Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-06-02

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:36 kitp joined #puppet-openstack
02:23 openstackgerrit David Moreau Simard proposed a change to stackforge/puppet-ceph: [WORKINPROGRESS] Add ceph::calamari  https://review.openstack.org/97128
02:35 kitp joined #puppet-openstack
03:25 kitp joined #puppet-openstack
03:39 mmaglana joined #puppet-openstack
04:18 tfz joined #puppet-openstack
04:32 starmer joined #puppet-openstack
05:34 kitp joined #puppet-openstack
05:36 kitp joined #puppet-openstack
05:37 ric` joined #puppet-openstack
05:53 sanjayu joined #puppet-openstack
06:09 mkoderer joined #puppet-openstack
06:21 thumpba_ joined #puppet-openstack
06:37 kitp joined #puppet-openstack
06:52 marun joined #puppet-openstack
07:03 EmilienM good morning
07:08 artem_panchenko joined #puppet-openstack
07:16 michchap EmilienM: morning
07:21 arnaud_orange joined #puppet-openstack
07:37 kitp joined #puppet-openstack
07:55 mmagr joined #puppet-openstack
08:03 admin0 joined #puppet-openstack
08:38 kitp joined #puppet-openstack
08:59 openstackgerrit A change was merged to stackforge/puppet-swift: Avoid duplication error when ringserver and storage are on the same node  https://review.openstack.org/96505
09:26 mattymo hi EmilienM michchap
09:27 mattymo we'll have our first meeting tonight? (morning for US)
09:27 michchap mattymo: Morning :) I don't know if I'll be there since it's at about 1am.
09:27 aldavud joined #puppet-openstack
09:28 EmilienM michchap: damn
09:29 mattymo michchap, I know you need your beauty rest
09:29 EmilienM michchap: next time we could do it late for me, I don't mind
09:29 EmilienM michchap: but you will have to wake up at 8am :P
09:29 mattymo late is fine for me too. most of our meetings are at night already
09:30 michchap mattymo: lol yes. It actually has taken me nearly 2 weeks to get my sleeping cycle roughyl back to normal and I am still falling asleep earlier than I usually would.
09:30 mattymo michchap, I travel often enough so it takes usually 3 days
09:31 michchap mattymo: Ah...I only really do the summits and the odd holiday. This time has been worse than usual.
09:32 michchap EmilienM: we'll see - I'm sure I'll be fine for late meetings soon.
09:32 EmilienM michchap: your presence is more than required to be honest
09:33 EmilienM s/requied/really wanted/
09:33 EmilienM :)
09:33 mattymo EmilienM, michchap in case you guys are interested, we've got a list of proposed blueprints for Fuel 5.1 https://etherpad.openstack.org/p​/fuel-library-5.1-design-session
09:34 EmilienM mattymo: sorry, but as usual I would suggest pushing efforts in our upstream puppet modules
09:34 mattymo I understand. it's one of the blueprints
09:34 mattymo :)
09:34 EmilienM so you etherpad is cool, but I don't care, since I won't use this code
09:34 mattymo but we won't stop developing new features. what I'm looking for is what feature would you like to see in upstream puppet openstack
09:35 EmilienM what I would like to see, is patches in our upstream modules
09:35 michchap EmilienM: will be fine. You know what I am working on with the lib. I am writing a blog post about how your config-tools system works.
09:35 michchap EmilienM: I am trying to replicate what you have as much as I can, though without Jenkins. BTW do you have the Jenkins jobs themselves stored publically?
09:35 EmilienM michchap: awesome. I did not have time to test the builder yet
09:35 michchap EmilienM: no rush.
09:36 EmilienM michchap: our Jenkins is and will stay internal
09:36 michchap EmilienM: yeah...but when you are running a job, what is the script for the job?
09:36 michchap EmilienM: the jjb file
09:36 EmilienM michchap: but I can share all the jobs YAML files (we use JJB)
09:37 michchap EmilienM: yeah that's what I was after - so I can see what happens at the very start of a job.
09:39 kitp joined #puppet-openstack
09:40 michchap mattymo: Some work on the ML2 stuff in puppet-neutron would be useful. EmilienM has been mostly on his own on that - we could use more testing.
09:41 EmilienM i would love reviews by the way
09:41 sanjayu joined #puppet-openstack
09:42 mattymo yeah ml2 is really important for juno. trying to get our neutron guy to look at it
10:40 kitp joined #puppet-openstack
10:49 openstackgerrit Lukas Bezdicka proposed a change to stackforge/puppet-swift: Restore selinux context on swift mount  https://review.openstack.org/97195
11:41 kitp joined #puppet-openstack
11:54 morazi joined #puppet-openstack
12:41 kitp joined #puppet-openstack
12:49 dprince joined #puppet-openstack
13:10 openstackgerrit Yanis Guenane proposed a change to stackforge/puppet-swift: Improve readability of swift::xfs  https://review.openstack.org/94867
13:12 spredzy_ Guys could I have please some reviews on https://review.openstack.org/#/c/94395/, https://review.openstack.org/#/c/94399/, https://review.openstack.org/#/c/94412/
13:12 mjblack joined #puppet-openstack
13:12 spredzy_ All related to enable SSL on endpoints
13:42 kitp joined #puppet-openstack
13:44 daneyon joined #puppet-openstack
13:52 toonces joined #puppet-openstack
13:54 xingchao joined #puppet-openstack
14:00 mjblack joined #puppet-openstack
14:05 kitp joined #puppet-openstack
14:09 openstackgerrit Yanis Guenane proposed a change to stackforge/puppet-swift: Bump stdlib dependency from >=2.3.0 to >=3.2.0  https://review.openstack.org/97243
14:13 openstackgerrit Arnaud proposed a change to stackforge/puppet-nova: Adds vnc_keymap support  https://review.openstack.org/97244
14:13 EmilienM arnaud_orange: tests? ^
14:14 arnaud_orange EmilienM: what do you mean?
14:14 EmilienM arnaud_orange: https://wiki.openstack.org/wiki/Pu​ppet-openstack#Rspec_puppet_tests
14:16 arnaud_orange EmilienM: ok, sorry, didn't do that yet, should it be done before commiting?
14:16 EmilienM arnaud_orange: every patch has to include the tests.
14:17 EmilienM so you have to patch both manifests/* and spec/classes/*
14:17 EmilienM and also please add a commit message
14:17 mgagne joined #puppet-openstack
14:17 arnaud_orange the short commit message isn't enough?
14:19 EmilienM arnaud_orange: usually, developers write commit message with the goal to understand what the patch does without reading the coe
14:19 EmilienM code*
14:19 arnaud_orange Ok, will do that.
14:19 aldavud https://wiki.openstack.org/wiki/GitCommitMessages
14:20 aldavud best practice for commit messages ^
14:20 arnaud_orange thx
14:23 xingchao For a standard commit msg: a short summary, an detailed description, and an related bug/bp number is needed
14:25 openstackgerrit Yanis Guenane proposed a change to stackforge/puppet-swift: Improve readability of swift::xfs  https://review.openstack.org/94867
14:25 thumpba_ joined #puppet-openstack
14:25 mdorman joined #puppet-openstack
14:27 mjblack joined #puppet-openstack
14:27 openstackgerrit Yanis Guenane proposed a change to stackforge/puppet-swift: Improve readability of swift::xfs  https://review.openstack.org/94867
14:41 arnaud_orange EmilienM: is there any way to improve the review I did with tests only or do I need to commit again the whole thing?
