14:00:58 #startmeeting puppet-openstack 14:00:59 Meeting started Mon Jan 12 14:00:58 2015 UTC and is due to finish in 60 minutes. The chair is crinkle. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:00 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:03 The meeting name has been set to 'puppet_openstack' 14:01:32 #topic Vancouver summit 14:01:36 EmilienM: ? 14:01:39 #link https://etherpad.openstack.org/p/puppet-vancouver-summit-proposal 14:01:49 I started an etherpad 14:02:30 I would like to submit a talk with someone 14:02:39 about a deployment use-case with our Puppet modules 14:02:47 show our efforts 14:03:00 morning 14:03:03 the status, what you can do, how to contribute, etc 14:04:08 crinkle: though it's maybe a bit too early to talk about design dessions 14:04:13 sessions* 14:04:36 EmilienM: thats a good idea for a talk 14:05:01 it's a mix between ops/dev talk 14:05:35 mfisch: might be a good person to give that talk with :) 14:05:54 indeed, why not! 14:06:26 I need to discuss with my team but its possible yes 14:06:29 just cannot commit yet 14:06:29 i’d be happy to be a part of that too 14:06:43 #action EmilienM to start an abstract on the etherpad 14:06:57 hello all 14:07:40 I'm happy to help too I just don't have a real deployment to talk about 14:07:43 mdorman: ok. We will figure that out. 14:07:49 kk 14:08:17 mdorman: mfisch crinkle : I just created the etherpad to be sure we don't overlap much about Puppet submissions 14:08:29 and try to submit them efficiently, and together 14:08:35 sounds good. 14:08:46 cool. We'll follow up on IRC later 14:08:52 crinkle: done. 14:08:58 mmk 14:09:09 #topic Ops midcycle 14:09:34 I'm not sure if I will be there, but I know there will be a puppet session 14:09:49 mfisch has volunteered to moderate it 14:09:57 it would be good to come up with topics beforehand I think 14:10:00 EmilienM: will you be coming? 14:10:31 mfisch: where is it? 14:10:40 and when? sorry I missed the infos 14:10:40 philadelphia 14:10:42 Philadelphia, US 14:10:53 east coast-ish 14:10:53 #link https://etherpad.openstack.org/p/PHL-ops-meetup 14:10:57 thanks 14:11:20 hum. I don't think I'll go, I'm not OPS but I'll forward it to my OPS team. 14:12:07 but +1 to let mfisch moderating it :-) 14:12:32 But yes we do need an agenda 14:12:57 should we try to start one now or send to the mailing list? 14:13:01 crinkle, if you don't ago is anyone from PL going ? 14:13:14 spredzy: possibly one of our ops people 14:13:44 WIth an ops focus some discussion on things that are problems for ops is high on my agenda 14:14:15 crinkle, who ever that person is, probably add him/her in the loop for the agenda might be a good idea 14:14:25 sure 14:15:41 #action Ask mailing list for help preparing ops midcycle agenda 14:16:14 hi,all 14:16:22 hello 14:16:59 #topic openstackclient patches 14:17:14 We merged all the major work but it ended up wreaking a bit of havoc 14:17:35 Three patches still to do bugfixes, then all those patches should be backported: 14:17:45 yeah 14:17:47 #link https://review.openstack.org/145861 14:17:57 #link https://review.openstack.org/145986 14:18:07 #link https://review.openstack.org/145987 14:18:17 xingchao: could you review/merge them please? 14:18:27 since you're the only one core around today 14:19:05 hi all ... very late 14:19:15 or fc__ ... 14:19:22 hello 14:19:28 sbadia, ^^ 14:19:40 EmilienM: Ok, I will review them. sorry for absent recent meeting :( 14:19:47 crinkle: once those are merged and stable can you update the progres on the ML? 14:19:53 hi, late also :s 14:19:57 seems like some people woke up after my last e-mail :-) 14:19:57 mfisch: sure 14:20:02 we'd like to move our trees forward but were waiting for less chaos 14:20:07 +1 14:20:37 great. Enough core people to have a look at the backports. Please review them, they are blockers 14:20:59 #action crinkle will update the mailing list when openstackclient patches are merged and stable 14:21:59 #info Next steps are updating the patches to allow keystone v3 (richm) and updating the rest of the modules to use openstackclient (crinkle) 14:22:10 wouhouuuuu 14:23:04 Anything else to discuss on that topic? 