Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-06-04

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

All times shown according to UTC.

Time Nick Message
00:21 mmaglana joined #puppet-openstack
00:23 ryanycoleman joined #puppet-openstack
00:32 shakayumi joined #puppet-openstack
00:45 michchap aimon: yes
00:47 michchap aimon: that makes sense
00:47 morazi joined #puppet-openstack
00:47 bodepd michchap: and I were having on offline conversation that's best had here.
00:47 bodepd michchap: do repo setups belong in extras of lib. I had assumed they should be in lib
00:48 bodepd b/c they are required for performing individual acceptance testing of the individual components
00:48 michchap bodepd: I put a little thought to it and I think that we need a way to separate integration testing requirements from module requirements, because otherwise we end up with neutron requiring nova and vice versa
00:48 bodepd everything requires keystone
00:48 bodepd if they require it, they require it
00:49 bodepd Even beyond integration testing
00:49 michchap everything requires keystone, but an integration test for neutron or cinder or heat probably needs a whole cloud
00:49 bodepd the defaults deps should install things so that they will be generally useful
00:49 bodepd maybe this goes back to the argument of should everything just be in the same repo
00:50 michchap yeah it really does - we end up with all to all depepndencies, or close to it. I'm not sure if that's desirable.
00:50 dvorak seems kind of unavoidable
00:50 dvorak at least for integration tests
00:51 michchap but only for integration tests - they aren't module dependencies are they?
00:51 ozialien joined #puppet-openstack
00:51 bodepd michchap: I think they are.
00:51 michchap if we don't have a way to distinguish the two then I guess we just have to make the choice
00:51 dvorak dunno, I've never installed them with anything that does dep resolution :)
00:51 mmaglana joined #puppet-openstack
00:51 bodepd michchap: the user experience is pretty close to integration testing
00:52 bodepd dvorak: me niether. I even create librarian-puppet-simple for that exact purpose :)
00:52 bodepd michchap: I could go either way.
00:52 bodepd michchap: concensus woudl be nice, b/c I've already started on this stuff
00:52 rmoe joined #puppet-openstack
00:52 michchap bodepd: I don't think we have another choice do we?
00:58 bodepd markvoelker: hey, can we start making massive chnages to the openstack_builder?
00:58 bodepd markvoelker: do you guys want to cut an icehouse branch?
01:01 michchap bodepd: so the only real downside for having many dependencies for the modules is that if someone is installing from the forge and deploying multi-node, they can't just configure something like a network node without bringing in almost all of the modules for other services, even if they aren't needed.
01:01 michchap bodepd: I think that's a pretty minor issue
01:06 bodepd michchap: ah, deps as per node as opposed to per deployment
01:08 mmaglana joined #puppet-openstack
01:09 michchap bodepd: In that case extras and lib are both going to be dependencies, as you say.
01:11 bodepd michchap: I was thinking that everything depends on lib and extras depends on everything
01:11 bodepd michchap: in my mind that is how I was splitting things out
01:11 michchap bodepd: unless we define extras as being something that's never needed by any integration tests, but that rules out a lot of things
01:13 michchap bodepd: for example I don't think it's unreasonable to have an integration test with the firewall active. That means firewall rules now need to be in lib, and we have to bring in puppetlabs-firewall.
01:13 bodepd michchap: is that bad?
01:14 bodepd michchap: is that a reasonable line in the sand to decide what goes where?
01:14 michchap bodepd: It's probably also valid to have keystone behind haproxy as an integration test, so now load balancing rules are in lib
01:14 dvorak wouldn't you have integration tests specifically for extras functionality like that?
01:14 michchap bodepd: I don't think there's going to be much left in extras.
01:14 bodepd michchap: I would see those as being 'openstack' integration tests
01:15 bodepd yes there will: ha, monitoring, log aggregation
01:15 bodepd different types of ha, monitoring stuff, etc
01:16 michchap ah ok. None of that exists today afaik. The extras style module I have right now covers: repo mgmt, firewalling, load balancing+keepalived, provider networks fixups, and compatibility for older openstack releases.
01:17 xarses joined #puppet-openstack
01:17 michchap There was sentiment against me moving that into a lib module if it was going to be a dependency of the other modules.
01:17 bodepd what do you mean?
01:19 dvorak well, a lot of those things inherently have to be fairly opinionated
01:19 michchap When I proposed openstacklib as what it currently is, the response was that it wasn't appropriate.
01:19 dvorak repo mgmt makes sense to me to be in lib, but the rest of it not so much
01:20 dvorak for example, some people will use hardware load balancers, or LVS, and it's not fair to assume haproxy
01:21 michchap If we're saying that adding module dependencies that may not be needed isn't a big deal, why are we even proposing two repos for 'other' functionality.
01:21 bodepd rspec 3.x is a real pain
01:21 bodepd has anyone been mucking with that?
01:21 michchap I saw it yesterday and sighed loudly =/
01:21 bodepd did that just start happening?
01:22 bodepd dvorak: that is my argument. well actually I only care about repo management atm
01:22 michchap I guess 3 just came out
01:22 bodepd dvorak: b/c I can't test anything without those
01:22 dvorak well, repo management isn't really opinated and you can pretty easily turn it on or off
01:22 bodepd michchap: maybe a bug needs to be opened against spec_helper to not download 3?
01:23 dvorak although I'd really like to avoid a [repoclass] -> Package<||> type of dependency
01:24 michchap dvorak: we could possibly get around that by tagging every single package resource, but it sounds like a lot of work.
01:24 dvorak michchap: I don't think it'd be that bad really.  class level resource defaults could be used to handle the few classes that have lots of packages
01:25 dvorak could probably put together a rspec test to find all the packages that don't have a tag also
01:25 openstackgerrit A change was merged to stackforge/puppet-ceph: obsolete cuttlefish and add firefly  https://review.openstack.org/96451
01:25 michchap dvorak: I'm not against making that change if it makes repo mgmt easier.
01:25 bodepd michchap: https://github.com/bodepd/puppet-openstacklib
01:25 bodepd michchap: that is my place holder until the repo is created
01:26 michchap bodepd: https://github.com/michaeltchapman/puppet-​openstacklib/blob/master/manifests/repo.pp
01:26 bodepd michchap: I looked at that
01:26 bodepd michchap: and grabbed some stuff from it
01:26 michchap I needed proxy support
01:26 bodepd michchap: ah, you already noticed that I did not pull that in :)
01:27 bodepd michchap: I was thinking that should be in the rdo/uca classes
01:27 bodepd specifically, I want to get rid of as many compostions classes as possible
01:27 bodepd I would rather have something with smarts compose it for you
01:28 bodepd I did look at that method. Is there somewhere sensical to move it to?
01:28 xarses joined #puppet-openstack
01:28 bodepd michchap: there is probably something to be said for just grabbing it as-is, but now is a great time to change it
01:29 michchap bodepd: which method?