14:42 EmilienM arnaud_orange: sorry I'm not sure to get it
14:42 EmilienM arnaud_orange: if you changed something, git review again
14:42 arnaud_orange I am currently changing the spec/classses/* files
14:43 arnaud_orange and will commit + review
14:43 arnaud_orange but the previous commit is already done
14:44 EmilienM arnaud_orange: amend
14:45 EmilienM git commit --amend
14:45 arnaud_orange hum, ok, thanks
14:46 cbaesema joined #puppet-openstack
14:48 NuclearW0rm joined #puppet-openstack
14:51 openstackgerrit Lukas Bezdicka proposed a change to stackforge/puppet-swift: Add swift::proxy::crossdomain class  https://review.openstack.org/97259
14:54 Hunner Is the meeting today in #openstack-meeting-3? Sstackalytics has #openstack-meeting and satori #openstack-meeting-alt
14:54 mgagne almost forgot about it
14:57 kkirkpatrick joined #puppet-openstack
14:58 dachary will it be #openstack-meeting-3 ?
14:58 mmagr joined #puppet-openstack
14:58 hogepodge joined #puppet-openstack
14:58 mgagne afaik, no room has been reserved
14:58 dtalton joined #puppet-openstack
14:58 EmilienM oops
14:59 hogepodge This is the room.
14:59 EmilienM "For the first puppet-openstack IRC developer meeting, I'm proposing Monday June 2 at 15:00 UTC in the #puppet-openstack channel on freenode."
14:59 hogepodge Here is the etherpad. https://etherpad.openstack.org​/p/puppet-openstack-2014-06-02
14:59 EmilienM hogepodge: I would suggest using a OpenStack meeting room next time, because the have the bot
14:59 EmilienM #startmeeting stuff
15:00 dachary mouhahaha
15:00 thumpba_ joined #puppet-openstack
15:00 mgagne lets start, we will figure that out after
15:00 xingchao this room also has bot :)
15:00 EmilienM xingchao: really?
15:00 EmilienM Hunner: meeting-3 is busy by storyboard
15:01 hogepodge I was looking at the meeting rooms. It looked like main or 2 might be free, but theis seemed as good as place as any to start out at.
15:01 hogepodge log: http://irclog.perlgeek.de/puppet-openstack/today
15:01 xingchao EmilienM: It seems just has log function
15:01 EmilienM ah yeah
15:02 thumpba_ joined #puppet-openstack
15:02 EmilienM hogepodge: I guess you lead the meeting?
15:02 hogepodge I guess so EmilienM. I haven’t done an IRC meeting before, so tips and feedback encouraged.
15:03 hogepodge If you’re here for the meeting, can you drop your name into the attendance section in the Etherpad?
15:03 EmilienM let's start by #startmeeting
15:03 EmilienM then announce the topic with #Topic blabla
15:04 EmilienM I think it would be enough for today
15:04 hogepodge #Topic Libraries (puppet-openstacklib and puppet-openstack_common)
15:04 michchap I'm just typing out summaries for those.
15:05 kkirkpatrick joined #puppet-openstack
15:05 hogepodge Kind of a silly point to start with, but there’s been confusion about how to name. I like what Michael is writing, though. openstacklib is the low level library that the other modules will depend upon.
15:06 hogepodge Common is the library that builds out common functionality like HA (higher firewall).
15:06 hogepodge The summaries are a good place to start. The next steps should be to write up basic specs and create the projects. Thoughts on any of that?
15:07 EmilienM openstacklib will be mandatory for all openstack modules
15:07 EmilienM hogepodge: by using the new repo, right?
15:07 mgagne I didn't have much time to test or prototype something with aviator in openstacklib.
15:07 dvorak are those names set in stone?  I think part of the confusion is that those names sound like nearly the same thing.  using openstack_extras or openstack_supp might be less confusing than openstack_common
15:08 dmsimard I think real integration testing will become a necessity since many modules will be using the same core libraries and even the slightest modification could have repercussions
15:08 mgagne However we should come up with a solution on how to distribute aviator in our module: gem, bundled in openstacklib, own module, etc.
15:08 spredzy_ could we also gather the puppet common dependencies to the openstack module into openstacklib ? Since every module will inherit it, module dependency would take place in a single spot
15:08 dvorak spredzy_: +1
15:08 spredzy_ (ie. stdlib, concat, inifile, etc...)
15:08 michchap dvorak: we haven't even made the repo yet. It will be a merge of puppet-openstacklib (my module) and puppet-openstack-cloud (Emilien's module) and that merging is going to take at least a month to get right.
15:09 mgagne we shouldn't mix subjects if we want an efficient meeting...
15:09 michchap OK, so the topic right now should be just the first library.
15:10 hogepodge #Topic puppet-openstacklib
15:11 michchap dmsimard: we found that out when the mysql module changed drastically on master branch :D
15:11 hogepodge Aviator in openstacklib, my preference is to either include it, or build a new module that includes it. We can’t always rely on operators being able to install gems.
15:11 nfedotov joined #puppet-openstack
15:11 rickerc agreed - no gems requirement
15:12 michchap is there a way we could do a git submodule? how does that interact with the forge?
15:13 michchap I also like the idea of not having gems, but we shouldn't be reinventing libraries.
15:13 hogepodge As long as we don’t use symlinks it should build out fine. A submodule would be a good way to track changes, but I would want us to pin to a fixed point, and deliberately move it forward.
15:14 mgagne michchap: I had the idea to create a puppet module bundling aviator. This module could be surprisingly named puppet-aviator
15:14 michchap mgagne: I am overwhelmed with surprise.
15:14 xarses what pulls in aviator?
15:14 xingchao git submodule is a good way
15:14 michchap mgagne: so do you build that with CI somewhere?
15:15 dachary hogepodge: git submodule reference a commit hash and won't move forward
15:15 michchap does this break r10k/librarian?
15:16 dvorak aviator also pulls in a few deps.  a Gemfile with just aviator in it requires 23 gems
15:16 hogepodge Oh, right. Yuck.
15:16 michchap that's pretty bad
15:16 dvorak some of those are redundant, like activesupport ends up in there twice.
15:17 hogepodge dvorak That’s a problem. Do you have a list of them?
15:17 hogepodge maybe paste into the etherpad?
15:18 dvorak here is a gist - https://gist.github.com/dv​orak/cbec3139a48f23b8d0d4
15:18 xarses mgagne: dvorak what requires aviator ? mysql?
15:19 mgagne Have I been disconnected? not much activity in this meeting :-/
15:19 EmilienM mgagne: no
15:19 hogepodge xarses: we want to stop using the command line interfaces and start using the REST endpoints.
15:20 hogepodge So we need to pull in json, http libraries, and so on.
