo_meeting___2017_08_23'
15:00:37 <dmsimard> dprince: let's continue after the meeting
15:00:49 <dprince> dmsimard: sure
15:00:49 <jruzicka> #topic hello everyone
15:00:54 <jpena> o/
15:00:56 <jruzicka> o/
15:00:57 <rbowen> o/ Yo
15:00:59 <ykarel> o/
15:01:12 <dprince> hi
15:01:24 <amoralej> o/
15:01:33 <chandankumar> \o/
15:01:54 <jruzicka> feel free to add agenda items: https://etherpad.openstack.org/p/RDO-Meeting
15:02:02 <dmsimard> \o
15:02:21 <jruzicka> #chair dmsimard dprince jpena rbowen ykarel amoralej chandankumar
15:02:22 <openstack> Current chairs: amoralej chandankumar dmsimard dprince jpena jruzicka rbowen ykarel
15:02:26 <number80> o.
15:02:39 <number80> a lot of people today
15:02:39 <trown> o/
15:02:59 <jruzicka> and lots of code reviews flying around as well :)
15:03:04 <chandankumar> people has returned back from holidays :-)
15:03:30 <jruzicka> #chair number80
15:03:31 <openstack> Current chairs: amoralej chandankumar dmsimard dprince jpena jruzicka number80 rbowen ykarel
15:04:32 <jruzicka> #chair trown
15:04:33 <openstack> Current chairs: amoralej chandankumar dmsimard dprince jpena jruzicka number80 rbowen trown ykarel
15:05:02 <jruzicka> allright, let's get started
15:05:08 <jruzicka> #topic Pike release notes
15:05:12 <rbowen> #link https://github.com/redhat-openstack/website/blob/master/source/blog/2017-09-01-rdo-pike-released.html.md
15:05:18 <rbowen> That's the draft of our Pike release notes.
15:05:28 <rbowen> jpena has started with some things.
15:05:42 <rbowen> Please take a moment to reflect on what has been done in this cycle, and add something on there.
15:05:55 <rbowen> Also, I'd like to include a complete list of who has contributed, but I'm not sure how to get Git to tell me that.
15:06:19 <rbowen> If someone knows how to do that, please let me know, or just go ahead and do it.
15:06:32 <rbowen> It would be awesome if we end up with a script that does this for next time. Hint, hint.
15:06:41 <rbowen> /EOL
15:07:01 <jruzicka> #action everyone to help with Pike release notes: https://github.com/redhat-openstack/website/blob/master/source/blog/2017-09-01-rdo-pike-released.html.md
15:07:43 <jruzicka> as we don't have much agenda, we can stare into the doc and wonder how we can help for a while :)
15:08:20 <rbowen> Also see the excellent release notes for Ocata. That's what we're going for.
15:09:26 <jruzicka> we're doing a lots of great work, spare some time to raise visibility of your awesome contributions ;)
15:09:41 <rlandy> dmsimard: hello - see the last openstack manual run - you'll see all four master OVB jobs are green (including a tempest run)
15:10:08 <jruzicka> #topic PTG video interviews
15:10:22 <rbowen> Another one that's really just an announcement.
15:10:24 <rbowen> #link https://docs.google.com/spreadsheets/d/1KNHuo9Yb5kbjZAYGQ_PAo-YFndD8QTdaKzaPoct_aaU/edit#gid=0
15:10:38 <rbowen> If you'll be at the PTG, I want to talk with you about your involvement in the RDO project.
15:10:56 <rbowen> I'm doing a series of video interviews, to promote what's new in Pike.
15:11:01 <rbowen> Like I did in Atlanta.
15:11:08 <dmsimard> rlandy: yay ?
15:11:15 <dmsimard> let's chat after the meeting :)
15:11:31 <rbowen> Also, if you're at the PTG, I'll be paying the bar tab one night, like in Atlanta. Still waiting for details about the official evening event, so we can pick a different evening. :-)
15:11:48 <trown> sweet
15:11:53 <rlandy> dmsimard: sure - next we will have to check the stable/release jobs
15:12:08 <rbowen> That's all I've got.
15:12:41 <jruzicka> cool
15:12:44 <apevec> can we have a quick status on pike branching?