01:31 bodepd s/method/class/
01:31 bodepd repo.pp
01:32 michchap yep. I'm happy to remove it and use mappings to push out common variables.
01:32 bodepd michchap: yes.
01:33 michchap oh epel was irritating to deal with
01:33 kitp joined #puppet-openstack
01:33 michchap because lots of modules just define it themselves
01:34 michchap the other thing is that for the customer I deployed at, they had a local mirror of rdo anyway, so they weren't using the repo class.
01:35 bodepd we can add something if we need it
01:35 bodepd maybe that goes in extras ;)
01:48 djm68 joined #puppet-openstack
01:57 xingchao joined #puppet-openstack
02:06 mmaglana joined #puppet-openstack
02:10 kitp joined #puppet-openstack
02:19 gildub I'd like to get review/advice for github.com/purpleidea/puppet-gluster.git - My concerns is the systematic use of Exec instead of types/providers. Anyone?
02:33 openstackgerrit A change was merged to stackforge/puppet-ceph: Preparation for scenario_node_terminus tests  https://review.openstack.org/96580
02:35 djm68 joined #puppet-openstack
02:43 starmer joined #puppet-openstack
02:46 mmaglana joined #puppet-openstack
03:32 comptona joined #puppet-openstack
03:32 djm68 joined #puppet-openstack
03:43 starmer joined #puppet-openstack
04:07 kitp joined #puppet-openstack
04:10 daneyon joined #puppet-openstack
04:19 gildub Anyone to merge already reviewed: https://review.openstack.org/#/c/94066 ?
04:22 mgagne gildub: I'm not familiar with the inner working of tempest yet and won't be able to merge it tonight without an other +2. Where can I find info about the required packages for tempest?
04:25 gildub mgagne, I don't know, I've never used tempest myself. Just trying to get that merge for dependent bug.
04:29 gildub mgagne, I doesn't look like there is any openstack tempest related packages available at least on Fedrora/CentOS/RHEL
04:29 mgagne gildub: ok, I looked at old notes I had regarding the installation of tempest and database dev libraries aren't in the list of packages I had to install
04:31 gildub mgagne, ah yeah, I checked that too.
04:31 michchap gildub: ugh that gluster module is AGPL :(
04:31 mgagne gildub: I +2 the change
04:31 gildub magne, great, thanks!
04:35 gildub michchap, this is supposed tobe FOSS but maybe not explicitly AGPL
04:36 michchap gildub: the feature list is certainly impressive.
04:37 michchap gildub: some of it is very strange to me, like this class: https://github.com/purpleidea/puppet-g​luster/blob/master/manifests/again.pp
04:39 gildub michchap, the way I understand it, the author has worked out orchestration puppet issues this way. He mentioned that in his blog: http://ttboj.wordpress.com/2014/03​/24/introducing-puppet-execagain/
04:39 mgagne michchap: hahaha, I like it
04:40 michchap gildub: that's a very interesting solution to the orchestration problem :)
04:40 mgagne michchap: I like this section "Help, it won’t stop!"
04:41 daneyon joined #puppet-openstack
04:42 michchap mgagne: haha yep
04:43 gildub michchap, mgagne, besides advanced features like that, the module is relying on Exec. That works fine. But my concern is for this to evolve more to more generic modules using, where possible, suitable Types/Providers.
04:44 gildub The author point is that on the long run, types/providers are gone. I don't know about that but for now my concern is to fit within the rest and at least some sort of standards.
04:44 gildub michchap, mgagne, what do you guys think?
04:45 mgagne gildub: I should read more than the headers
04:46 michchap gildub: I'm not sure about types/providers being gone in the long run. Is there a blog post about that as well?
04:46 gildub michchap, mgagne, when I said relying on Exec, I mean there are a *lot*
04:46 michchap gildub: but to be honest I have no issue with properly written exec resources.
04:46 gildub mgagne, fair enough!
04:47 mgagne gildub: why isn't he writing a Makefile and trigger its execution in a cron then?
04:47 mgagne gildub: shell scripts + dependencies
04:47 gildub michchap, I haven't find any discussion about that, but maybe purpleidea has a pointer, will ask him.
04:47 michchap gildub: I guess he would need all the file system and block device layer stuff to be written as native types. I have looked at that in the past and it's a large task.
04:48 gildub mgagne, that's a very good point!!!
04:48 mgagne gildub: I didn't read anything about types/providers. Is it in an other blog post?
04:49 mgagne gildub: or some random rant he made on twitter?
04:49 gildub michchap, this is why I'm asking, my understanding of puppet philosophy, at least when going generic enough, is by this type/provider/function refactorisation. But end of the day, I'm trying to fit in.
04:50 michchap gildub: I think your understanding is correct - PuppetLabs will always recommend using native types over exec.
04:50 mgagne gildub, michchap: why can't a wait logic be used just like the one in puppetdb? Is it because some resources aren't available until the next run?
04:51 michchap mgagne: This is exactly what I was talking about at the summit. See: https://github.com/stackforge/puppet-keyst​one/blob/master/manifests/service.pp#L108
04:51 gildub michchap, ok so I should say "recommendation" not "standard".
04:51 michchap mgagne: the difference is that his method is going to converge, mine might timeout and fail.
04:51 mgagne michchap: yep, I understand that specific use case. Is it the same for gluster?
04:51 michchap mgagne: but if you want to fail fast that seems desirable.
04:52 mgagne michchap: One example I can think of is disk partitions not available or discovered by puppet until the next run
04:53 gildub mgagne, this gluster module works with puppetdb, but it also has to works without, that's one of the use case for me :(
04:53 michchap mgagne: why wouldn't they be, though. Either it's a matter of time, or it requires another action (like a reboot). Both methods are reliant on it being a matter of time.
04:53 mgagne michchap: +/-
04:53 mgagne michchap: aren't resources discovered once when the provider is initialized?
04:54 gildub mgagne, exactly, setting up gluster server is a one of many cluster scenario
04:54 gildub ^server(s)
04:54 mgagne gildub: if you need broader node orchestration, puppet won't be able to do all of the job by itself.
04:55 mgagne gildub: you will probably need another tool for that
04:55 gildub mgagne, well purpleidea has managed to get the cluster working with this approach
04:56 mgagne gildub: with again?
04:56 gildub mgagne, with this gluster module
04:58 mgagne gildub: sure. If your plan is to touch your left ear, you can do it with your right hand by passing it over your head. Or you can touch it with your left hand. It just happens that one method is silly and not the best.
04:58 michchap mgagne: Right. The resources are part of the catalog, but their state can be discovered at any point - we were having this issue with keystone, right?
04:59 mgagne michchap: I was referring to the idea or retriggerring a puppet run from within puppet with again
05:00 michchap mgagne: ok, so on the second run there might be a different catalog? That doesn't make sense though - the first one isn't going to change the manifests, just the system state.