15:20 rickerc some of the aviator dependencies are only for when using it as a cli, right? and aren’t needed if all we care about is a library?
15:20 * mgagne got bombed by backlog
15:20 xingchao mgagne: puppet-aviator modules is aim to install aviator ?
15:20 xarses ah, so once fuel ends up consuming it we will start building packages for it, since we don't like gems ( for whats its worth )
15:20 mgagne xingchao: yes so other people can use it outside puppet-openstack
15:20 dvorak hogepodge: we'll have to dig in and find out how many of those can be pulled out, for example, things like minitest don't need to be in there
15:20 michchap xarses: yeah I'll probably end up throwing it in an omnibus
15:21 hogepodge michchap Lots of git submodules.
15:21 Pax joined #puppet-openstack
15:21 openstackgerrit Arnaud proposed a change to stackforge/puppet-nova: Adds vnc_keymap support  https://review.openstack.org/97244
15:21 mgagne git submodules won't be compatible with forge
15:22 tfhartmann joined #puppet-openstack
15:22 mgagne would they?
15:22 badiane_ka joined #puppet-openstack
15:22 EmilienM xarses: I don't get the fuel stuff here
15:22 xingchao mgagne: I think it could be in openstacklib::deps::aviator
15:23 hogepodge mgagne I don’t see why whey wouldn’t be. It’s worth checking out. Symlinks are bad, everything else should just like like files to PMT.
15:23 hogepodge s/like like/look like/
15:23 mgagne hogepodge: I don't suggest using symlink, I suggest creating a module with aviator bundled in it so people don't have to gem install it
15:24 mmaglana joined #puppet-openstack
15:24 hogepodge mgagne No symlinking. Sub modules (confusing teminolog) just build out.
15:24 admin0 joined #puppet-openstack
15:24 hogepodge Ok, so actions:
15:25 hogepodge * puppet-aviator project to capture library code and remove gem dependency.
15:25 hogepodge * puppet-openstacklib project added to stackforge
15:26 hogepodge (project proliferation here, let’s be mindful about not throwing out too many more)
15:26 EmilienM it makes sense
15:26 hogepodge * initial specs for puppet-openstacklib design, including testing plan.
15:27 hogepodge Does anyone have initial code that they want to use for the seeds? Otherwise I can create skeleton projects and get them both added to Stackforge this week.
15:27 michchap hogepodge: do I have it right on the pad that Aimon was keen for this?
15:27 mgagne hogepodge: I have nothing yet
15:28 hogepodge ping aimon ?
15:28 hogepodge I’ll check in with him.
15:28 hogepodge Any further comments on this?
15:29 michchap so the conclusion is that we're not using git submodules, or we are?
15:29 hogepodge We are using git submodules.
15:29 mgagne hogepodge: I don't think git submodules are a good idea
15:30 hogepodge What would be the other solution?
15:30 hogepodge Maintain forks?
15:30 mgagne hogepodge: what I suggested above
15:30 mgagne hogepodge: we don't need a fork
15:30 hogepodge Ok, I see above. Would aviator bundled in it be a fork of the original code, though?
15:31 mgagne hogepodge: what's the definition of fork for you?
15:31 xingchao mgagne suggest using puppet-aviator to manage it
15:31 dvorak hogepodge: depends on how successful we are in getting our changes into aviator I imagine
15:31 xingchao and we should package aviator
15:31 hogepodge Code copied out of the project into the new project.
15:31 mgagne hogepodge: devs want the ease of use of git and submodules. Operators can't afford that in production. modules have to be self contained and autonomous
15:32 social__ Hi I have question about commit tag Closes-Bug:. Could it be possible to extend it so it adds external gerrit review link also to redhat bugzilla?
15:32 mgagne hogepodge: yes, that ia what I suggested by "bundling" it
15:32 mgagne hogepodge: and we use a script to update it whenever we need it
15:32 michchap mgagne: I don't want git modules unless  they work with librarian - I'm not sure they do.
15:32 EmilienM social__: we are in meeting now, stay online, we will get you support after the meeting.
15:32 dmsimard social__: You might have a better chance in #openstack-infra
15:33 michchap mgagne: s/modules/submodules
15:33 Hunner librarian could always be updated; r10k supports submodules
15:34 hogepodge I see them as similar means to achieve the same end, but don’t have my heart set on submodules.
15:34 Hunner Actually, I should probably double-check that before claming so :/
15:34 xingchao mgagne: we need packaging aviator into rpm and deb for production use, and puppet-aviator support multi OS,  is that right ?
15:34 michchap Hunner: I was frantically googling for r10k submodules and couldn't find anything.
15:34 hogepodge I do worry about code drift, though.
15:34 mgagne xingchao: I don't know
15:34 michchap xingchao: would you be able to host a yum/apt repo?
15:35 abramley left #puppet-openstack
15:35 hogepodge xingchao one of the benefits to just sticking library code into the module is that pluginsync will make everything available to all of the agents without having to worry about dependencies or collisions with other installed libraries.
15:36 xingchao michchap: no, but I could afford a spec git repo. people can packaging  by using it.
15:36 Hunner michchap: https://github.com/adrienthebo/r10k/blob/ma​ster/lib/r10k/cli/environment/deploy.rb#L13
15:36 michchap hogepodge: that pretty much rules out everything other than a puppet module, doesn't it.
15:36 mgagne hogepodge: using git submodule is the same, you have to update the commit to which your git submodule points.
15:36 dvorak Hunner: that's not in a relased version yet
15:36 michchap xingchao: sure. I think fpm should be able to package it pretty easily.
15:36 aimon hogepodge yo!
15:37 michchap Hunner: awesome thanks.
15:37 aimon you rang?
15:37 Hunner But submodules are more annoying than in 90% of the times people use them
15:37 Hunner than useful*
15:37 hogepodge aimon we’re having the weekly meeting. Can you take a look a the log? We’re talking about aviator right now. http://irclog.perlgeek.de/puppet-openstack/today
15:37 xarses xingchao: michchap hogepodge:  Like I said, fuel team will have to start making Ubuntu and RHEL packages for aviator, so those should be available this release cycle as we pull up on the main line puppet-openstack modules
15:38 EmilienM xarses: upstream in Ubuntu & RHEL?
15:38 michchap EmilienM: I'd imagine on the Fuel repo.
15:38 hogepodge xarses which repos would that live in? Would they conflict with RDO and UCA?
15:39 EmilienM michchap: which is not the best way IMO for our modules.
15:39 xingchao hogepodge: I understand the benefits, but should we pin a gem package by using pluginsync?
15:39 xarses EmilienM: We might be able to get them upstream with Ubuntu
15:39 xingchao xarses: good news
15:39 aimon michchap what item did you mean?
15:39 EmilienM michchap: hogepodge: #openstack-infra uses only cloud archive afik
15:39 xarses hogepodge: I'm not sure about RDO/UCA it would depend on the deps chain
15:39 rickerc xarses: if you got into Ubuntu and EPEL that’d probably cover
15:39 hogepodge xingchao pluginsync is guaranteed to work regardless of platform.