14:23:25 nop 14:23:46 mmk 14:23:53 #topic rabbitmq 14:24:26 puppet labs just released a 5.0.0 of the rabbitmq module that contains a security fix 14:24:38 I would really like to get the modules to be using that 14:24:43 crinkle: any pointer? 14:24:57 EmilienM: pointer? 14:25:03 crinkle: URL 14:25:14 crinkle: what does it change? 14:25:23 ah I don't think there is a cve yet for the security fix 14:25:29 if it's the way to setup RabbitMQ server, I would argue to drop this code 14:25:39 there are some api changes that make it incompatible with nova::rabbitmq and cinder::rabbitmq 14:25:47 and just manage RabbitMQ resources (users/permissions/vhosts) 14:26:01 #link https://github.com/puppetlabs/puppetlabs-rabbitmq/blob/master/CHANGELOG.md 14:26:05 crinkle: server side? 14:26:16 thx for URL 14:26:25 I don't know the best way to deprecate those clases 14:26:47 classes 14:27:01 crinkle: 5.0.0 changes are for server or resources? 14:27:16 mainly for the server setup 14:27:25 so it wouldn't really affect the rabbitmq_user and such 14:27:27 ok so my position here is to drop this code 14:27:35 because it's not up to our modules to setup rabbitmq server 14:27:51 we already had this discussion I think 14:28:12 we did but I'm wondering if we can just drop it or if we have to go through a deprecation cycle first 14:28:14 Agree with EmilienM about not up to the puppet-openstack modules to set up rabbit 14:28:32 crinkle, that would be one way to go 14:28:36 deprecation first 14:28:36 communicate on the ML 14:28:40 and drop the code in next cycle 14:29:03 iirc, bogdando was working on it. 14:29:18 on the other hand given that the release contains a security fix maybe we could sidestep the deprecation ocycle 14:29:36 crinkle, defined sidestep please 14:29:36 hello 14:29:43 define* 14:29:46 bogdando: is still working on it, I don't think all his latest patches are in this release though 14:30:06 spredzy: just dropping it as opposed to notifying and then dropping in six months? 14:30:13 spredzy: esquiver! 14:30:17 crinkle, EmilienM I submitted a patch which allow rabbitmq not being installed from nova anymore 14:30:27 good. 14:30:40 bogdando: link? 14:30:44 crinkle, EmilienM thx 14:31:19 #link: https://review.openstack.org/144745 14:32:11 so I think we'd rather drop that class entirely instead of modifying it 14:33:13 crinkle: I would drop the class and use https://github.com/stackforge/puppet-openstacklib/blob/master/manifests/messaging/rabbitmq.pp 14:33:47 crinkle: I mean, keep nova::rabbitmq but using openstacklib 14:34:31 okay, my only concern is whether we can do that on master now or if we need to add a notify{} to it and then fix it in six months 14:34:49 crinkle: I would warning 14:35:02 okay 14:35:44 #action crinkle will submit patches to puppet-nova and puppet-cinder to add deprecation warning for setting up the rabbitmq server 14:36:13 crinkle and bogdando : just make sure you're sync on this one 14:36:29 crinkle: also, change the way to create resources to use openstacklib. 14:36:36 yes 14:37:41 crinkle, EmilienM so, looks like the new versions of rabbitmq module install should go to this messaging/rabbitmq.pp as well? 14:38:17 yes openstacklib should be updated to use the new module version 14:38:20 bogdando: there is no change AFIK since it only affect server configuration 14:38:25 crinkle: isn't? 14:38:25 If we want remove rabbitmq_class we should first deprecate it and keep with False option, IMHO 14:38:51 crinkle, bogdando isn't the prupose by the end of next sprint being to not manage rabbitmq via puppet-modules anymore ? 14:38:58 bogdando: makes sense 14:39:07 so class rabbitmq shouldn't be moved to message/rabbitmq.pp 14:39:30 sprint -> cycle 14:39:32 sorry 14:39:35 openstacklib/messaging/rabbitmq.pp is only for users/permissions/vhosts resources 14:39:40 spredzy, yes, it should. I suggest 1st to deprecate this param and provide a new value for it (false) 14:40:05 +1 14:40:10 ack 14:41:21 #action also modify the rabbitmq class to not require setting up the server (bogdando - https://review.openstack.org/144745) 14:41:49 anything else on rabbitmq? 