05:00 gildub mgagne, nice analogy!
05:01 michchap mgagne: but the prefetching will happen again, so if you're using native types you could paper over issues where data is only loaded once at that stage and never updated
05:01 mgagne gildub: I have no credit for that one, our netadmin came up with it to explain how silly our ideas were.
05:01 michchap mgagne: or am I totally misunderstanding you...
05:02 gildub michchap, sounds like puppet metaprogramming!
05:02 mgagne michchap: I'm just saying that because it works doesn't make it the best and most elegant way to fix an issue.
05:03 gildub mgagne, Exactly I'm concerned that using too much Exec is going something very hard to maintain down the road. Now maybe this a catch 22 situation, because without Exec maybe there isn't any gluster module.
05:03 michchap mgagne: I think we agree that it's not a good idea.
05:03 mgagne michchap: yep
05:03 michchap gildub: you'll need exec for anything storage related.
05:03 michchap gildub: It's just how things are until someone writes a really nice interface for mkfs and parted uisng native types
05:03 mgagne gildub: unfortunately, from my reading, the blog post does not explain what use case required the use of again
05:04 mgagne gildub: so I can only make supposition
05:04 gildub michchap, sure but that's the role of type/provider, they can do %x
05:04 shakamunyi joined #puppet-openstack
05:05 michchap gildub: Are you deciding whether to use this module or write your own?
05:06 gildub michchap, I was trying to convince purpleidea to maybe doe less Exec :(
05:06 gildub michchap, it's a lost cause
05:08 michchap gildub: Would he accept patches using native types?
05:08 kitp joined #puppet-openstack
05:08 gildub michchap, mgagne, one related issue is about using installing python modules to be able to run things like this:
05:08 gildub https://github.com/purpleidea/puppet-glus​ter/blob/master/manifests/volume.pp#L203
05:09 gildub michchap, I believe he will, he said => you're of course welcome to patch this or fork it, but i won't be working on those features or merging them upstream. tbh, i doubt they will work the way you're expecting
05:10 gildub michchap, by upstream I understand puppetlabs/stackforge
05:10 michchap gildub: if he says he won't merge them upstream...
05:11 michchap gildub: this is on stackforge?
05:11 gildub michchap, well I need to make sure we understand each other properly on this one, he considers his module upstream but I said there upstream to it as stackforge
05:12 gildub michchap, I don't think so...
05:12 gildub michchap, there is an entry in puppet forge
05:13 michchap gildub: ah that makes more sense.
05:15 daneyon_ joined #puppet-openstack
05:16 gildub michchap, what do you think about the python library I mentioned ?
05:17 michchap gildub: putting scripts in files, setting them as a resource somewhere and then running them is pretty much the opposite of what I think is good practice in puppet, but maybe he has a good reason.
05:19 michchap gildub: I don't know enough about gluster to give a very good answer as to whether it's a good decision or not. Without context I would say it's a bad one, but maybe it's necessitated by something in gluster.
05:19 gildub michchap, I agree with you. well I suppose once on the Exec road, doesn't matter much. What bothers me most is having to have to pull other packages just to be able to run the module but that has nothing to do with the installed product. Ideally it should even remove it at the end.
05:20 michchap gildub: yep. The risk of losing idempotency is extremely high in that situation.
05:21 gildub michchap, well the module needs python-argparse to filter exec commands of gluster -xml for things like status.
05:23 mmaglana joined #puppet-openstack
05:23 michchap gildub: right.
05:25 gildub michchap, what about renaming it ansible-gluster?
05:25 gildub michchap, and then add a puppet-ansible to use it? LOL
05:26 michchap gildub: Write some native types for ansible so we can do orchestration from puppet :)
05:32 gildub michchap, mgagne, trying to get a pointer from purpleidea about this 'types/providers gone in the long run' discussion
05:33 gildub michchap, will try to clarify the upstream/stackforge side of things along recommendations with him too. But I understand this is really tricky with storage.
05:33 gildub michchap, mgagne - Thank you for your help and time!
05:34 mgagne =)
05:34 michchap gildub: no problem. good luck 'fixing' the module :)
05:34 gildub :)
05:44 mgagne michchap: hardest part about starting a blog is always finding a good theme/skin
05:44 michchap mgagne: agree. Also deciding on a SSG
05:45 mgagne michchap: I guess SSG doesn't mean SevenSuperGirls
05:45 michchap mgagne: static site generator?
05:45 mgagne michchap: oh, I found one or two
05:45 michchap mgagne: I used middleman for stacktira.aptira.com - I found it quite nice
05:48 mgagne michchap: I'm planning in using pelican this time
05:48 mgagne on*
05:53 michchap mgagne: looks nice
05:54 xingchao mgagne: are you planning to write blog ?
05:54 mgagne xingchao: trying (for months)
05:54 xingchao mgagne: great :0
05:55 mgagne xingchao: I used to write articles and tutorials for PHP (I know) years ago but stopped writing since =(
05:56 djm68 joined #puppet-openstack
05:56 xingchao mgagne: keep writing is a good habit, it's quite right to pick it up again.  lol
05:57 mgagne xingchao: =)
05:59 tfz joined #puppet-openstack
05:59 bodepd anything interesting going on?
06:02 mgagne it's raining outside =)
06:08 michchap I am testing this: http://techs.enovance.com/6797/how​-to-use-edeploy-roles-with-vagrant
06:09 kitp joined #puppet-openstack
06:10 bodepd michchap: interesting...
06:10 bodepd michchap: you can do actually bare metal image deploys with that?
06:11 michchap bodepd: Yeah. I have only tested pxe booting vagrant boxes but I don't see why not
06:12 michchap bodepd: but pushing the roles into vagrant boxes or openstack images makes it easier to test.
06:16 bodepd michchap: define: pushing the roles? You mean the base boxes contain the package installs?
06:17 bodepd michchap: I'm ok with that. I just want to make sure we have packer scripts to build out those boxes.
06:17 bodepd in a way that is configurable
06:17 michchap bodepd: I was typing out exactly that
06:17 bodepd in my head, I know what it looks like
06:17 michchap bodepd: there is a disconnect atm - role definitions should be in only a single place so there is no way to drift.
06:17 bodepd I spent a day playing with packer, have you seen the veewee to packer scripts?
06:18 bodepd michchap: I don't have enough date to follow that point
06:18 bodepd this whole effort is massive. Mayve we should start coordinating some of the tasks.
06:19 michchap bodepd: the roles are built from bash scripts defining what packages are needed from which repos - we want them to come from the catalog so that we have only a single list of packages for a role: the one in puppet.
06:19 bodepd michchap: do you even know of good tools for modeling ths?
06:19 bodepd yeah, that should be pretty easy.
06:20 michchap bodepd: the workflow? I guess I'd just use a vector graphics thing like sketch
06:20 mkoderer joined #puppet-openstack
06:20 bodepd michchap: can you send me a link?