15:39 michchap aimon: I recall you were looking into aviator for native types.
15:40 dtalton joined #puppet-openstack
15:40 aimon I asked if y’all wanted me to give it a shot.. i got no reply sos I did not proceed
15:40 aimon but I certainly can
15:41 michchap so it seems there's going to be two camps: people who push out a deb/rpm package to every node for aviator, and people who are installing everything using r10k/librarian from git sources and thus would like pluginsync. Is that accurate? Is there another group?
15:42 hogepodge I’m firmly in the pluginsync camp unless aviator goes mainline in Ubuntu/Debian and RH.
15:43 Hunner +1 pluginsync
15:43 aimon yes +1 to that
15:43 aimon maintaining debs/rpms is a no win for us
15:43 dachary +1
15:43 mgagne we can't package it, create a repo and expect the end user to use it. people might not have that liberty.
15:43 aimon install/lib locations vary way too much to make it sane
15:43 openstackgerrit Francois Deppierraz proposed a change to stackforge/puppet-nova: Introduce nova::config + Update nova::config example description.  https://review.openstack.org/97272
15:44 EmilienM +1 pluginsync also
15:44 michchap so lets assume we make puppet-aviator, with the aviator code under lib either via submodule or by copy, how are we dealing with Aviator's dependencies in this scenario?
15:45 hogepodge michchap It sounded like we were going to enumerate them as needed. notfun.
15:46 michchap aimon: do you know how many of the aviator deps are actually needed?
15:47 michchap https://github.com/aviator/avia​tor/blob/master/aviator.gemspec
15:48 michchap I wonder if we can get away without thor and terminal table since we dont' need output
15:48 hogepodge +1 to that
15:48 aimon michchap I am on call with mmaglana now.. I’ll slip in the question
15:49 michchap aimon: we'll definitely need at least faraday and probably activesupport.
15:49 hogepodge We’re coming up on an hour, I can stay until 16:45 UTC, but want to make sure we don’t go too long for others who have obligations.
15:50 dachary +1 on timeboxing ;-)
15:50 hogepodge Ok, so let’s revisit this next week. I can investigate how to best package. Anyone else want to look also?
15:51 mmaglana hello! thor and terminal are only needed for the CLI. :-)
15:51 michchap hogepodge: just use fpm. tis easy
15:51 EmilienM michchap has to sleep, we have a lot of work coming with the lib :P
15:51 michchap haha
15:51 nfedotov left #puppet-openstack
15:51 michchap mmaglana: thanks! That's good news
15:52 mmaglana sidenote: will be deleting that aviator repo linked above. making https://github.com/stackforge/aviator the official one sometime this week.
15:52 hogepodge mmaglana please, feel free to drop in feedback about how we’re looking at this into the etherpad.
15:53 hogepodge new topic? I want to quickly touch on the others in the list.
15:53 michchap sure
15:53 hogepodge #Topic blueprints/specs
15:54 ryanycoleman joined #puppet-openstack
15:54 hogepodge bodepd suggested we do a review on what’s actually necessary. https://review.openstack.org/#/c/96876/ It looks like there aren’t any changes, but you can expand the view and leave comments.
15:54 hogepodge An alternative would be to just do editing in an Etherpad. I just copied it right out of Nova with some changes, which is why it’s so heavy.
15:55 hogepodge Is this still something that we want to do? There hasn’t been much response to it.
15:55 social__ dmsimard: EmilienM: thanks
15:55 michchap There are 3 blueprints on launchpad right now
15:55 michchap I'm going to admit I've never created one. I just add things that I need.
15:56 hogepodge I’m not much of a user either, admittedly. It’s come up at both summits, though.
15:56 michchap I think there's only a small number of situations where we actually get value out of blueprints. How do others feel?
15:56 spredzy_ IMO we need blue prints for changes that applies to all module (a given change/pattern that we need to apply)
15:57 michchap spredzy_: right. To enforce consistency so we don't end up with different interfaces.
15:57 aimon joined #puppet-openstack
15:57 hogepodge can attendees drop their opinion into the etherpad as a poll? +-1/0?
15:57 michchap spredzy_: like the SSL stuff that just went up?
15:58 xingchao michchap: +1
15:58 spredzy_ michchap, for example ;)
15:58 sputnik13 joined #puppet-openstack
15:58 spredzy_ we could have all agreed on how to apply SSL, and then implementing it is easy
15:58 spredzy_ it would avoid the back and forth reviews
15:58 spredzy_ if everyone know which direction we are taking and how it will be accomplished
15:58 hogepodge I think it would be valueable, especially with the coordination that’s going to be needed for a new lib.
15:59 openstackgerrit Francois Deppierraz proposed a change to stackforge/puppet-glance: Introduce glance:config to manage custom options  https://review.openstack.org/97276
15:59 michchap ok, that sounds good to me. So if you're adding a parameter/feature to a module that's going to be repeated (ie. anything exposed in Oslo) then it may need a BP
15:59 EmilienM yep
15:59 hogepodge Plus there was a lot of interest from the operators group, and they’re the consumers of our work.
15:59 EmilienM like the way to configure MySQL or RabbitMQ
15:59 hogepodge Getting started is the hardest part.
16:00 michchap EmilienM: right. That's one we've already messed up in a few ways.
16:00 hogepodge I don’t like to spend all my time designing, I like to code. :-) I also don’t like fixing my bad design decisions. :-/
16:00 spredzy_ won't be a one man's fault ever again ;)
16:01 michchap so if we see reviews that are on something to be added to many modules, we should add a -1 if we feel there should be a review, to prevent back and forth?
16:01 EmilienM we should not spend too much time on spec, just enough to have something good
16:01 hogepodge It also gives a point of reference for code review.
16:02 michchap so that we stop discussion on a patchset and move to a spec?
16:02 hogepodge +1 michchap
16:02 EmilienM if the discussion land in the patchset it's too late imo
16:02 rickerc michchap: yes
16:02 michchap well that seems actionable at least.
16:03 openstackgerrit Francois Deppierraz proposed a change to stackforge/puppet-keystone: Introduce keystone:config  https://review.openstack.org/97278
16:03 EmilienM how do we handle people sending patches without the blueprint?
16:03 EmilienM -1 + a comment to say "write spec" ?
16:03 rickerc yes
16:03 * dachary leaves the room trying to make as little noise as possible
16:03 EmilienM or -2 :P
16:03 michchap if it only affects one module then I would review it as normal
16:03 hogepodge +1 EmilienM
16:03 spredzy_ well first you need to see if it impacts more than one module
16:04 spredzy_ if it does, kindly send him back to 'write spec before please'
16:04 michchap if it affects multiple modules then -1 and say we should write a spec
16:04 EmilienM yep ok make sense
16:04 hogepodge I’m all for just starting to submit specs to the project right now, regardless of the template.
16:04 michchap We should be wary of scaring off new contributors with this requirement.
16:05 EmilienM i fill out the etherpad with this
16:05 openstackgerrit Francois Deppierraz proposed a change to stackforge/puppet-cinder: Introduce cinder:config to manage custom options  https://review.openstack.org/97279
16:05 EmilienM need a blueprint ^^^^ \o/
16:05 aimon I’v got yo get to the office.. michchap I will give creating an aviator module a shot and report back how it went
16:05 michchap aimon: cool! thanks!
