15:00:06 <EmilienM> #startmeeting puppet-openstack
15:00:07 <openstack> Meeting started Tue Jan 12 15:00:06 2016 UTC and is due to finish in 60 minutes.  The chair is EmilienM. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:08 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:11 <openstack> The meeting name has been set to 'puppet_openstack'
15:00:15 <EmilienM> #link agenda URL: https://etherpad.openstack.org/p/puppet-openstack-weekly-meeting-20160112
15:00:19 <EmilienM> o/
15:00:23 <mwhahaha> hi
15:00:27 <chem> o/
15:00:27 <aderyugin> hi
15:00:51 <clayton> o/
15:01:01 <_ody> o/
15:01:01 <degorenko> hey o/
15:01:04 <iurygregory> hey
15:01:42 <EmilienM> #topic Review past action items
15:02:04 <EmilienM> the only action we haven't updated yet is " consider a rewrite design for https://bugs.launchpad.net/puppet-nova/+bug/1506066 "
15:02:06 <openstack> Launchpad bug 1506066 in puppet-nova "Scope-related warnings from nova::vncproxy::common" [Medium,In progress] - Assigned to Nate Potter (ntpttr)
15:02:27 <EmilienM> nate potter is around?
15:02:34 <EmilienM> mwhahaha: I guess we haven't found time on it right?
15:02:41 <mwhahaha> no
15:03:09 <richm> hello
15:03:41 <EmilienM> ok no prob
15:03:43 <EmilienM> we'll postpone it
15:03:56 <EmilienM> ok let's start the agenda
15:04:10 <EmilienM> #action postpone consider a rewrite design for https://bugs.launchpad.net/puppet-nova/+bug/1506066
15:04:12 <openstack> Launchpad bug 1506066 in puppet-nova "Scope-related warnings from nova::vncproxy::common" [Medium,In progress] - Assigned to Nate Potter (ntpttr)
15:04:25 <EmilienM> #topic Virtual sprint agenda / work items
15:04:33 <EmilienM> so for those who'll attend it
15:04:56 <EmilienM> #link sprint agenda https://etherpad.openstack.org/p/puppet-happy-new-year-2016
15:05:16 <EmilienM> we have some topics, maybe we could just put a name on each so we have a point of contact
15:05:34 <EmilienM> I volunteer to lead Reno (release management) work
15:05:57 <EmilienM> we have a topic for Mitaka deprecations & removals
15:06:14 <EmilienM> regading the color I'm not sure who wrote it, maybe mfisch or degorenko
15:06:16 <degorenko> i can help mfisch if he needs help :)
15:06:26 <degorenko> it was mfisch :D
15:07:37 <EmilienM> we also have bug triage
15:07:45 <clayton> mfisch said to say he's running late
15:08:02 <EmilienM> clayton: cool! thanks to let us know
15:08:23 <EmilienM> degorenko: maybe you can lead the deprecation thing and mfisch the bug triage like usual?
15:08:38 <degorenko> EmilienM, ok, no problem
15:09:38 <EmilienM> we have an item for having real CI on our puppet-openstack-cookiecutter CI
15:09:39 <_ody> I've got other things I could do but think Reno just needs to happen so I am happy to just follow someone on that.
15:10:01 <EmilienM> _ody: please suggest the other things!
15:10:02 <degorenko> EmilienM, +++1 for real ci for cookiecutter :)
15:10:30 <EmilienM> _ody: you can use the etherpad maybe
15:10:46 <EmilienM> who is volunteer to lead the puppet-openstack-cookiecutter work?
15:10:56 <degorenko> for last time we already have a few new modules and we faced with some problems, so real ci will really help
15:10:58 <richm> will there be cookies?
15:11:08 <EmilienM> richm: I'll send them to your place, yeah
15:11:10 <degorenko> and milk :)
15:12:02 <EmilienM> nobody?