14:42:21 nop 14:42:43 #topic puppet-group members 14:42:46 EmilienM: 14:42:54 #link https://groups.google.com/a/puppetlabs.com/forum/#!topic/puppet-openstack/t8hYgzceOdM 14:43:12 so I would like to drop people not involved from the list: https://review.openstack.org/#/admin/groups/134,members 14:43:47 I like hogepodge words 14:44:03 people not involved lately, may come back later but for now, we need to update the list 14:44:09 +1 14:44:16 +1 14:44:33 especially for people who did not even reply to the email 14:44:51 so the first step is to drop some people, and the next step would be to promote some others that contributed well in the last cycle 14:45:00 mfisch: indeed :D 14:45:41 let me quote chris words "I’d like to think of being a core developer as more of a responsibility than as a reward. If you’re not able to contribute to reviews, coding, planning meetings, and other community efforts, there should be no shame in stepping down and no barriers for reentry when you can commit yourself again." 14:46:13 sbadia: crinkle xingchao fc__ : since you are the 4 core here, can you vote? 14:46:41 +1 14:47:02 +1 too 14:47:28 :) 14:47:33 ok, sounds like everyone (core and non-core) agree with it. 14:47:41 +1 other openstack team also follow this rule 14:48:21 #action EmilienM to update the list and communicate on the ML about the update 14:48:23 crinkle: done. 14:48:52 #topic open discussion 14:48:58 any other topics to bring up? 14:49:17 dprince: hey 14:49:26 EmilienM: hello 14:49:33 I think Dan wanted to give an update about Heat/Puppet integration 14:50:13 sure, the update is that things are (slowly) progressing. Patches have started to land in the TripleO Heat Templates. 14:50:32 we are tracking progress on this etherpad: https://etherpad.openstack.org/p/puppet-integration-in-heat-tripleo 14:50:41 #link https://etherpad.openstack.org/p/puppet-integration-in-heat-tripleo 14:51:19 there are good notes there about how this work, etc. But today we are very close to having a working devtest_overcloud.sh which can use Puppet. 14:51:31 dprince: this is awesome 14:51:32 devtest_overcloud.sh is a script we use in TripleO to dev/test things 14:51:52 * dprince figures this probably doesn't mean much here... it is a TripleO thing 14:52:06 it's okay 14:52:17 One thing I would mention is I'm really liking the ability to deploy to baremetal w/ Heat and Puppet 14:52:21 it's another interesting use-case here 14:52:44 Heat gives you the ability to use OpenStack APIs to create the basic deployment resources, via Ironic, Neutron, etc. 14:52:51 And PUppet is great for the configuration side. 14:54:11 I've already got some good feedback from nibalizer, and EmilienM on the reviews. But more thoughts would be appreciated. 14:54:27 awesome 14:54:33 As an example this is what a puppet storage node ended up looking like: https://review.openstack.org/#/c/146129/2/puppet/overcloud_storage.pp,cm 14:54:35 dprince: xingchao is very familiar with puppet-swift AFIK 14:54:45 fc__ to 14:55:02 I would ask them to have a look at the manifests if they want/can 14:55:11 EmilienM: ++ that would be cool 14:56:07 As much as possible I'm trying to keep the manifests themselves to just 'include' statements, and then do everything else in Hiera 14:56:31 I'm missing some Heat features to be able to do some conversions however... so I need to do a bit of that on the puppet side for now. 14:56:40 +1 14:56:52 Eventually, when some of the Heat features land I think we'll end up in a much better place... 14:56:53 exactly how people should use the modules 14:57:02 good progress, thanks for the update 14:57:10 we're about out of time, any other updates? 14:57:25 dprince: don't hesitate to poke us on IRC if you need reviews 14:57:39 EmilienM: cool, thanks 14:58:42 crinkle: I think we are done :-) 14:58:47 woot 14:58:52 #endmeeting