16:05 hogepodge aimon thanks!
16:05 aimon np at all
16:05 aimon ttyl
16:06 michchap EmilienM: lol...I think we already implemented the config class to every other module though.
16:06 EmilienM hogepodge: are we good with spec?
16:06 hogepodge yes.
16:06 hogepodge #Topic Meeting Schedule.
16:06 EmilienM I would suggest alternate with michap and me
16:06 EmilienM I don't mind staying late
16:07 michchap seriously dw about me. Sometimes I will come and sometimes I won't.
16:07 hogepodge This meeting time is terrible for michchap.
16:07 EmilienM hogepodge: we can switch with my timezone, but I care about xingchao
16:07 hogepodge How many other people is 15:00 UTC terrible for?
16:07 openstackgerrit Francois Deppierraz proposed a change to stackforge/puppet-neutron: Introduce neutron:config to manage neutron configuration  https://review.openstack.org/97281
16:07 michchap yeah but even if you made it the perfect time for me, I still might have to be somewhere else.
16:07 hogepodge xingchao ?
16:07 EmilienM hogepodge: i care about core developpers
16:07 xingchao It's 0:07 here. I'm ok
16:08 EmilienM and major contributors
16:08 hogepodge EmilienM me too.
16:08 EmilienM hogepodge: i would suggest switching the next time to have a good time for xingchao and michchap
16:08 EmilienM and late for me
16:08 michchap I think we just need to accept that not everyone is going to make every meeting.
16:08 EmilienM michchap: OpenStack project use to switch the time
16:08 mgagne sorry guys, I'm in a meeting at the office
16:09 francois joined #puppet-openstack
16:09 EmilienM mgagne: about hours
16:09 hogepodge So if 15:00 UTC this week, what time next week?
16:10 EmilienM afik, major contributors are in europe, US/CA and AUS/China, so we need to switch between AUS/China & Europe
16:10 michchap xingchao: how early is OK for you?
16:11 EmilienM 10pm UTC ?
16:11 xingchao michchap: how about 13:00  UTC?
16:12 EmilienM I vote for 10pm UTC
16:12 EmilienM http://www.worldtimebuddy.com/?qm=1&amp​;lid=2988507,100,6077243,2147714&h=​2988507&date=2014-6-2&sln=0-1
16:13 michchap xingchao: ah you went in the other direction. I think that's early for US West? I was thinking something like 00:00 UTC so it's about 8am for you
16:13 EmilienM if michchap can wake up at 8am, I know he's a late guy :P
16:13 xingchao EmilienM:  actually I sleep late, so what's your best hours ?
16:13 michchap EmilienM: I think that wuld be 6am for xingchao
16:13 EmilienM michchap: 00 UTC is 2am to 3am \o/
16:13 michchap EmilienM: ouch
16:13 EmilienM i may be drunk at this time
16:14 xingchao oh, I can't get up so early :)
16:14 michchap I think we should stick with the current time.
16:14 EmilienM so don't consider my words
16:14 michchap xingchao and myself are both late waking up...
16:14 mgagne I don't think we will be able to accommodate everyone
16:14 mgagne we have people working in Singapore and scheduling a meeting is near impossible
16:14 EmilienM mgagne: no but at least switching for next time would be fair for michchap
16:14 mgagne EmilienM: ^
16:14 EmilienM and xingchao also :-)
16:14 EmilienM ok
16:14 EmilienM we should just vote, no?
16:15 michchap let's stick with this time for now and if it becomes a problem then xingchao can start -2 all our patches until we change it
16:15 EmilienM time is running out, we shoud go ahead
16:15 EmilienM and discuss on it later
16:15 michchap EmilienM: yep.
16:15 hogepodge xingchao ?
16:15 xingchao 15:00 UTC is ok, no need for adjusting time for me :)
16:15 EmilienM you rock guys
16:16 xingchao next topics :)
16:16 hogepodge I’ll have some coffee or tea shipped to you.
16:16 hogepodge #Topic reviews
16:17 hogepodge This can be quick, if you’re waiting on a review that’s important to you, please drop it in the etherpad.
16:17 EmilienM the one with ML2 is important to me
16:17 EmilienM https://review.openstack.org/#/c/82353/
16:17 hogepodge Keeping in mind that all reviews are equal, just some more equal than others.
16:17 EmilienM thx francois for the comment btw
16:18 EmilienM hogepodge: lol
16:18 hogepodge next topic?
16:18 michchap hogepodge: yep
16:18 hogepodge #Topic Integration tests
16:18 francois EmilienM: no pb, thanks for the patch
16:19 hogepodge Who’s running integration tests against gerrit? Does anyone have a replacement for smokestack?
16:19 michchap I don't want to put words in his mouth, but bodepd is working on building public CI.
16:19 dmsimard hogepodge: We are, with ceph, with rspec-system
16:20 dmsimard hogepodge: rspec-system was axed by puppetlabs and it looks like the way forward would be beaker but we haven't made the jump yet
16:20 hogepodge dmsimard let me know if we can help out with beaker.
16:21 hogepodge CI is on my schedule to build out in June/July.
16:21 michchap I think one of the challenges I've struggled with is building sane rpms and debs from openstack.
16:21 dmsimard hogepodge: The main pain point is this one: https://github.com/puppetlabs/beaker/issues/155
16:22 dmsimard but it looks like it's something that can be done with vagrant without extending beaker
16:22 michchap It really requires consumption of the three major downstreams (debian/ubuntu/redhat) in an abstracted way, and that doesn't exist atm afaik.
16:23 hogepodge It’s about to get harder too, with the 12/14 and 6/7 changeover happening.
16:24 hogepodge Maybe RH and Canonical will make it easier for us by just targeting one system for their repos.
16:24 michchap I don't think that's such a big deal. The problem is you have to replicate the package building infra of 3 very different orgs
16:25 hogepodge Why can’t we just consume their packages?
16:25 michchap hogepodge: I don't think that's going to happen :(
16:25 michchap hogepodge: because we want to run master against master.
16:25 mattymo I can comment that RH gave up on nightly master builds some time ago
16:25 mattymo and we're finding it incredibly difficult to build off master consistently
16:26 mattymo stable master is a lot easier
16:26 hogepodge I’d just make a new installer that pulls from source and leverages virtenv
16:26 michchap hogepodge: you won't have any init scripts.
16:26 mattymo better with venvs, but it's hard for puppet
16:26 michchap the init scripts are written and maintained by the downstreams
16:27 michchap if you don't need init scripts, then you can use anvil or just wrap fpm.
16:27 rickerc that’s maybe a plus as they’re somewhat broken :-)
16:27 michchap but without init scripts puppet is going to have issues managing services
16:28 hogepodge How many init scripts are just cut and paste jobs? Nearly all of them from what I could tell. I dropped my own init scripts with templates.