15:12:12 <degorenko> i can do it :D
15:12:21 <EmilienM> degorenko: well, you already have an item
15:12:28 <degorenko> i can help somebody :)
15:12:31 <chem> I can give it a shot
15:13:01 <EmilienM> chem: cool! so the idea is to have a CI job that generate a module and run unit & functional tests
15:13:12 <chem> EmilienM: ack
15:13:14 <EmilienM> because for now, puppet-openstack-cookiecutte is running noop jobs
15:13:17 <EmilienM> which is useless
15:13:29 <EmilienM> chem: are you familiar with openstack-infra?
15:13:35 <chem> will need help on this side
15:13:46 <mwhahaha> i can help with the ci for cookiecutter
15:13:50 <EmilienM> nibalizer is infra core and can help a lot
15:13:56 <EmilienM> mwhahaha, chem : ok I put your names
15:14:00 <chem> oki thanks :)
15:14:10 <EmilienM> and we have a last item for puppet-ceph CI
15:14:21 <EmilienM> I think xarses (offline) is the best guy to lead it
15:14:36 <degorenko> agree
15:14:36 <EmilienM> feel free to contact him if you're interested to hack on it
15:15:05 <EmilienM> I think we're good for the agenda, _ody feel free to add more topics, but asap so we can groom and find people working on it
15:16:38 <EmilienM> #action _ody complete sprint agenda with more topics if needed
15:17:04 <EmilienM> anything else for the sprint? any question?
15:17:14 <EmilienM> let's move ahead then
15:17:21 <EmilienM> #topic OpenStack client for nova providers
15:17:28 <EmilienM> degorenko: o/
15:17:38 <degorenko> yep :)
15:17:39 <EmilienM> #link https://review.openstack.org/#/c/226862/
15:17:45 <degorenko> want discuss acceptance for my patch
15:18:18 <degorenko> there is sometime failures
15:18:33 <degorenko> because in each run we can have different order for resources
15:18:33 <EmilienM> random failures?
15:18:53 <EmilienM> degorenko: that's something you can fix in your types, right?
15:19:00 <degorenko> well
15:19:11 <degorenko> i'm not sure about fixing it in types
15:19:26 <EmilienM> degorenko: can you give more context?
15:19:28 <degorenko> but as way, it can be, yes
15:19:32 <degorenko> let me sec
15:20:23 <degorenko> we can use requires in types, but it will not work, if we want run it separatly from deployment
15:20:26 <degorenko> is it ok?
15:20:55 <degorenko> so, we can use some require for nova
15:21:00 <degorenko> but we can't use require on keystone
15:21:21 <degorenko> only if in acceptance file require for resource in place
15:21:24 <EmilienM> autorequire will find the resource in the catalog
15:21:50 <degorenko> yes, but if we will deploy nova in one node and keystone on another one - it will not work
15:22:04 <EmilienM> yep
15:22:04 <_ody> ...and autorequires will only set up the relationship if it finds the resources
15:22:15 <EmilienM> right, it won't fail if no find
15:22:16 <degorenko> so, failures here are: can't find nova package/api whatever
15:22:26 <degorenko> or, keystone api is not available
15:22:40 <degorenko> the second one is more important
15:22:43 <_ody> ah, different problem.
15:22:52 <degorenko> because we can use cross project autorequire
15:22:53 <EmilienM> that's why healthcheck would be a nice use case here
15:22:58 <EmilienM> we discussed that at the summit
15:23:21 <degorenko> well, yes
15:23:27 <degorenko> but we don't have it now :D
15:23:45 <degorenko> as workaround we can use autorequire for nova in types
15:23:53 <degorenko> and require for keystone in acceptance
15:23:55 <degorenko> is it ok?
15:24:06 <degorenko> require for nova_aggregate resource only, of course
15:24:20 <_ody> seems fine
15:24:24 <EmilienM> +1
15:24:33 <_ody> the latter issue is a puppet problem
15:24:42 <degorenko> ok, that's all what i want to know :) i'll update my patch
15:25:17 <_ody> it is after the new year...time to start bugging product about application orchestrator againg
15:26:01 <_ody> (also another way to fix the split node issue)
15:26:31 <EmilienM> _ody: is it open-source?