06:21 michchap bodepd: it's not free. http://bohemiancoding.com/sketch/
06:22 michchap bodepd: I use it for diagramming but there might be something free+better
06:23 michchap bodepd: the google docs image thing is ok for diagrams these days if you want something we can both add to
06:24 daneyon joined #puppet-openstack
06:27 bodepd I have omnigraffle for diagrams and I'm darn happy with it
06:27 bodepd I'm just trying to visual all of the tasks and where they meet
06:28 bodepd it would be nice to be able to create a diagram and have each node link to a list of tickets or trello tasks or something
06:28 michchap bodepd: that would be neat.
06:33 aldavud joined #puppet-openstack
06:38 EmilienM good morning
06:49 michchap EmilienM: morning
06:57 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-neutron: Configure OVS Agent when using ML2 plugin  https://review.openstack.org/82353
06:57 artem_panchenko joined #puppet-openstack
07:04 openstackgerrit A change was merged to stackforge/puppet-ceph: Initial classes/tests for a roles/profiles pattern  https://review.openstack.org/94195
07:10 kitp joined #puppet-openstack
07:20 aldavud dmsimard: I just commented on the use of tags
07:22 aldavud dmsimard: let me know if that makes sense to you. https://review.openstack.org/96635
07:48 aldavud joined #puppet-openstack
07:50 aldavud_ joined #puppet-openstack
07:59 derekh_ joined #puppet-openstack
08:03 tfz joined #puppet-openstack
08:10 kitp joined #puppet-openstack
08:18 arnaud_orange joined #puppet-openstack
08:21 social__ Hi, anyone around to approve this backport? https://review.openstack.org/#/c/97195/
08:22 social__ EmilienM: ^^ please
08:34 sanjayu joined #puppet-openstack
08:36 djm68 joined #puppet-openstack
08:41 mmagr joined #puppet-openstack
08:44 EmilienM social__: sure!
08:45 EmilienM social__: done
08:46 social__ EmilienM: thanks a lot
08:48 rmoe joined #puppet-openstack
08:48 sanjayu joined #puppet-openstack
08:48 mkoderer joined #puppet-openstack
08:48 goneri joined #puppet-openstack
08:49 rickerc joined #puppet-openstack
08:50 openstackgerrit joined #puppet-openstack
08:51 openstackgerrit A change was merged to stackforge/puppet-cinder: Enable cinder server to be run in SSL mode  https://review.openstack.org/94395
08:52 openstackgerrit A change was merged to stackforge/puppet-nova: Enable nova server to be run in SSL mode  https://review.openstack.org/94399
09:00 openstackgerrit A change was merged to stackforge/puppet-swift: Add client_package_ensure parameter and swift::client class  https://review.openstack.org/96287
09:03 xenolog joined #puppet-openstack
09:04 openstackgerrit A change was merged to stackforge/puppet-swift: Include swift::params in swift::client  https://review.openstack.org/96284
09:04 openstackgerrit A change was merged to stackforge/puppet-swift: Include swift::params in swift::client  https://review.openstack.org/96289
09:06 shakamunyi joined #puppet-openstack
09:11 dgollub joined #puppet-openstack
09:11 kitp joined #puppet-openstack
09:36 shakamunyi joined #puppet-openstack
09:37 aldavud_ joined #puppet-openstack
09:52 holser joined #puppet-openstack
10:12 kitp joined #puppet-openstack
10:37 shakamunyi joined #puppet-openstack
10:58 berendt joined #puppet-openstack
11:12 kitp joined #puppet-openstack
11:14 openstackgerrit A change was merged to stackforge/puppet-swift: Include swift::params in swift::client  https://review.openstack.org/96286
11:23 openstackgerrit A change was merged to stackforge/puppet-ceph: Support for scenario_node_terminus for ceph::mon  https://review.openstack.org/96635
11:23 openstackgerrit A change was merged to stackforge/puppet-ceph: cephx support for scenario_node_terminus  https://review.openstack.org/97102
11:30 openstackgerrit A change was merged to stackforge/puppet-tempest: Tempest does not requires DB devel libs  https://review.openstack.org/94066
11:38 shakamunyi joined #puppet-openstack
11:38 sputnik13 joined #puppet-openstack
11:50 openstackgerrit A change was merged to stackforge/puppet-swift: Restore selinux context on swift mount  https://review.openstack.org/97195
11:54 dmsimard merge storm
11:56 openstackgerrit David Moreau Simard proposed a change to stackforge/puppet-swift: Avoid duplication error when ringserver and storage are on the same node  https://review.openstack.org/97763
12:06 morazi joined #puppet-openstack
12:09 EmilienM dmsimard: morning!
12:10 dmsimard EmilienM: oi
12:13 kitp joined #puppet-openstack
12:20 dmsimard EmilienM:  ^ ?
12:20 dmsimard Or you are just being nice? :)
12:20 EmilienM dmsimard: i use to be nice
12:39 shakamunyi joined #puppet-openstack
12:49 dprince joined #puppet-openstack
12:51 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-neutron: Configure OVS Agent when using ML2 plugin  https://review.openstack.org/82353
12:51 djm68 joined #puppet-openstack
13:03 shakamunyi joined #puppet-openstack
13:14 kitp joined #puppet-openstack
13:35 britthouser joined #puppet-openstack
13:41 djm68 joined #puppet-openstack
13:50 djm68 joined #puppet-openstack
13:52 mjblack joined #puppet-openstack
14:15 mmagr joined #puppet-openstack
14:15 kitp joined #puppet-openstack
14:17 djm68 joined #puppet-openstack
14:40 rwsu joined #puppet-openstack
14:41 kkirkpatrick joined #puppet-openstack
14:49 kitp joined #puppet-openstack
14:51 morazi joined #puppet-openstack
14:52 mmagr joined #puppet-openstack
15:00 mgagne Can a core review that one? https://review.openstack.org/#/c/97559/
15:00 mgagne it's about fixing unit tests with RSpec 3.0.0
15:01 EmilienM mgagne: right now
15:01 mgagne EmilienM: cool
15:01 mgagne EmilienM: I though the initiative to create a wiki page for our meeting: https://wiki.openstack.org/wiki/Meetings/Puppet
15:01 EmilienM mgagne: done
15:01 EmilienM cool !