15:12:59 <jruzicka> #topic quick status on pike branching
15:13:04 <amoralej> sure
15:13:06 <apevec> we are still bumping reqs afaict
15:13:08 <jruzicka> #chair apevec
15:13:09 <openstack> Current chairs: amoralej apevec chandankumar dmsimard dprince jpena jruzicka number80 rbowen trown ykarel
15:13:19 <amoralej> we are doing some reqchecks
15:13:21 <rbowen> There. It's on the agenda. :-)
15:13:32 <number80> and some branching is still happening
15:13:39 <amoralej> yes and building
15:13:49 <apevec> #link https://review.rdoproject.org/r/#/q/topic:pike-branching+status:open
15:13:50 <amoralej> most libraries and clients are already built in cbs
15:14:02 <chandankumar> I have updated requirements sync for tempest plugins
15:14:04 <amoralej> #link https://review.rdoproject.org/etherpad/p/pike-release-preparation
15:14:11 <chandankumar> once merged branching will be done
15:14:13 <amoralej> apevec, ^ has status updates
15:14:13 * jruzicka afk 2 mins
15:14:29 <apevec> thanks
15:14:54 <amoralej> gimme a while and i'll give you some numberes
15:15:18 <apevec> ok, I see I need to re-try requests updated
15:15:31 <amoralej> we have 115 branched distgit
15:15:34 <rdogerrit> morsik proposed openstack/neutron-dynamic-routing-distgit rpm-master: Add neutron-dynamic-routing package  https://review.rdoproject.org/r/7935
15:15:37 * Duck o/
15:15:42 <amoralej> and most of them are already built
15:15:56 <apevec> last attempt was failing unittests iirc
15:16:21 <apevec> amoralej, 115 out of total=?
15:16:34 <amoralej> from what's not branched, we have 92 puppet modules
15:16:40 <amoralej> and 77 non-puppet
15:16:51 <amoralej> from those 77 we have reqcheck done for around 40
15:16:59 <amoralej> that i'd like to get branched today
15:17:23 <amoralej> that reminds me, please review https://review.rdoproject.org/r/#/q/topic:pike-branching+status:open+sync
15:17:42 <amoralej> we have glare specs merged and package reviews approved
15:17:56 <amoralej> but maintainer didn't send rdoinfo review to remove under-review
15:18:12 <amoralej> i'll send it
15:18:28 <amoralej> and we are reviewing packages for karbor, i don't know if we will arrive on time
15:18:47 <apevec> amoralej, try contact him, maybe he doesn't want it in Pike ?
15:18:55 <amoralej> i already tried
15:19:01 <jruzicka> rdopkg reqcheck is helping, woohooo \o/
15:19:03 <amoralej> mfedosin, are you around?
15:19:21 <apevec> is anything using Glare?
15:19:22 <amoralej> so, that's the main status
15:19:26 <adarazs> jpena: I reviewed the /api/promotions change -- looks good as far as I can tell, just found a small docs mistake. thanks for posting the link yesterday. I hope we'll have the client updated for it soon as well so that we can use that. will the first returned result be the last promotion?
15:19:30 <apevec> amoralej++ thanks
15:19:32 <amoralej> dunno
15:19:44 <amoralej> apevec, we already moved to queens queens-uc tags in rdoinfo, btw
15:19:53 <jruzicka> on related note, is `rdpokg reqcheck` output included in jobs as discussed?
15:20:19 <amoralej> jruzicka, i have a different opinion about how to automate dependencies
15:20:19 <apevec> amoralej, yeah, I'd say glare is missing Pike if maintainer is absent
15:20:26 <rdogerrit> Merged openstack/neutron-lbaas-dashboard-distgit rpm-master: Requirements sync for pike  https://review.rdoproject.org/r/8779
15:20:34 <jruzicka> amoralej, I'm all ears ;)
15:20:38 <jruzicka> well, eyes...