16:28 rickerc things like neutron with *aaS stuff and w/ or w/o ml2 transition is where the inits get fun
16:28 hogepodge Setting that aside, though, are the packaging systems themselves open source? Could be just appropriate those?
16:28 michchap the swift one also broke recently on ubuntu.
16:29 michchap hogepodge: consider the Ubuntu one is building things from Bazaar.
16:29 hogepodge (just he says)
16:29 EmilienM mattymo: could say to Sergey Vasilenko to not push in my patchsets ???
16:29 EmilienM https://review.openstack.org/#/c/82353/
16:30 mattymo EmilienM, he said it looks good
16:30 rmoe joined #puppet-openstack
16:30 mattymo you mean "to not push my patchset"? He didn't add a new patchset (push in)
16:30 EmilienM mattymo: gerrit says yes
16:30 michchap EmilienM: I don't see a push from anyone other than you?
16:30 hogepodge Can we start a mailing list thread on the CI issue?
16:30 EmilienM jenkins is unhappy
16:30 hogepodge I think I need to call the meeting so I can prep for my next.
16:31 michchap hogepodge: talk to Dan.
16:31 aimon joined #puppet-openstack
16:31 mattymo hogepodge, go ahead. I'm not our CI expert. We're a while away before we can use our existing system to test CI for puppet-openstack directly
16:31 mattymo but I hope we can get to that in 1-2 months where we can cherry pick and do a full deployment and validate
16:31 michchap hogepodge: he's researching it now, and everything he builds he is going to open up so others can replicate it.
16:32 hogepodge I’ll ping him. Thanks.
16:32 hogepodge EmilienM is there anything special I do to end a meeting?
16:32 EmilienM #endmeeting
16:32 mattymo EmilienM, I just see  "Sergey Vasilenko2:03 PM Patch Set 6: Code-Review+1"
16:32 EmilienM mattymo: ok I misread, new gerrit excuse works ? :-)
16:32 hogepodge Thanks everybody!
16:33 EmilienM thanks! see you guys
16:33 michchap Later folks.
16:33 EmilienM michchap: night :P
16:33 mmaglana joined #puppet-openstack
16:33 michchap EmilienM: it sure is!
16:33 EmilienM michchap: you can sleep tomorrow morning :)
16:33 xingchao see you guys :)
16:36 shakamunyi joined #puppet-openstack
16:36 mattymo EmilienM, why did you think he pushed?
16:37 EmilienM mattymo: jenkins is not happy
16:38 EmilienM i misread and thought about a new patchset
16:38 mattymo EmilienM, well, I would be unhappy if I got a -1 from jenkins too
16:39 EmilienM mattymo: more when test pass locally
16:39 * mattymo pats EmilienM. It's going to be okay, buddy.
16:40 EmilienM lol
16:40 EmilienM mattymo: btw, when we can see some patches from you guys? :)
16:41 mattymo EmilienM, this week is still design phase! We just had a meeting about discussing which modules we'll each take
16:42 mattymo I think I'll take nova. xenolog/Sergey neutrno, xarses ceph/cinder (or maybe me cinder).
16:42 mattymo the rest are still up for debate
16:43 mattymo there's a lot in neutron this round to merge, mainly focused on 3rd party plugins
16:43 mattymo not quite so exciting
16:43 xarses mattymo: i thought you cleaned up cinder mostly already
16:43 mattymo I did but then stopped
16:43 mattymo I will pick it up in a few days and replay gitlog and see where it lands
16:46 daneyon joined #puppet-openstack
16:46 rwsu joined #puppet-openstack
16:52 berendt joined #puppet-openstack
16:53 dtalton joined #puppet-openstack
16:55 mmaglana joined #puppet-openstack
17:00 aldavud_ joined #puppet-openstack
17:02 kkirkpatrick joined #puppet-openstack
17:04 mmaglana hogepodge: will be working on that in-memory configuration bit for Aviator this week. also, I don't know if this is a blocker but I'm still in the process of merging downstream code that adds Keystone v3 API support to aviator. I estimate 2 weeks of work for that.
17:04 hogepodge Thanks mmaglana Also, any thoughts on how to break out avaiator-lite to reduce dependencies would also be great. Packaging with dependencies is going to be a neat trick.
17:05 mmaglana hogepodge: i was thinking about that too on my way to the office.
17:06 mmaglana i think i can easily pull out the cli part into aviator-cli or something. i'm also planning to remove its dependency on activesupport but i don't know how much work that will entail. i want to do it though since i'm not too fond of HashWithIndifferentAccess anymore.
17:07 hogepodge heh
17:11 mmaglana ...and when that's done, i'll start thinking about removing aviator's dependency on faraday so that it will just use built-in classes at production time. but...that's far into the future right now.
17:20 mmaglana joined #puppet-openstack
17:24 bodepd after I wrote that email, I also went ahead and just reviewed what was there: https://review.openstack.org/#/c/96627/
17:25 bodepd mgagne: woudl I be crazy if I revived rspec-system from the dead instead of using beaker?
17:25 bodepd mgagne: I'm very concerned with the PE specificness of beaker, the fact that it already works for puppet-ceph, as well of the fact that it seems like a tool that does everything related to PE testing.
17:26 spredzy_ mgagne, when you have few mins I'd like your review on https://review.openstack.org/#/c/97243/
17:27 xarses joined #puppet-openstack
17:29 dmsimard bodepd: PE stands for.. ?
17:29 bodepd Puppet Enterprise
17:29 dmsimard Oh, right.
17:31 bodepd if someone else wants to weigh in, that would be great. I'm also a little bit concerned that I will have to make lots of changes to whatever tool I pick to make it work for me (things like adding hiera support)
17:32 bodepd if I just grabbed rspec-system and *took it over* I could def have a greater change velocity. However, there is risk that beaker will add features that I need.
17:32 bodepd someone raised something about the risk of community fragmentation, but I think that issue already exists (if you look at download stats, more people use rspec-system than beaker)
17:32 xarses mattymo: i thought you cleaned up cinder mostly already
17:33 hogepodge bodepd rumor has it you’re working on public CI?
17:34 bodepd I need to build CI
17:34 bodepd so I might as well gate with it and make a push for embedded integration tests at the same time
17:37 aldavud one issue with integration tests currently is runtime
17:38 aldavud puppet-ceph already takes 1.5h for a complete run.
17:38 zastern joined #puppet-openstack
17:40 anode joined #puppet-openstack
17:40 anode i heard that people were talking about beaker in here… and i can’t resist that :)
17:41 aldavud :)
17:41 aldavud so... this: https://github.com/puppetlabs/beaker/issues/155
17:41 Hunner bodepd: We're trying to keep the puppet stuff as separate from the VM management stuff as possible so that beaker never requires puppet
17:42 Hunner Short of actually separating it into a separate gem like rspec-system-puppet did, though
17:42 anode aldavud: cool, i’ll get that specced out and into next weeks sprint (or this weeks, if i can get my other work done)
17:42 Hunner But you can actually use beaker to test chef/shellscripts/whatever if you want :)
17:43 Hunner beaker is already beyond the abilities of rspec-system, and we would love to have more external input on its adoption :)
17:44 Hunner bodepd: Also, have you seen beaker-rspec? It's almost exactly the same DSL as rspec-system, but beaker's vm wrangling abilities is better
17:44 bodepd Hunner: it doesn't have openstack support, and it just feels really really heavy.