15:03 mgagne EmilienM: we should ask Chris to prepare the next agenda I guess
15:06 openstackgerrit A change was merged to stackforge/puppet-neutron: Fix puppet-neutron rspec test for RSpec >= 3.0.0  https://review.openstack.org/97559
15:24 openstackgerrit A change was merged to stackforge/puppet-tempest: removing the orphaned os_concat function  https://review.openstack.org/96667
15:25 arnaud_orange joined #puppet-openstack
15:26 mmaglana joined #puppet-openstack
15:30 sputnik13 joined #puppet-openstack
15:39 djm68 joined #puppet-openstack
15:47 badiane_ka joined #puppet-openstack
15:50 daneyon joined #puppet-openstack
15:53 anode joined #puppet-openstack
15:56 mmaglana joined #puppet-openstack
15:58 ryanycoleman joined #puppet-openstack
16:06 djm68 joined #puppet-openstack
16:06 anteaya joined #puppet-openstack
16:07 anteaya does anyone know who maintains this gerrit automated account? puppetceph    Puppet Ceph Integration
16:07 anteaya and this one? puppet-openstack-ci-user    Puppet OpenStack CI
16:15 rmoe joined #puppet-openstack
16:28 xarses joined #puppet-openstack
16:34 djm68 joined #puppet-openstack
16:39 mgagne anteaya: dachary is maintaining it
16:39 mgagne anteaya: for Ceph
16:39 mgagne anteaya: for Puppet OpenStack CI, I think it is bodepd
16:40 anteaya mgagne: great thank you
16:40 anteaya dachary and bodepd when you are around
16:40 anteaya we are standardizing format for automated gerrit accounts
16:41 anteaya https://etherpad.openstack.org/p/aut​omated-gerrit-account-naming-format
16:41 anteaya CI for if it votes or ever will and Bot for if it will never vote
16:41 anteaya we can't use openstack in the name
16:42 anteaya I think Puppet Ceph CI should be okay
16:42 anteaya and Puppet CI for the other one?
16:42 anteaya let me know
16:54 djm68 joined #puppet-openstack
16:57 mmaglana joined #puppet-openstack
17:00 comptona joined #puppet-openstack
17:03 aimon__ joined #puppet-openstack
17:06 djm68 joined #puppet-openstack
17:09 djm68 joined #puppet-openstack
17:21 djm68 joined #puppet-openstack
17:24 djm68 joined #puppet-openstack
17:34 aldavud joined #puppet-openstack
17:39 bodepd I was using hte puppet-openstack-ci-user and I plan to use it again
17:39 bodepd anteaya: meme
17:40 bodepd anteaya: if you want to change it, go ahead (is there anything I would have todo?)
17:42 anteaya first we agree to a new name
17:42 anteaya then if you need to do anything on your end, we will let you know
17:42 anteaya we need to get rid of openstack and user
17:42 anteaya puppet-ci okay?
17:48 mtl1 Hi again. Is it a know issue that openstack::auth_file sets OS_TENANT_NAME to "openstack" instead of to "admin" ?
17:50 djm68 joined #puppet-openstack
17:51 bodepd anteaya: I'm totally happy to defer nameing to someone else :)
17:52 anteaya bodepd: well I don't need it to be a long conversation
17:52 anteaya does puppet-ci offend anyone?
17:56 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-keystone: enabling tokens flushing using cron  https://review.openstack.org/97865
17:57 mgagne anteaya: it does not offend me
17:57 ryanycoleman joined #puppet-openstack
17:59 dvorak has anyone else seen an issue with nova::rabbitmq using the old cluster_disk_nodes parameters for rabbit?
17:59 dvorak because of some questionable choices in the rabbitmq module, it's causing puppet to restart rabbitmq every time it runs
17:59 bodepd anteaya: works for me
18:00 anteaya awesome thank you
18:00 mgagne bodepd: nova::rabbitmq is a manifest we should probably move to openstack_extras
18:00 dvorak mgagne: yeah, it struck me as odd that it's there.
18:02 dvorak anyway, the simple fix for this problem is to change nova::rabbitmq to use the new parameter for cluster_nodes, instead of cluster_disk_nodes
18:02 djm68 joined #puppet-openstack
18:04 djm68 joined #puppet-openstack
18:11 bodepd mgagne: I'm still so lost on what goes in extras vs. lib
18:11 mgagne bodepd: lib => type/provider and stuff. extras -> manifests
18:11 bodepd mgagne: needless to say, it shoudl go somewhere.
18:11 bodepd mgagne: ok.
18:11 mgagne bodepd: for example, repo shouldn't go in lib
18:11 bodepd mgagne: so all modules depend on extras and lib
18:12 mgagne bodepd: ideally, lib shouldn't be depend on the version of the modules
18:12 mgagne bodepd: extras should be optional IMO
18:12 mgagne bodepd: otherwise they aren't extras
18:12 bodepd mgagne: repo setup and message queue config don't seem optional
18:12 mgagne bodepd: puppet-nova can work by itself without those
18:13 mgagne bodepd: puppet-nova might not work if lib is not there if we create a base ini provider in there
18:15 ryanycoleman joined #puppet-openstack
18:15 mgagne bodepd: that's my theoretical vision
18:16 mgagne bodepd: if rabbitmq goes in extras, what goes in puppet-openstack?
18:18 shakayumi joined #puppet-openstack
18:19 morazi joined #puppet-openstack
18:23 ryanycoleman joined #puppet-openstack
18:25 bodepd mgagne: my expecation is that dies
18:25 bodepd mgagne: or at least from my perspective
18:26 bodepd maybe it is actually common and not extras then
18:26 bodepd or I guess some of the things are common and some are extras :)
18:26 bodepd it's ok anyways, I can't get any of these repos merged b/c of the sort gate check
18:33 guimaluf joined #puppet-openstack
18:37 imcsk8 joined #puppet-openstack
18:40 sputnik13 joined #puppet-openstack
18:58 Waao joined #puppet-openstack
19:02 aldavud_ joined #puppet-openstack
19:06 openstackgerrit Donald Talton proposed a change to stackforge/puppet_openstack_builder: Temporary fix to correct ml2 bug.  https://review.openstack.org/97368
19:07 dprince joined #puppet-openstack
19:10 anode joined #puppet-openstack
19:18 mgagne bodepd: I should know that is as been replaced by puppetlabs/puppet-openstack (on the forge)
19:19 mgagne bodepd: I guess Chris kind of forgot to tell us about it or I missed the memo: https://forge.puppetlabs.com/puppetlabs/openstack
19:19 mgagne bodepd: https://github.com/puppetlabs/puppet​labs-openstack/tree/master/manifests
19:23 bodepd mgagne: yeah, it was killed, so the assumption is to remove the good bits out of there and out of stuff that was in ciscos repos into extras
19:23 bodepd I already started, but I was putting things that I guess should be in extras (repos) in lib
19:23 mgagne bodepd: what should we do about the new puppet-openstack? it's not in stackforge
19:24 bodepd mgagne: what new puppet-openstack?
19:24 mgagne bodepd: and it's already wrapping around https://github.com/puppetlabs/puppetlabs-opens​tack/blob/master/manifests/profile/rabbitmq.pp
19:24 mgagne bodepd: the one I mentionned 4 lines above
19:24 bodepd mgagne: AFAIC, it's dead
19:24 mgagne bodepd: you* should know that is as been replaced by puppetlabs/puppet-openstack (on the forge)
19:25 bodepd I'm confused, I though Chris hijacked the name for his stuff
19:25 mgagne bodepd: it looks that all the "extra" bits have been moved/implemented there
19:25 mgagne bodepd: puppet-openstack 4.0.0 on forge is his stuff
19:26 mgagne bodepd: ours got replaced
19:26 bodepd ah. they should/will be moved to extras.