15:26:57 <_ody> it is suppose to be.  I have need to bug them about release dates again now that we are in the new year.
15:27:02 <EmilienM> we need something that people can use without puppet master, without hiera and on Puppet 3 or 4
15:27:20 <EmilienM> well puppet master is another story though
15:27:34 <EmilienM> #topic Open Discussion, Bug and Review triage
15:27:40 <EmilienM> submit modules to triage here
15:28:00 <EmilienM> degorenko and chem submitted bunch of patches
15:28:03 <EmilienM> you want to discuss?
15:28:07 <chem> yep
15:28:14 <degorenko> no, just some of them stucked :)
15:28:27 <degorenko> because of ci or anything else
15:28:34 <chem> this one is critical and stuck "Neutron parses the wrong json revert to shell/csv."
15:28:51 <mfisch> here finally
15:28:53 <mfisch> commute took forever
15:29:05 <chem> I got a -1 from Sergey Valsilenko, but I'm not sure he got why I did the patch
15:29:06 <EmilienM> degorenko: do we have acceptance for murano?
15:29:21 <degorenko> EmilienM, not for dashboard
15:29:46 <degorenko> because of lack this package in repos
15:29:52 <degorenko> for liberty though
15:30:40 <EmilienM> degorenko: https://review.openstack.org/#/c/254181/ is not good
15:31:09 <degorenko> oops, missed it :)
15:31:42 <EmilienM> someone has to investigate https://review.openstack.org/#/c/242718/
15:31:45 <EmilienM> it fails
15:31:49 <EmilienM> I can look later though
15:32:05 <EmilienM> but failure on gate-tripleo-ci-f22-nonha is not something we need to consider "normal"
15:32:21 <EmilienM> #link TripleO CI status http://tripleo.org/cistatus.html
15:32:40 <degorenko> but it always failed actually for this patch
15:32:46 <EmilienM> yeah
15:32:51 <EmilienM> so the patch has something :)
15:33:25 <EmilienM> http://logs.openstack.org/18/242718/19/check-tripleo/gate-tripleo-ci-f22-nonha/2c65c80/console.html#_2016-01-12_10_06_03_556
15:33:25 <degorenko> that's strange
15:33:43 <EmilienM> it's obvious it breaks tripleo, glance does not look started
15:33:48 <EmilienM> let's look at overcloud logs
15:34:04 <EmilienM> undercloud sorry
15:34:15 <EmilienM> I'll take it offline
15:34:18 <EmilienM> anything else for today?
15:34:28 <chem> well the patch about neutron
15:34:44 <chem> https://review.openstack.org/#/c/262809/
15:34:45 <EmilienM> chem: Sergey does not seem online now, damn
15:35:02 <chem> ack, I will have to discuss with him at some point
15:35:05 <degorenko> xenolog - offline :(
15:35:29 <mattymo> degorenko, I sit near him.. he's in another meeting
15:35:40 <mattymo> can you guys discuss after the meeting with him?
15:35:44 <chem> basically neutron is currently broken, waiting secretly to blow off
15:35:45 <degorenko> oh, thanks
15:36:22 <chem> EmilienM: what is Sergey nick ?
15:36:42 <degorenko> chem, xenolog
15:36:50 <degorenko> as i mentioned above :)
15:36:53 <chem> ah, oki, make sense
15:37:00 <aglarendil> chem: well, I can be an avatar of Sergey today, because I agree with him
15:37:06 <aglarendil> his nick is @xenolog, btw
15:37:26 <chem> aglarendil: the problem is not json or not json, the problem is wrong json
15:37:32 <aglarendil> why would we revert the patch instead of fixing the output format in json parser?
15:37:40 <chem> aglarendil: cannot be done
15:37:50 <chem> aglarendil: you have to parse 2 json output
15:38:02 <chem> aglarendil: and one of them is a bug
15:38:28 <aglarendil> chem so you mean that neutron server returns some of the stuff in one format and some of the stuff in another format ?
