Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-09-15

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

All times shown according to UTC.

Time Nick Message
00:00 zeus joined #puppet-openstack
00:03 cjellick_ joined #puppet-openstack
00:03 dvr077 joined #puppet-openstack
00:06 dalgaaf___ joined #puppet-openstack
00:08 mgagne joined #puppet-openstack
00:13 EmilienM_ joined #puppet-openstack
00:14 sileht joined #puppet-openstack
00:21 holser joined #puppet-openstack
00:22 JayJ___ joined #puppet-openstack
00:49 JayJ___ joined #puppet-openstack
00:55 sputnik13 joined #puppet-openstack
01:07 JayJ____ joined #puppet-openstack
01:33 sputnik13 joined #puppet-openstack
01:51 JayJ____ joined #puppet-openstack
01:55 xingchao joined #puppet-openstack
02:21 openstackgerrit Xingchao Yu proposed a change to stackforge/puppet-designate: Add designate_api_paste_ini type  https://review.openstack.org/121358
02:22 openstackgerrit Xingchao Yu proposed a change to stackforge/puppet-designate: Introduce designate:config to manage custom options  https://review.openstack.org/121359
02:26 openstackgerrit Xingchao Yu proposed a change to stackforge/puppet-glance: Move os_region_name param into glance::api  https://review.openstack.org/121283
02:35 openstackgerrit Xingchao Yu proposed a change to stackforge/puppet-designate: Introduce designate:config to manage custom options  https://review.openstack.org/121359
02:36 openstackgerrit Colleen Murphy proposed a change to stackforge/puppet-openstacklib: (WIP) Implement base aviator provider  https://review.openstack.org/116753
02:47 openstackgerrit A change was merged to stackforge/puppet-nova: Add ability to set floating IP Pool names  https://review.openstack.org/117047
03:44 imcsk8 joined #puppet-openstack
04:26 gildub joined #puppet-openstack
04:54 sanjayu joined #puppet-openstack
04:58 sputnik13 joined #puppet-openstack
05:15 walkleyn joined #puppet-openstack
05:55 ddmitriev joined #puppet-openstack
06:15 TorLaedre joined #puppet-openstack
06:16 mmagr joined #puppet-openstack
06:22 ddmitriev joined #puppet-openstack
06:43 francois joined #puppet-openstack
06:48 gchamoul joined #puppet-openstack
06:49 gcha joined #puppet-openstack
07:01 tfz joined #puppet-openstack
07:20 arnaud_orange joined #puppet-openstack
07:48 qba73 joined #puppet-openstack
08:05 mmagr joined #puppet-openstack
08:06 gchamoul joined #puppet-openstack
08:08 derekh joined #puppet-openstack
08:11 mmagr EmilienM, Hi, do you got a minute?
08:18 sileht joined #puppet-openstack
08:49 alokj joined #puppet-openstack
08:56 gcha joined #puppet-openstack
09:33 sileht joined #puppet-openstack
09:41 berendt joined #puppet-openstack
09:44 walkleyn__ joined #puppet-openstack
10:18 gildub joined #puppet-openstack
10:34 Krast joined #puppet-openstack
11:06 JayJ____ joined #puppet-openstack
11:25 openstackgerrit David Moreau Simard proposed a change to stackforge/puppet-nova: virsh returns a list of secret uuids, not keyring names  https://review.openstack.org/119043
11:25 walkleyn__ joined #puppet-openstack
11:34 EmilienM good morning
11:34 EmilienM mmagr: yep
11:56 mmagr EmilienM, Hi Emilien, I just wanted to ask if I'm missing something in that patch https://review.openstack.org/#/c/113164/ ... I don't get why you guys want to move that conf to other class
11:57 EmilienM mmagr: will look
11:59 EmilienM mmagr: I think these options can be used without rbd backend
11:59 EmilienM so they have to move in libvirt class
12:01 mmagr EmilienM, ok ... I suppose those options should be as parameters of nova::compute::libvirt class right?
12:01 EmilienM yes
12:05 mmagr ok, will modify that patch
12:05 sputnik13 joined #puppet-openstack
12:12 sputnik13 joined #puppet-openstack
12:48 JayJ____ joined #puppet-openstack
12:58 rharrison joined #puppet-openstack
13:01 zeus joined #puppet-openstack
13:04 morazi joined #puppet-openstack
13:32 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-nova: Support workers number of EC2 service  https://review.openstack.org/121555
13:49 francois joined #puppet-openstack
13:52 ddmitriev joined #puppet-openstack
13:53 xingchao_ joined #puppet-openstack
14:00 francois joined #puppet-openstack
14:11 xingchao_ joined #puppet-openstack
14:17 mdorman joined #puppet-openstack
14:21 rwsu joined #puppet-openstack
14:26 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-cinder: Use $::processorcount to compute API workers number  https://review.openstack.org/121585
14:28 rmoe joined #puppet-openstack
14:31 prad joined #puppet-openstack
14:36 xingchao_ joined #puppet-openstack
14:37 rmoe joined #puppet-openstack
14:51 crinkle morning
14:57 spredzy_ hey
14:58 michchap_ morning
14:58 EmilienM o/
15:01 Hunner o/
15:03 crinkle I started making an agenda but didn't get terribly far: https://etherpad.openstack.org/p/puppet-openstack-agenda-9.15.2014
15:04 crinkle regarding migrating the modules to the new namespace in the puppet forge, I wanted to confirm that the plan for that is to do it on the Juno release?