15:21:13 <amoralej> maybe not the best moment, but, short version is that most of the times updating the spec is not the most important point when updating dependincies
15:21:18 <amoralej> but building them
15:21:18 <PagliaccisCloud> o/
15:21:26 <amoralej> that should be the starting point to automate dependencies
15:21:52 <amoralej> so when a dependency is updated in global-requirements.txt, we should automate to build it
15:22:02 <amoralej> via... i dunno, maybe patch to rdoinfo
15:22:19 <apevec> yeah. maybe
15:22:28 <amoralej> i think that should be the starting point to automate dependencies, not updating specs
15:23:02 <jruzicka> that's one step ahead
15:23:19 <jruzicka> just including rdpokg reqcheck output in jobs would be an improvement, no?
15:23:20 <rdogerrit> Merged openstack/keystone-tempest-plugin-distgit rpm-master: Requirements sync for pike  https://review.rdoproject.org/r/8782
15:23:20 <amoralej> it's not usefull to update the spec if there is no build
15:23:20 <rdogerrit> Merged openstack/designate-tempest-plugin-distgit rpm-master: Requirements sync for pike  https://review.rdoproject.org/r/8783
15:23:24 <rdogerrit> Merged openstack/tempest-horizon-distgit rpm-master: Requirements sync for pike  https://review.rdoproject.org/r/8784
15:23:26 <rdogerrit> Merged openstack/patrole-distgit rpm-master: Requirements sync for pike  https://review.rdoproject.org/r/8785
15:23:26 <jruzicka> automatically acting is a next step
15:23:55 <jruzicka> I'm talking about having up-to-date reqcheck easily available
15:25:15 <jruzicka> apevec, amoralej: so some info to sum up?
15:25:18 <amoralej> we could publish it in the repo dir for each dlrn build
15:25:29 <rdogerrit> Merged openstack/neutron-vpnaas-distgit rpm-master: Requirements sync for pike  https://review.rdoproject.org/r/8780
15:25:31 <jruzicka> amoralej, yeah, something like that
15:26:13 <amoralej> jruzicka, i think we need to think in the automation for the full workflow and start implementing step-by-step, but i'm not sure how having a reqcheck there would help, to be honest
15:26:14 <dmsimard> publish what, the output of reqcheck ?
15:26:29 <amoralej> yeah
15:26:46 <jruzicka> allright :)
15:26:54 <rdogerrit> Merged openstack/neutron-lbaas-distgit rpm-master: Requirements sync for pike  https://review.rdoproject.org/r/8778
15:27:11 <dmsimard> eh, we need something like automated upper constraints bumps which is where reqcheck might help
15:27:22 <jruzicka> it was just a suggestion, I'm not a very active user of the workflow anymore, really.
15:27:29 <dmsimard> a post job on upstream project merge that runs reqcheck if there was a change in reqs.txt
15:28:58 <amoralej> dmsimard, let's think there is a bump in u-c in oslo-config, we really don't need to bump it in every spec
15:29:11 <amoralej> what we need is to build that new version, and that's what we have
15:29:13 <amoralej> right now
15:29:17 <dmsimard> amoralej: that's not what I mean
15:29:36 <jruzicka> monitoring u-c and building accordingly sounds worth investigation
15:29:37 <dmsimard> amoralej: I mean if a patch in nova lands that modifies reqs.txt, we run reqcheck just to match if there's anything relevant to us that changed
15:29:48 <dmsimard> amoralej: flepied had a POC on that like 6 months ago I think
15:30:28 <dmsimard> amoralej: https://github.com/fredericlepied/rdo-packaging-failure-detection
15:30:38 <dmsimard> almost a year ago actually :P
15:31:10 <dmsimard> it was around the time we had started implementing DLRN-rpmbuild as a third party job to help warn us of incoming missing deps
15:31:15 <number80> Issue with requirements.txt and spec requirements is that there is no 1:1 mapping
15:31:22 <number80> we still need human review
15:31:25 <dmsimard> but we never really had a good system for notifying us of those failures
15:31:40 <dmsimard> number80: so what if it's a 0.80:1 mapping, it's better than nothing :)
15:32:02 <number80> dmsimard: it's not even up to that ratio depending the packages
15:32:06 <sfbender> Javier Peña created dlrnapi_client master: Add support for the new /api/promotions-get call  https://softwarefactory-project.io/r/9338
15:32:16 <dmsimard> number80: unless you're saying humans are 1:1 and have infinite time
15:32:24 <jruzicka> I was writing in detail about the problem of reqs.txt -> Requires: in some mail, I believe solving that problem is a key to all others.