15:38:50 <mwhahaha> yea it has to do with the available json formatters on the system, right?
15:38:50 <chem> aglarendil: it depends on the order of the json plugin load
15:39:01 <chem> aglarendil: mwhahaha yep
15:39:18 <aglarendil> so is it possible to somehow detect which format we are dealing with and use the right parser?
15:39:36 <chem> aglarendil: rather difficult
15:39:51 <chem> aglarendil: and you have to support 2 parser
15:40:07 <chem> aglarendil: the Field, Value and the plain json
15:40:31 <chem> aglarendil: they don't have the same start tag neither, so output detection is going to be hairy as well
15:40:42 <aglarendil> well, can we just fix neutron? for example, add an option to request particular format?
15:40:52 <chem> aglarendil: cannot be done neither ...
15:41:08 <clayton> even if you could, how likely would it be to be merged for mitaka?  And that doesn't help for backports
15:41:10 <chem> aglarendil: http://lists.openstack.org/pipermail/openstack-dev/2015-December/083172.html
15:41:54 <chem> aglarendil: it's done automatically by a plugin and depend on ... inode number or something :)
15:42:03 <chem> aglarendil: not user configurable
15:42:37 <EmilienM> +1 with clayton, we need a workaround in puppet-neutron I guess
15:43:02 <EmilienM> we need to help neutron upstream to be fixed but in the meantime workaround in our module.
15:43:27 <chem> for mitaka, clifftab dependency is going to disapear from openstackclient dependency and the neutron module will be broken
15:43:36 <chem> as it parse json from clifftab
15:44:01 <EmilienM> chem: it's already broken, see https://etherpad.openstack.org/p/puppet-openstack-ci-mitaka
15:44:07 <EmilienM> scenario002 on trusty
15:44:13 <EmilienM> failing to manage neutron subnets
15:44:14 <chem> EmilienM: that's normal
15:44:21 <EmilienM> not sure that's 100% related but it might be
15:44:23 <mwhahaha> so we switched away from csv for a reason, are we going to be reintroducing the issue by going back?
15:44:33 <xarses> mwhahaha: yes
15:44:33 <chem> cliff json != clifftab json
15:44:45 <chem> for me it break idempotency
15:44:54 <xarses> the csv order changed and was not guaranteed
15:44:56 <chem> but creation is still ok
15:45:04 <aglarendil> yeah, I read that already. so, I would pass the stuff we discussed to Sergey. I guess there is currently no way than to go with revert until neutron itself is fixed.
15:45:11 <EmilienM> right, it tried to create again
15:45:35 <xarses> can we at least use a CSV parser instead of regex?
15:45:36 <chem> EmilienM: that's the behavior with the broken parser.
15:45:50 <chem> xarses: the csv parser is used
15:45:52 <EmilienM> chem: cool so I confirmed the bug with mitaka packaging :)
15:46:08 <chem> EmilienM: not packaging, packaging is ok
15:46:09 <EmilienM> chem: this ? http://logs.openstack.org/35/255635/15/check/gate-puppet-openstack-integration-scenario002-tempest-dsvm-trusty/ca1990b/console.html#_2016-01-12_01_32_00_817
15:46:40 <chem> xarses: the regex is there to differenciate the error the client can output from the beginning of the real output
15:46:51 <chem> xarses: the same kind of stuff is used currently
15:47:04 <chem> EmilienM: looking
15:47:14 <chem> EmilienM: exactly
15:47:30 <EmilienM> ok cool
15:47:43 <chem> EmilienM: packager have remove clifftab as it's no longer required by openstackclient
15:47:51 <chem> see it requirements.txt
15:47:54 <chem> see its requirements.txt
15:48:31 <chem> EmilienM: or if it's still there it's just that the cliff parser is loaded after the clifftab parser
15:48:38 <chem> EmilienM: random :)
15:49:20 <mwhahaha> so the original bug was due to warning screwing up parsing
15:49:29 <EmilienM> should I test again with your patch in Depends-On and see how it behaves?