15:05 michchap_ seems reasonable to me
15:05 Hunner And the namespace is "stackforge"?
15:05 crinkle yes
15:06 crinkle and it is already created
15:06 crinkle i believe
15:06 spredzy_ crinkle, yep https://forge.puppetlabs.com/stackforge
15:06 Hunner Will they continue to use the same version sequencing?
15:06 crinkle i think they should
15:07 xingchao_ joined #puppet-openstack
15:07 Hunner It makes sense; just double-checking
15:07 crinkle also, should the puppetlabs-openstack module be moved there as well?
15:07 Hunner Not unless it is going to be THE stackforge openstack module
15:08 Hunner Since it is actually produced by puppetlabs
15:08 crinkle I agree
15:08 michchap_ it's not on stackforge so it doesn't make sense at this point.
15:09 crinkle okay
15:09 crinkle any other questions or comments on that topic?
15:10 michchap_ are you going to take care of the releases crinkle?
15:10 EmilienM crinkle: from OpenStack side, I don't know if we should use github/openstack namespace
15:10 crinkle michchap_: yes I'll take care of that
15:10 michchap_ crinkle: great.
15:11 crinkle EmilienM: could you clarify?
15:11 crinkle we're using github/stackforge now
15:11 EmilienM yes and I was thinking about moving them to github/openstack
15:12 Hunner As "official" projects?
15:12 EmilienM *one day*
15:12 EmilienM Hunner: yes
15:12 crinkle I don't really agree with making them official openstack projects - puppet is one of many options for managing openstack
15:13 crinkle i think it fits well under stackforge
15:13 michchap_ I don't have any problem with it, but I don't feel like it meets the quality requirement until we have public CI. We also tend to struggle to keep up with trunk. If those were fixed then I have no issue.
15:13 EmilienM ok, I think we can talk about this topic at the Summit and when we will have functional testing QA
15:13 EmilienM michchap_: true
15:14 crinkle that brings us to the other agenda item, EmilienM do you have an update on implementing CI?
15:18 crinkle we can table that for a moment
15:19 crinkle someone mentioned backwards compatibility policy on the agenda?
15:19 michchap_ yeah that was me
15:19 EmilienM crinkle: no
15:19 EmilienM I don't have patches
15:19 EmilienM crinkle: I started some work on "github/stackforge/dox" to support Puppet
15:19 EmilienM https://review.openstack.org/#/c/119943/
15:20 EmilienM but really WIP
15:20 michchap_ I have been thinking about whether maintaining backwards compatibility is actually helpful. Firstly, following semver we are doing major releases every six months which by definition are changed interfaces. Secondly and more importantly, we frequently need to add parameters that are realistically needed by a deployment.
15:21 michchap_ An example of this is nova-neutron notifications. Without this set up properly, no notifications are sent and VMs won't get past spawning. We didn't break back compat, but the user needs to update their data layer (hiera or enc)
15:21 michchap_ so given that these types of situations are going to occur because of upstream changes, we have to expect users to review their data layer on every major release
15:22 michchap_ and given that, I don't see much benefit in maintaining backwards compatibility as strictly as we do
15:22 michchap_ I don't really know how to manage the changes any better than a changelog and release notes, though.
15:24 michchap_ What will definitely help is a common data layer driving CI, so we can track changes that have to go in there.
15:24 mdorman i would argue that most people do a review, anyway, when going to a new major release.  just because that’s kind of the expectation from OS
15:24 mdorman so don’t think this would be a major problem.
15:25 michchap_ mdorman: That's exactly my thoughts - there is no prod cloud that wouldn't test out new module versions anyway. There isn't a use case of deploy havana, upgrade the modules, set ensure => latest and apply puppet to upgrade to icehouse.
15:25 mdorman yup.
15:25 michchap_ I will bring this up at the summit because I recall the response being very different last time.
15:27 michchap_ anyway, that's all I have on the topic. Perhaps it would be better suited to the mailing list or a summit discussion.
15:28 crinkle I think either or both would be appropriate
15:29 JayJ____ joined #puppet-openstack
15:29 crinkle I agree that using semver should imply that we don't need to maintain backwards compatibility on major releases
15:29 crinkle anything else to discuss? any pet reviews that need attention?
15:30 michchap_ the hash based repo thing is ready
15:31 michchap_ I'm not using stahnma/epel because it doesn't cover my use case, which means it likely doens't cover other people's use case.