19:26 bodepd or there will be some picking and choosing. Questions, such as, who will do the work, are definitely worth asking
19:26 mgagne bodepd: I'm confused. what's puppet-openstack? who maintains it? why should we maintain an other set of "extras" if they are already available in puppet-openstack?
19:27 bodepd mgagne: b/c it is a dead project that contains tons of code that needs to be deprecated
19:27 mgagne bodepd: I think you misunderstand my concerns
19:27 bodepd mgagne: I'm afraid to start ripping stuff out of it
19:27 mgagne bodepd: I know that puppet-openstack (ours) is dead
19:28 bodepd yes!
19:29 mgagne bodepd: let me rephrase that: what's puppetlabs/puppet-openstack? Why has it replace puppet-openstack on Forge? Who will maintain it? Why should we maintain another set of "extras" (in stackforge/puppet-openstack_extras) if they are already available in puppetlabs/puppet-openstack and puppetlabs/puppet-openstack could be the one pushed by PuppetLabs marketing?
19:29 bodepd that is a composition layer that I won't be using
19:30 bodepd it is driven by Chris/Puppet Labs and focuses on being easy to get started with
19:30 ryanycoleman joined #puppet-openstack
19:30 bodepd AFAIK, it is not intended to have the same types of functionality that will be in extras
19:30 bodepd although, it's not a statement I can make for sure
19:30 mgagne bodepd: right, it's not all about you. I could as well say that I won't use openstack_extras.
19:30 mgagne bodepd: the bits you suggested putting in extras are already available in puppetlabs/puppet-openstack
19:31 bodepd mgagne: give an example
19:31 mgagne bodepd: https://github.com/puppetlabs/puppetlabs-openst​ack/blob/master/manifests/resources/firewall.pp https://github.com/puppetlabs/puppetlabs-open​stack/blob/master/manifests/resources/repo.pp https://github.com/puppetlabs/puppetlabs-openst​ack/blob/master/manifests/resources/database.pp
19:31 guimaluf joined #puppet-openstack
19:31 mgagne bodepd: https://github.com/puppetlabs/puppetlabs-opens​tack/blob/master/manifests/profile/rabbitmq.pp
19:31 mgagne bodepd: https://github.com/puppetlabs/puppetlabs-open​stack/blob/master/manifests/profile/mysql.pp etc
19:32 bodepd in generaly, the reason that I would choose not to use things from there is b/c it embeds hiera lookups inside of the manifest code
19:32 bodepd I can see that there is repo related code, but I can also tell that it is missing parameters that I need
19:33 mgagne bodepd: why always talking in "I" ?
19:33 bodepd or more specifically, it calls classes from within classes, which makes it impossible to override
19:33 mgagne bodepd: it only looks like a selfish conversation
19:33 Waao Hi, can anyone help me ? I'm a newbie to ceph but i rly want to learn about it. Especially your puppet ceph module which was introcued to me by a teacher of mine. Where can i find a clear explanation for dumbies ? x)
19:34 Waao (sry for my average english haha :D)
19:34 bodepd mgagne: I'm actually a bit confused. I looked at this repo, and IIRC, it was totally different yesterday
19:34 mgagne Waao: As far as I know, dachary is the main maintainer of puppet-ceph. I unfortunately don't use it nor do I have any experience with ceph.
19:36 bodepd repos is a bad example. I would not use that, but I could use: https://github.com/puppetlabs/puppetlabs-openst​ack/blob/master/manifests/resources/repo/uca.pp
19:36 bodepd which is the same as code that I though I was moving to extras
19:36 bodepd and as for the question of why I would move it to extras, I thought that was a decision that had already been made
19:36 openstackgerrit Donald Talton proposed a change to stackforge/puppet_openstack_builder: Temporary fix to correct ml2 bug.  https://review.openstack.org/97368
19:37 bodepd b/c puppet-openstack was killed and logic that was contained there needed to be moved somwhere
19:37 mgagne bodepd: What I'm saying is there are duplicated efforts. We have people (including Chris) here suggesting creating an extras repo to group common code and we have Chris pushing his own composition layer to the forge with the same extra bits he suggests putting in openstack_extras.
19:37 mgagne bodepd: how is that not confusing?
19:37 zastern left #puppet-openstack
19:37 Waao mgagne, do you know anybody able to help me ? I guess "Dachary" is my teacher !
19:37 mgagne bodepd: plz stop talking about puppet-openstack, I know it's dead.
19:38 bodepd bodepd: I had never looked at his code. I've been working on moving things in what I though was the decided direction.
19:38 dachary Waao: hi
19:38 Waao dachary, Hi !
19:38 dachary am I ?
19:38 morazi joined #puppet-openstack
19:38 mgagne Waao: unfortunately, I don't contribute to puppet-ceph and don't know whoelse contribute to the project. dachary is the only name I can think of.
19:38 Waao mgagne, thanks a lot for yr help :)
19:39 Waao dachary, i guess you are ! I received a mail from you today..
19:39 bodepd mgagne: yes, the question of why are things that we agreed will go into extras appearing in puppet-openstack
19:39 mgagne bodepd: then there are 2 directions happening at the same time.
19:39 dachary Waao: oh, nice of you to drop by :-)
19:39 mgagne bodepd: leaded by the same guy
19:40 Waao dachary, was curious ! But i'm pretty lost now haha
19:40 ryanycoleman joined #puppet-openstack
19:40 bodepd mgagne: I'm not sure what to say. I'm working on getting automated deployments testing and working on some fixes that had been discussed by the builder crew
19:40 dgollub joined #puppet-openstack
19:40 bodepd mgagne: and I pretty quickly ran into an issue of needing that repo code somewhere for deployments
19:41 bodepd mgagne: so I starting working on extras/lib b/c I though that was the agreed upon direction
19:42 bodepd mgagne: I hear your concerns, and I not aware of them, but I'm not sure what is actionable.
19:42 bodepd s/I was not aware of them/
19:42 bodepd I'm still just in the stage of getting those repos up and accepted, so it'll be next week before any code goes in
19:43 mgagne bodepd: ok, I guess you can use your own repo for now?
19:43 bodepd mgagne: I was actually just working on a patch for librarian-puppet-simple so that you can respecify an override repos
19:43 mgagne bodepd: override?
19:44 mgagne bodepd: you mean having layers above an existing Puppetfile?
19:44 bodepd yeah. a core one.
19:44 bodepd then the ability for anyone to place a new one in a known location and have it override
19:44 bodepd b/c people are using their own modules sets and revisions during development
19:45 mgagne bodepd: btw, do you know any tool to partially mirror rubygems.org?