17:45 bodepd Hunner: isn't that the gem that has absolutely no documentation
17:45 Hunner bodepd: Can you quantify "heavy"? Perhaps we can slim it up :)
17:45 Hunner And I believe openstack support is actually on the list
17:45 bodepd Hunner: if you go to the docs, it mostly talks about PE testing
17:45 Hunner Speaking of which, anode: https://github.com/puppetlabs/rspec-system/blob​/master/lib/rspec-system/node_set/openstack.rb
17:46 anode Hunner: we have plans to roll openstack into beaker, we just don’t have a date yet
17:46 bodepd Hunner: I'm very concerned about basing community stuff on a tool that is so tied to a commercial prodcut
17:46 Hunner bodepd: That's one of my issues with it too... the testing framework usage gets tangled up with the PE testing usage, but that's because we haven't sicced the docs team on it yet...
17:47 Hunner bodepd: I think we're in agreement about that :)
17:47 bodepd Hunner: I did a doc comparison between the two: rspec-system looks like exactly what I want
17:47 anode yes, beaker supports PE.  but beaker is also the tool being used for supported modules.
17:47 bodepd Hunner: it looks like what I want may or may not be hidden in beaker
17:47 bodepd anode: I'm not working on PE supported modules thoug, I'm working on community stuff
17:48 Hunner bodepd: I've tried to get 100% of the capabilities in rspec-system to be available in beaker (where they made sense), so the docs may lie about available functionality
17:48 bodepd I'm also a little concerned about velocity. I probably need a pretty large set of change applied to whatever tool I choose
17:48 anode you may want to talk to electical in puppet-dev, he’s got a whole separate CI system doing testing of his own module using beaker
17:48 anode and was integral in our incorporation of docker support
17:48 bodepd what is an example of a feature related to module testing in beaker what I can't get from rspec-system?
17:49 anode maintenance and support? ;)
17:49 anode we’ve deprecated rpsec-system, and will no longer be doing any support or accepting new PRs
17:49 bodepd anode: I'm comparing beaker to reviving rspec-system and maintaining it
17:51 Hunner multi-node support in rspec-system doesn't really work
17:51 aldavud https://github.com/puppetlabs/beaker/wiki​/How-to-Write-a-Beaker-Test-for-a-Module <- looks like porting tests from rspec-system-puppet is straight forward though
17:51 djm68 joined #puppet-openstack
17:51 bodepd Hunner: I'm sure I'll have to write my own stuff for multi-node support.
17:52 bodepd Hunner: b/c it will use hiera data to define roles
17:52 Hunner It's very straight forward; basically change puppet_apply() to apply_manifest() and update your nodesets :P
17:54 Hunner I'd be really sad if beaker didn't fit your needs as a community member, since that is one of our goals :(
17:55 bodepd Hunner: the documentation leads me to believe that it does not fit my needs
17:55 bodepd Hunner: especially compared to the README from rspec-system.
17:55 bodepd Hunner: community support is a goal thoug?
17:55 Hunner Yes!
17:55 bodepd (I got the impression it either was or would become a PE thing)
17:55 hogepodge bodepd we’re here, right?
17:55 anode if you could give us a list of the shortcomings that would be great
17:56 anode and yes, we want to have One Harness To Rule Them All
17:56 djm68 bodepd: beaker already has multi-node support
17:56 bodepd the docs make it seem very PE, and make me leary of using it long term for community stuff
17:56 bodepd I'll start with beaker. at least you guys have proven that community usage is important.
17:56 djm68 bodepd: beaker is not PE specific, it is used by the puppet open source team, and modules team
17:57 bodepd I'm a little terrified at the amount of patches that I may need to submit though
17:57 djm68 bodepd: beaker is a fully open sourced project now
17:57 djm68 bodepd:vs the amount of work need to write your own?
17:57 bodepd djm68: vs. the amount of work to patch those into rspec-system
17:57 djm68 bodepd: I seem to recall going down this path previously
17:58 imcsk8 joined #puppet-openstack
18:01 djm68 joined #puppet-openstack
18:03 anode i also have trouble with the ‘heavy weight’, but because i’d need to know what counts as heavy
18:03 ryanycoleman joined #puppet-openstack
18:03 anode beaker does ride along with support for a big set of hypervisors, but i don’t consider that to be an issue for basic testing.  it’s just code that doesn’t get executed for everyone
18:07 bodepd the dependencies for one. Also, I guess it felt heavy in that most of the stuff it does, was not what I was looking for.
18:07 anode we do pin everything and check for dependency hell before we role out - so it should be pretty self contained
18:08 bodepd I guess that comment mainly comes from a doc comparison between it and rspec-puppet. That one does exactly what I needed vs. the other one doing that as a subset of a much larger set of functionality.
18:08 bodepd uggh. s/rspec-puppet/rspec-system/
18:09 anode that’s the double edged sword of being the one-and-only-one tool, it has to handle a lot of situations, including supporting open source testing
18:09 anode along with PE specific calls
18:09 bodepd yep
18:10 anode though, thinking it over, the specifics to PE are pretty much on the installation front - the rest is all just wrappers around calling ‘puppet’ of any origin
18:10 bodepd as I said before though, I'm happy to at least go through the process of setting up beaker and submitting some patches before I do anything drastic :)
18:10 * anode nods
18:10 anode sure, sounds good
18:11 bodepd patch acceptance is probably the biggest concern, but it looks like most of those will be in the puppet plugin for beaker
18:11 anode beaker people mostly hang out in #puppet-dev, so you can direct questions there or to me directly
18:11 bodepd and stop highjacking #puppet-openstack :)
18:12 anode exactly :)
18:13 anode if it helps, i did a beaker 101 talk at the Portland PUG that was recorded… though it is 101 level and thus may not have much of interest to you
18:19 bodepd anode: yeah, I'm sure I can get that far. I've already read the docs.
18:19 anode cools
18:20 rstarmer joined #puppet-openstack
18:21 Mr_D joined #puppet-openstack
18:45 openstackgerrit Lars Kellogg-Stedman proposed a change to stackforge/puppet-ceilometer: Fixes agent_notification_service_name  https://review.openstack.org/95519
18:47 djm68_ joined #puppet-openstack
18:48 EmilienM mgagne: maybe you missed it but I let you a comment on https://review.openstack.org/#/c/92993/
18:48 ryanycoleman joined #puppet-openstack
18:49 djm68__ joined #puppet-openstack
18:57 djm68 joined #puppet-openstack
19:01 ken_barber Hunner: multi-node in rspec-system did work btw. Lets not make up stuff :-).
19:02 kkirkpatrick joined #puppet-openstack
19:04 Hunner It did work. There might have been parts that were rough, but maybe until they were fixed :)
19:18 kitp joined #puppet-openstack
19:19 djm68 joined #puppet-openstack
19:23 thumpba_ joined #puppet-openstack
19:24 mgagne I'm having serious issues with IRC, I missed 52 messages :-/
19:24 djm68 joined #puppet-openstack
19:31 EmilienM mgagne: any bouncer?