15:31 crinkle michchap_: what is your use case that isn't covered?
15:31 michchap_ crinkle: I want to set particular parameters to the yumrepo resources.
15:32 michchap_ crinkle: It's not pratical to create a parameter for every single one.
15:32 michchap_ crinkle: So to use stahnma's epel is very easy - just do manage_epel = false, and then include the class yourself.
15:33 cjellick joined #puppet-openstack
15:33 michchap_ crinkle: but specifying all the repositories for a site via a hash from hiera is extremely useful where the user needs control over parameters like the proxy, or any of the obscure ones.
15:33 Hunner michchap_: Are you saying the actual epel module doesn't support all parameters, or the OS modules' use of it doesn't?
15:34 michchap_ Hunner: I'm saying that the epel module/class doesn't wrap every parameter for the yumrepo resource type.
15:34 michchap_ Hunner: which is totally fine in the majority of cases, but the reason I made this patch was to get around that limitation.
15:39 crinkle michchap_: I seem to be missing something, you're using baseurl, descr, gpgkey, and failovermethod which I think are all covered in the module
15:39 michchap_ crinkle: but it's passed to create_resources - I could specify any parameter for the yumrepo type in that hash and it would work.
15:40 michchap_ crinkle: like, these are all the options: https://docs.puppetlabs.com/references/latest/type.html#yumrepo
15:40 crinkle but you're using $epel_hash to manage epel, which the user doesn't seem to have control over
15:41 michchap_ crinkle: You can trivially specify it yourself in $repo_hash
15:41 michchap_ eg: https://github.com/michaeltchapman/puppet_openstack_builder/blob/master/data/hiera_data/user.stacktira.yaml#L157
15:42 crinkle if you specify it via $repo_hash then you would set $manage_epel to false anyways
15:42 alokj joined #puppet-openstack
15:43 sputnik13 joined #puppet-openstack
15:44 michchap_ crinkle: I can also add other parameters via $repo_defaults, for example to add a proxy to every yumrepo resource being created.
15:45 crinkle oh okay
15:46 crinkle okay I can see the value in doing it this way
15:46 daneyon joined #puppet-openstack
15:47 michchap_ It's worth noting that spredzy_ and I differ pretty drastically in our preferred way to manage repos, but I think the way I've done it accommodates both his preference, which is to make a module for each repo, and my preference, which is a giant hash.
15:49 michchap_ I realised I had comments on that review explaining this at least in part, but they were sitting in draft =/
15:51 crinkle anyone else want to review https://review.openstack.org/#/c/113922 or have other items to discuss?
15:53 JayJ____ joined #puppet-openstack
15:54 crinkle notes are on https://etherpad.openstack.org/p/puppet-openstack-agenda-9.15.2014
15:55 crinkle have a good morning everyone
15:55 * crinkle -> bagels
15:55 michchap_ Thanks all.
15:55 markvoelker1 joined #puppet-openstack
15:58 EmilienM thanks
15:59 xingchao_ joined #puppet-openstack
16:12 JayJ____ joined #puppet-openstack
16:13 rmoe joined #puppet-openstack
16:14 daneyon joined #puppet-openstack
16:15 tfz joined #puppet-openstack
16:26 JayJ____ joined #puppet-openstack
16:40 alokj_ joined #puppet-openstack
16:42 sanjayu joined #puppet-openstack
16:52 thumpba joined #puppet-openstack
16:54 JayJ____ joined #puppet-openstack
16:59 morazi joined #puppet-openstack
17:18 JayJ____ joined #puppet-openstack
17:19 openstackgerrit Colleen Murphy proposed a change to stackforge/puppet-openstacklib: (WIP) Implement base aviator provider  https://review.openstack.org/116753
18:33 tfz joined #puppet-openstack
18:33 openstackgerrit joined #puppet-openstack
18:36 EmilienM I would like a review on michchap_ patch (old) : https://review.openstack.org/#/c/111729/
18:36 EmilienM any core around may have a look please ^
18:57 thumpba joined #puppet-openstack
19:08 openstackgerrit A change was merged to stackforge/puppet-keystone: Fix usage of admin_endpoint  https://review.openstack.org/111729
19:21 mmaglana joined #puppet-openstack
19:24 openstackgerrit Richard Megginson proposed a change to stackforge/puppet-keystone: setup keystone using apache mod_wsgi  https://review.openstack.org/109676
19:28 thumpba joined #puppet-openstack
20:03 aimon joined #puppet-openstack
20:04 sputnik13 joined #puppet-openstack
20:07 sputnik13 joined #puppet-openstack
20:11 pradk joined #puppet-openstack
20:11 sputnik13 joined #puppet-openstack
20:16 thumpba joined #puppet-openstack
20:56 sputnik13 joined #puppet-openstack
22:22 dachary joined #puppet-openstack
22:57 JayJ____ joined #puppet-openstack
23:07 dtalton joined #puppet-openstack
23:46 imcsk8_ joined #puppet-openstack

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