19:45 bodepd I'm also working on a Vagrant file to can be overridden by another vagrantfile
19:45 bodepd no, seems like something google would know :)
19:45 bodepd and soon enough a hiera data directry that can be overridden
19:45 mgagne bodepd: I looked for it already, there is a bunch of silly solutions available, one being to clone the 5745TB of rubygems available
19:46 bodepd what's wrong with a proxy?
19:46 dachary Waao: you made it that far which is good :-)
19:46 mgagne bodepd: could be a solution
19:46 bodepd dachary: as an FYI, I'm working on a pretty major rewrite of builder
19:47 dachary Waao: did you reply to my mail ? I can provide you with pointers to start contributing right now ;-)
19:47 bodepd dachary: we haven't sorted when we are going to be pushing non-backwards-compat code
19:47 mgagne bodepd: but I would like to be able to control when rubygems are updated and not wait for someone to download them before caching them
19:47 dachary bodepd: that's good news (or is it not ?) !
19:48 dachary bodepd: as a user i'm quite confortable with non compatible changes at this point ;-)
19:48 bodepd https://github.com/rubygems/rubygems-mirror
19:48 Waao dachary, not yet, i saw the bug list you send us and decided to first understand what's that puppetceph module before reply anything !
19:48 bodepd mgagne: that is probably not the link you are looking for
19:48 mgagne bodepd: I tried that one already =)
19:48 mgagne bodepd: I don't have that much space on my server to host all of rubygems
19:48 bodepd dachary: I'll help guide you through updating when I pull the trigger, just a warning/testing the waters
19:49 bodepd mgagne: yeah, I was being a littel snarky with that link
19:49 dachary Waao: you need a lot of context to figure out what puppet-ceph is. In a nutshell, it's an installer for a software defined storage. I'm not sure how much of that actually makes sense to you ;-)
19:49 bodepd mgagne: did you look at stickler?
19:50 dachary bodepd: cool, thanks :0)
19:50 daneyon joined #puppet-openstack
19:50 mgagne bodepd: will take a look
19:50 djm68 joined #puppet-openstack
19:50 bodepd mgagne: that kind of looks like it might be what you are looking for
19:51 Waao dachary, it's a good sum up to begin with, i assure you haha. I understood ceph is a way to store and access data. Now it's okay to me to say this module is kind of 'setup' for ceph ?
19:51 djm68 joined #puppet-openstack
19:52 dachary Waao: in windows parlance it is, indeed.
19:52 bodepd mgagne: to answer your question from before. Ideally, we would not combine things that are generally useful with a composition layer that is only used by folks with certain requirements
19:52 mgagne bodepd: I always fear ending up with a kludge
19:52 bodepd mgagne: so therefor, although there is initial duplication, the intention is to resolve that
19:52 bodepd mgagne: given that there is consensus, the intention must be to resolve the duplication
19:54 mgagne bodepd: proxy solution: https://gist.github.com/1218276/851c​1b43f219f23f5847747fb68b74a2970268bd
19:54 Waao dachary, haha i agree ! It keeps things simplier at first
19:56 dachary Waao: have you used git before ?
19:57 dachary I must say your english is surprisingly good for a CS student. No offense intended to french citizens.
19:59 Waao dachary, yeah ! I've yet cloned the module repos haha. But i forgot my ruby install' was broken.. so i got to fix that first..
19:59 badiane_ka joined #puppet-openstack
20:00 dachary Waao: cool, ping me if you need anything.
20:00 Waao dachary, thanks ! Haha we know..frenchs are not good at all, and our accent is horrible
20:04 hogepodge joined #puppet-openstack
20:05 dachary mais pas du tout !
20:07 Waao C'est un avis personnel mais si quand même ! Le Français est plat et notre accent anglais du moins en premier lieu est moche :p
20:11 hogepodge This is about the extent of my French. https://www.youtube.com/watch?v=X5hrUGFhsXo
20:11 hogepodge I’ll be watching that video on the flight to Paris the autumn.
20:12 Waao LOL
20:12 Waao Great, such usefull sentences and phrases !
20:21 dachary everybody needs to learn some french for the next summit I'm afraid :-P
20:27 aldavud joined #puppet-openstack
20:32 tfz joined #puppet-openstack
20:51 Waao pink dachary
20:51 Waao ping dachary
20:51 Waao xD
20:51 dachary ponk ?
20:51 Waao haha fails.
20:51 Waao Question !
20:52 Waao The bug list u gave us are mainly typo bugs
20:52 dachary very perceptive of you.
20:52 Waao is that normal or..other question what is this lesson meaning ? How to contribut to an open source software ?
20:53 dachary for most people it's correct. Now if you feel up to a more challenging contribution, we can figure something out ;-)
20:55 Waao So it's really how to contribute, like propose a push, waiting for replies and acceptation ?
20:56 dachary in a nutshell, yes.
20:56 Waao Okay, thought it was a total technical lesson ! I'm surprised
20:57 dachary most people are. There is a lot of "soft skills" involved in Free Software.
20:57 dachary s/is/are/
20:58 Waao That's okay, it's interesting to figure out how communication and team effort is merged in Free Software
20:59 Waao and btw, we dont have enough time at school to go deeper on a totaly new software
20:59 dachary Waao: I'm curious how long it would take you to go thru the review process, technically speaking. Starting from https://wiki.openstack.org/wiki/Puppet​-openstack#Contributing_to_the_modules
21:00 dachary my be is that you can sort it out within less than one hour but I may be mistaken
21:00 dachary s/my be/my bet/
21:00 dachary s/sort it out/figure it out/
21:01 Waao with you answering our questions may be..
21:02 aldavud hmmm, if you are on a windows machine it might be tricky to get the whole ssh key thing sorted though ;)
21:02 ryanycoleman joined #puppet-openstack
21:03 ryanycoleman joined #puppet-openstack
21:03 Waao oh I give up to install properly things on windows when ruby said "ERROR loading zlib" Argh. I'm downloading cygwin. =D
21:07 Waao So if i fix typo bug (Bug #1326251) i can propose my fix to contributors and if its upvoted it could be merged to the current trunk ?
21:20 derekh_ joined #puppet-openstack
21:21 djm68 joined #puppet-openstack
21:22 hogepodge Waao We had a similar discussion at the OpenStack Summit.
21:23 hogepodge There’s a tension on how to encourage contributions from everyone, including new users with narrow patches, while maintaining standards we developed as a community.
21:25 hogepodge I encourage you to take the extra steps. It feels steep at the beginning, but is so rewarding in the longer run. If it’s too steep of a hill, we can have a more experienced contributor assume ownership of the patch.