15:49:31 <mwhahaha> https://bugs.launchpad.net/puppet-neutron/+bug/1508511
15:49:32 <openstack> Launchpad bug 1508511 in puppet-neutron "providers, like neutron_network, neutron_subnet, fails if Neutron CLI throw warnings" [Medium,Fix committed] - Assigned to Sergey Vasilenko (xenolog)
15:49:38 <chem> mwhahaha: this is fixed there https://review.openstack.org/#/c/263874/3
15:49:44 <chem> mwhahaha: on top of the revert
15:50:10 <EmilienM> chem: can I try with https://review.openstack.org/#/c/263874/3 or https://review.openstack.org/#/c/262809/2 is enough?
15:50:26 <EmilienM> I guess https://review.openstack.org/#/c/263874/3 is the one to test
15:50:29 <mwhahaha> sure as long as we make sure not to reintroduce that bug, I think it would be OK to switch back to csv
15:50:32 <chem> EmilienM: you should take both to not be bitten by warning output
15:50:49 <EmilienM> chem: ok
15:50:52 <chem> mwhahaha: I added the necessary tests, so that should be robust
15:50:59 <EmilienM> chem: they are dependents, right?
15:51:05 <mwhahaha> which since json wasn't added until liberty (i think) would make it more compatible with previous versions anyway
15:51:08 <chem> EmilienM: yep
15:51:13 <EmilienM> cool
15:51:18 <EmilienM> I'll test it today
15:51:33 <EmilienM> in the meantime, I would like to see some discussions between folks involved in that, and maybe a concensus
15:51:56 <chem> EmilienM: yeap, that why I send a lot of noise on the ML :)
15:52:13 <chem> EmilienM: will get in touch with Sergey
15:52:19 <EmilienM> cool
15:52:22 <EmilienM> thanks folks
15:52:38 <EmilienM> I updated https://etherpad.openstack.org/p/puppet-openstack-ci-mitaka
15:52:49 <EmilienM> good progress on getting mitaka!
15:53:08 <EmilienM> do we have anything else for today? I'll close the meeting in 30s otherwise
15:53:42 <EmilienM> gchamoul: did you resolve your issues with rspec on puppet-swift ?
15:53:55 <mwhahaha> just to announce, i'm working on rspec-puppet-facts
15:54:09 <EmilienM> mwhahaha: thanks a lot for this work, I did some reviews already
15:54:09 <mwhahaha> https://review.openstack.org/#/c/264830/ and https://review.openstack.org/#/q/topic:rspec-puppet-facts
15:54:35 <mwhahaha> comments welcome, i've got patches for puppet-cinder and puppet-aodh
15:54:40 <gchamoul> EmilienM: nope, there are a lot of issue due to some changes in puppet 4.3.0
15:54:43 <mwhahaha> so everyone can see what it looks like
15:55:03 <EmilienM> gchamoul: please let the group know if you need some help
15:55:34 <gchamoul> EmilienM: for sure, I need help ... others couple of eyes ...
15:55:35 <EmilienM> specially Hunner or _ody if that's puppet core things
15:56:11 <EmilienM> #action chem to discuss about https://review.openstack.org/#/c/263874/ and find concensus with the group
15:56:53 <EmilienM> #action emilien to test https://review.openstack.org/#/c/263874/ with mitaka repos
15:57:25 <EmilienM> Hunner, _ody: can you give a hand on gchamoul about his work on getting rspec jobs working on latest versions?
15:57:33 <EmilienM> AFIK only puppet-swift is blocking now
15:57:55 * _ody doesn't even know where Hunner is
15:57:59 <_ody> Like a ghost lately.
15:58:04 <_ody> gchamoul: Yeah I can help.
15:58:20 <EmilienM> thanks guys
15:58:23 <EmilienM> I'm closing the meeting
15:58:24 <gchamoul> _ody: thanks!
15:58:31 <EmilienM> #endmeeting