15:32:59 <dmsimard> jruzicka: relevant (from 2 years ago) https://trello.com/c/woaK75me/132-create-a-tool-to-match-requirements-between-reqstxt-and-specfiles
15:32:59 <jruzicka> *most others we have
15:33:00 <number80> dmsimard: nope, but we need better log analysis tooling.
15:33:00 <dmsimard> :P
15:33:00 <openstackgerrit> Merged openstack/packstack master: Disable ceilometer middleware by default  https://review.openstack.org/494645
15:33:06 <dmsimard> number80: ++
15:33:34 <number80> well, I'm not against experimenting anyway or be proven wrong
15:33:39 <amoralej> dmsimard, i'll take a look on that but we shouldn't be adding a lot of info that nobody will check :)
15:33:44 <jruzicka> note that I already wrote pymod2pkg and renderspec that deal with it but much more work is needed for full automation
15:33:46 <dmsimard> number80: what I'm saying is that if we have reqcheck complaining, we can have humans review the reqcheck complaints -- rather than have humans review *everything* manually
15:34:00 <jruzicka> furthermore, someone would still need to maintain the reqs.txt -> Requires map
15:34:02 <jpena> number80: with pymod2pkg we're slowly getting to that 1:1 mapping
15:34:34 <jpena> if we try to use it, it'll take us some time to map the missing modules, but it could stabilize quickly
15:34:47 <number80> jpena: I mean, requirements.txt is missing information like: optional deps, sometimes upstream mixes build time/runtime/test-only deps etc..
15:34:57 <jpena> number80: ohhh those
15:35:00 <number80> yep
15:35:02 <amoralej> there are a lot of corner cases in requirements.txt, we should be careful with automatic updates
15:35:09 <number80> pymod2pkg is good tool though
15:35:18 <jpena> yes, we'll need human reviews anyway
15:35:39 * number80 thinks that now we have minimal eyeball coverage for these :)
15:36:14 <jruzicka> number80, we need to describe all these cases in pymod2pkg or elsewhere but that's doable
15:37:06 <number80> jruzicka: pymod2pkg can't guess if a requirement is only there to satisfy an obscure cinder backend that nobody uses except 2 or 3 people having very specific hardware
15:37:31 <jruzicka> We simply need a reliable, maintainable and complete map of reqs->Requires. It is now maintained redundantnly in individual packages an moreso in the heads of maintainers
15:37:58 <jruzicka> number80, of course, there will always be per-package overrides
15:38:02 <number80> we can off course bundle everything, but we also have to think that these RPMs will be now consumed by docker images where size matters
15:38:12 <jruzicka> but things that are same for all packages should be in central map
15:38:44 * number80 wonders how we dockerized installation will manage optional backends btw
15:38:44 <jruzicka> then you just put an optional file in the distgit with the irregularities for specific projects.
15:38:55 <jruzicka> you combine the two and boom, automation :)
15:39:14 <number80> jruzicka: could be something like that
15:39:23 <jruzicka> RPM now has weak requires, no?
15:39:40 <amoralej> dmsimard, i just added a topic on behalf of you
15:39:51 <dmsimard> ಠ_ಠ
15:39:59 <dmsimard> oh, sure
15:40:09 <amoralej> jruzicka, i think weak requires are not in centos7, right?
15:40:26 <jruzicka> amoralej, not but it's gonna take a while until the map is complete :)
15:40:27 <dmsimard> I briefly discussed it with apevec to bring him up to speed so we can definitely chat about it when we're done with this topic :)
15:42:11 <jruzicka> I think we're done... this is an ongoing goal and discussion.
15:42:13 <number80> jruzicka: not on EL7
15:42:42 * jpena just added another topic
15:43:02 <jruzicka> #topic We still haven't done any work towards shipping stable release of containers
15:43:17 <jruzicka> dmsimard, stage is yours ^
15:43:44 <rdogerrit> Alfredo Moralejo proposed openstack/tacker-distgit rpm-master: Requirements sync for pike  https://review.rdoproject.org/r/8765
15:43:47 <dmsimard> I don't really have anything to add :/
15:44:05 <jruzicka> heh, well observed phenomena of RDO meeting emergent agenda
15:44:11 <dmsimard> We need to rope in folks like dprince, EmilienM or mwhahaha
15:44:18 * mwhahaha hides
15:44:32 <dmsimard> apevec: you around ?
15:46:09 <dmsimard> mandre: you there ?
15:46:29 <dmsimard> ¯\_(ツ)_/¯
15:46:38 <jruzicka> ¯\_(ツ)_/¯
15:46:41 <amoralej> not luck
15:46:50 <jruzicka> next topic? not much time left.
15:46:52 <dmsimard> Basically, we need to agree on who's doing what and where it will end up
15:47:01 <number80> dmsimard: mailing list
15:47:03 <amoralej> The question is "who/how/where must create and publish container images using cloudsig RPMS usable for tripleo"?
15:47:03 <jruzicka> oh but next is just annoucement... nevermind
15:47:16 <dmsimard> amoralej: that's a good summary
15:47:31 <jruzicka> #info The question is "who/how/where must create and publish container images using cloudsig RPMS usable for tripleo"?
15:47:54 <dmsimard> I can take it to the mailing list if necessary, I've just been raising red flags and warnings
15:48:33 <dmsimard> Ok, I'll do that.
15:50:19 <jruzicka> #topic Everything .rdoproject.org (except planet?) migrated to RDO Cloud
15:50:30 <jruzicka> jpena, ^
15:50:37 <jpena> thanks
15:50:56 <jruzicka> nice
15:50:57 <jpena> yesterday, blogs.rdoproject.org was migrated to RDO Cloud. With that, we've mostly completed our migration
15:51:03 <rbowen> There was a discussion a few weeks back about folks blogging about RDO Cloud and what's there. Can someone do a writeup of the .rdoproject.org stuff that's on RDO Cloud, and why it's interesting? Just a couple paragraphs would be fine.
15:51:45 <jpena> Looking at the DNS entries, I see planet.rdoproject.org is "somewhere", but not sure if that is still active
15:51:54 <jpena> rbowen, is that machine still valid?
15:52:06 <rbowen> It is, but it's really just a tool that only I use.
15:52:09 <rbowen> It's on my personal machine.
15:52:14 * jruzicka likes writing blags but he doesn't know much about RDO Cloud
15:52:23 <jpena> ok, so no need to care about it, I guess
15:52:32 <rbowen> It's a simple Venus/Planet install, and could live anywhere, but I'm happy to keep running it.
15:52:48 <dmsimard> rbowen: oh yeah I had an etherpad for that
15:52:57 * dmsimard looks
15:53:05 <jpena> let's keep it as-is then. If you want to migrate it anytime, just ping me
15:53:23 <rbowen> Thanks
15:53:39 <dmsimard> rbowen: https://review.rdoproject.org/etherpad/p/rdo-cloud-blog-series
15:54:00 <rbowen> Yeah, that was it. I knew I'd seen that somewhere.
15:54:12 <rbowen> Should I just start bugging people about that, then? :-)
15:54:45 <rbowen> I can interviews, as suggested there, to facilitate.
15:55:01 <dmsimard> rbowen: yeah, I have too little bandwidth for that :(
15:55:09 <rbowen> We're out of time, though. I'll take it to the mailing list, and see what we can make happen.
15:55:58 <rdogerrit> Merged openstack/tacker-distgit rpm-master: Requirements sync for pike  https://review.rdoproject.org/r/8765
15:56:15 <jruzicka> #topic Chair for next meeting
15:56:51 <ihrachys> dmsimard, sorry for late response. still need help?
15:58:07 <jpena> if nobody else volunteers, I'll chair the next meeting
15:58:20 <jruzicka> #info jpena to chair next RDO meeting
15:58:27 <jruzicka> #topic open floor
15:59:13 <jruzicka> someone with mad python3 RPM packaging skillz please eye my rdopkg py3 attempt ;) https://softwarefactory-project.io/r/#/c/9332/
15:59:48 <amoralej> i'll be off for next three weeks, if i owe you something, let me know asap
16:01:42 <jruzicka> #endmeeting