21:28 Waao which extra steps. I figured out 2 step : 1 - Fixing bugs on my local version 2 - propose a patch to other. They can test, comment and upvote patch. Only members of core_contributors can decide to merge my path to the trunk. Is that it ? :O
21:30 hogepodge Waao I guess by extra steps I mean following code formatting guidelines, testing, going through the iterative review process.
21:30 djm68 joined #puppet-openstack
21:32 Waao Oh ! okay. It feels like too much for a typo bug, but it's seems awesome for technicals issues :p
21:35 guimaluf joined #puppet-openstack
21:36 hogepodge A typo bug would receive less scrutiny because it doesn’t need it. If a review languishes, especially if it’s a simple one, dropping it into this channel will usually result in an immediate review from core.
21:36 dvorak speaking of which, I need reviews on this: https://review.openstack.org/#/c/94120/ :)
21:45 hogepodge dvorak. thanks
21:46 aldavud dvorak: awesome! that will finally let me do flat networks :)
21:47 dachary hogepodge: I agree on principle. I have a distinct feeling that Waao  is going the extra mile already. He is learning a number of new conventions and tools. It's difficult to judge without knowing the person's background though.
21:48 dachary and I don't know Waao, it's just a gut feeling ;-)
21:49 Waao Seeing you speak about me but not understanding the crucial phrase that make all sense.. Argh !
21:50 hogepodge dachary agree Waao thanks for getting involved. I want to make sure that everyone who wants to contribute has the opportunity and is supported in that.
21:50 dachary Waao: :-)
21:51 hogepodge irc is a bit deficient in that regard.
21:51 hogepodge Waao it’s all favorable speaking about you.
21:52 Waao i dnt doubt about it ! :p
21:52 dachary Waao: yes. Don't worry. We're constantly trying to find better ways to welcome new contributors while maintaining the quality standards of the projects.
21:54 Waao I go to sleep ! Thanks all for your support. I think i'll read all the contributing documentation before searching to contribute really even if it's typobug or wtver. It seems important ! :D
21:54 dachary Waao: good night !
21:55 Waao Thanks, bye :)
21:56 djm68 joined #puppet-openstack
21:56 dachary hogepodge: in two weeks from now two dozens of CS students will visit puppet-openstack and my hope is that they will learn how to contribute. For most of them it will be their first time. Do you have any advice on how this should be handled ?
22:00 hogepodge Will they know basic git workflow?
22:03 hogepodge I would give an example in class, then maybe find easy patches for them to practice on.
22:05 dachary a number of them probably don't know git
22:05 ryanycoleman joined #puppet-openstack
22:06 dachary there will be technical training during class indeed
22:06 dmsimard dachary: two dozen, yeesh
22:06 dmsimard that's a lot :)
22:07 dachary dmsimard: it would be extraordinary if all of them turn into long term contributors. But even if just one does, it's great ;-)
22:08 mkoderer joined #puppet-openstack
22:09 openstackgerrit Clayton O'Neill proposed a change to stackforge/puppet-neutron: Added '' as a valid value for gateway_ip  https://review.openstack.org/97957
22:12 jamespd joined #puppet-openstack
22:16 djm68_ joined #puppet-openstack
22:20 mjblack joined #puppet-openstack
22:24 mmaglana joined #puppet-openstack
22:29 aldavud dachary: dmsimard: what is your opinion on zapping disks when removing an osd (i.e. 'test -b #{data} && ceph-disk zap #{data}')
22:31 dmsimard aldavud: I'm biased because I'm not comfortable with managing the cluster with puppet beyond the initial installation
22:31 aldavud but then just never specity ensure => absent ;)
22:31 dmsimard I trust puppet but not too much.
22:32 aldavud I don't plan to ever use it, but I can't have integration tests without the ability to clean up
22:33 ryanycoleman joined #puppet-openstack
22:34 aldavud i'm stuck on scenario_node_terminus osd removal. either i have a fully running system, or all the ceph binaries are gone and I would need to manually clean the partition table
22:36 aldavud which is anoying as GPT can't be cleared by just dd'ing the beginning of the disk
22:39 djm68 joined #puppet-openstack
22:43 daneyon joined #puppet-openstack
22:50 dachary aldavud: I'm ok with that. And I'm quite sure removing OSD is not in the workflow of any system administrator at the moment so it won't scare anyone ;-)
22:51 ryanycoleman joined #puppet-openstack
22:55 exed_ joined #puppet-openstack
22:59 prad joined #puppet-openstack
23:01 aldavud dachary: I was already looking into porting ceph-disk zap to ruby :)
23:02 aldavud https://gist.github.com/al​davud/e5e2f47234135fe193b3
23:09 daneyon joined #puppet-openstack
23:23 djm68 joined #puppet-openstack
23:25 djm68_ joined #puppet-openstack
23:35 dvorak it looks like we're now using rspec 3.0.0 on stable/icehouse also?
23:48 michchap joined #puppet-openstack
23:52 bodepd michchap: good morning! I've been slightly tab stalking you :)
23:53 michchap bodepd: morning :)
23:53 bodepd I've been busy. I've got code I want you to look at
23:53 michchap sure
23:54 bodepd michchap: https://github.com/bodepd/lib​rarian-puppet-simple/pull/28
23:54 dvorak funny, me too
23:54 dvorak I can wait though :)
23:54 bodepd mine are piles of code to look at :)
23:54 bodepd michchap: that first one is pretty basic obvious. The next one starts to add overrides to the builder
23:55 bodepd https://github.com/bodepd/pup​pet_openstack_builder/pull/1
23:55 dvorak so is the reason that we're getting rspec 3 because the Gemfile's don't require something lower?
23:55 bodepd michchap: I've also got a ton of questions for you. They are pretty builder specific, so I might hit you elsewhere
23:55 bodepd dvorak: I'm pretty sure its coming from spec_helper
23:56 bodepd dvorak: lemme check
23:56 dvorak bodepd: doesn't appear to be
23:56 dvorak it's happening on stable/icehouse also
23:56 dvorak and I don't think it makes sense to upgrade those tests to rspec 3
23:56 michchap the first one looks good. I can test
23:56 bodepd I'm guessing that is the issue: https://github.com/puppetlabs/puppetla​bs_spec_helper/blob/master/Gemfile#L6
23:56 michchap the second I might need to have some breakfast
23:57 bodepd michchap: yeah. basically, the magic is hte special directory and an evil eval
23:57 michchap I'll ping you shortly
23:57 bodepd you may want to ask in #puppet-dev or #puppet, I saw that it got mentioned last night
23:57 bodepd dvorak: more likelly this: https://github.com/puppetlabs/puppetlab​s_spec_helper/blob/master/Rakefile#L32
23:58 bodepd dvorak: does that make sense? we defer to spec_helper to pull in test deps so that we get the right versions of what puppetlabs uses
23:58 bodepd but their dep range wasn't specific enough. I'm surpirsed it hasn't been updated upstream
23:59 dvorak actually, that makes even less since, I think that should say it needs 2.10.x

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