19:34 mgagne EmilienM: yep
19:34 EmilienM mgagne: maybe network outage on your bouncer server
19:34 mgagne EmilienM: I'll blame rackspace =)
19:35 dmsimard mgagne: blasphemy !
19:41 djm68 joined #puppet-openstack
19:42 spredzy_ guys quick question
19:42 spredzy_ I just realized rspec-puppet now installs rspec 3.0.0
19:43 spredzy_ did any of you face speficic issues with that move ?
19:45 aldavud joined #puppet-openstack
19:45 jkrish joined #puppet-openstack
19:54 EmilienM Hunner: hey, any idea about this ^ ?
19:55 spredzy_ EmilienM, Hunner took it on #puppet
19:55 EmilienM spredzy_: oops, my bad
20:11 openstackgerrit joined #puppet-openstack
20:13 kkirkpatrick joined #puppet-openstack
20:14 aldavud joined #puppet-openstack
20:14 anode joined #puppet-openstack
20:20 hogepodge joined #puppet-openstack
20:24 djm68 joined #puppet-openstack
20:30 kitp joined #puppet-openstack
20:32 thumpba_ joined #puppet-openstack
20:46 djm68 joined #puppet-openstack
20:49 rstarmer joined #puppet-openstack
20:58 djm68 joined #puppet-openstack
20:58 thumpba_ joined #puppet-openstack
21:01 thumpba_ joined #puppet-openstack
21:05 thumpba__ joined #puppet-openstack
21:09 abramley joined #puppet-openstack
21:09 abramley left #puppet-openstack
21:10 Hunner I didn't see any issue
21:10 Hunner Well, I don't know if I was using rspec 3.0 yet :P
21:11 ricardorocha joined #puppet-openstack
21:13 thumpba joined #puppet-openstack
21:17 admin0 joined #puppet-openstack
21:20 mgagne hogepodge: won't be easy for aviator :-/
21:26 hogepodge :-/
21:29 thumpba joined #puppet-openstack
21:29 hogepodge mgagne: mmaglana is already thinking about ways to ease that pain
21:29 mjblack joined #puppet-openstack
21:29 mgagne hogepodge: oh cool. Did I missed it? I had issues with IRC a couple hours ago
21:30 hogepodge Maybe so? Right after the meeting we chatted a little it about breaking out the cli and library bits, and also simplifying some other libraries.
21:30 hogepodge Not too long of a converstaion
21:30 hogepodge s/converstaion/conversation/
21:30 mmaglana what's not going to be easy? :-)
21:31 hogepodge packaging aviator and dependencies to not rely on gems.
21:31 mmaglana ah.
21:31 hogepodge It’s not going to be impossible either.
21:31 mmaglana aimon: is helping me with that right now.
21:31 mmaglana the bigger problem we're fixing right now is compatibility with ruby 1.8
21:31 djm68 joined #puppet-openstack
21:32 mmaglana 1.8 does not know this syntax ==> {key: value}
21:32 mmaglana that just needs some of aimon's sed fu
21:33 djm68 joined #puppet-openstack
21:35 bodepd did we ever sort out openstack_common vs. openstack_lib for reusable bits like repos? I want to go ahead and just submit the patches for that repo
21:35 dvorak it came up a bit in the meeting earlier today
21:36 bodepd darn. are those recorded somewhere?
21:36 dvorak url in the topic
21:37 bodepd I'll just assume common and post the patch to this channel to give folks a chance to argue.
21:38 bodepd I'm going to start it out as blank. Folks can feel free to add whatever patches they want. I'm just going to start by putting the repo code in there
21:42 dmsimard Awesome to see you back in action here bodepd :)
21:42 bodepd dmsimard: mostly just causing trouble
21:43 bodepd even after reading the meeting notes, I'm still a little confused about the library naming.
21:43 bodepd it looks like is
21:44 bodepd it was decided as openstacklib, but it will contain both deps of modules as well as things that compose them, or am I confused?
21:51 hogepodge bodepd openstacklib should provide common functionality for basic tasks; setting up endpoints and databases for example. Repeated code across the modules needs to be refactored there
21:51 djm68 joined #puppet-openstack
21:51 hogepodge common is meant to do composition tasks
21:52 hogepodge common (a bad name) depends on puppet-* depends on puppet-openstacklib
21:52 bodepd https://review.openstack.org/#/c/97357/
21:53 bodepd I don't like common for that purpose. That seems synonomous with lib for me
21:54 bodepd michchap: ^^ check out that patch and see if it doesn't cause you too much pain to have it start as an empty repo as opposed to being based on your work
21:54 hogepodge ^^ (bad name)
21:54 hogepodge Floor is open for suggestions.
21:56 bodepd not sure if these are much better: helpers, utils, ...
21:56 bodepd uggh, naming is hard
21:57 ryanycoleman joined #puppet-openstack
21:59 sileht joined #puppet-openstack
22:02 ryanycoleman joined #puppet-openstack
22:05 hogepodge We could go the core route and just pick a non sequitur as a name. puppet-openstack_voyager for example
22:06 sileht joined #puppet-openstack
22:08 hogepodge Or just go all commercial, sell the name to the highest bidder. puppet-openstack_bodeco can be yours for a modest donation.
22:08 dmsimard get a sponsor
22:08 dmsimard and call it like puppet-openstack_coke or something
22:08 dmsimard :D
22:13 thumpba joined #puppet-openstack
22:16 thumpba_ joined #puppet-openstack
22:20 bodepd puppet-openstack_make_that_shit_useful_in_prod
22:20 xarses joined #puppet-openstack
22:20 bodepd is there are char limit for module names though?
22:22 xarses joined #puppet-openstack
22:39 dvorak bodepd: I suggested openstack_extras, or openstack_supp but that doesn't really capture what it does either
22:39 dvorak but I think either of them are better than common :)
22:39 admin0 joined #puppet-openstack
22:43 bodepd dvorak: I don't hate extras.
22:43 bodepd supp needs to be expanded to make it clearer. I assume supplemental?
22:47 dvorak yeah, but that's too long for a module name imo
22:49 bodepd dvorak: anyways +1 for extras
22:53 djm68 joined #puppet-openstack
22:55 rickerc puppet-openstackcomp is getting long but at least sorta says what it is
22:56 dvorak puppet-openstack_bottomglue and puppet-openstack_topglue
22:56 dvorak for the support and overarching glue pieces
22:56 dvorak ^^ that's a joke btw.
22:57 rickerc dvorak: at least do over and under to hopelessly confuse people hopping between this and tripleo
23:36 openstackgerrit Donald Talton proposed a change to stackforge/puppet_openstack_builder: Temporary fix to correct ml2 bug.  https://review.openstack.org/97368
23:47 hogepodge there you have it rickerc and bodepd, puppet-openstack_tripleo, hopeless confusion
23:49 hogepodge composer?
23:50 hogepodge federation?
23:55 aldavud joined #puppet-openstack

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary