Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-10-02

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

All times shown according to UTC.

Time Nick Message
00:16 ari joined #puppet-openstack
00:39 mjblack joined #puppet-openstack
01:05 ryanycoleman joined #puppet-openstack
01:46 bodepd mjblack: I've been doing a bit too much of that as well ;)
01:54 mjblack bodepd: its really too good, its becoming an addiction for me :/
02:14 otherwiseguy joined #puppet-openstack
02:20 ari joined #puppet-openstack
02:51 mjblack_ joined #puppet-openstack
03:05 mjblack joined #puppet-openstack
05:25 tvb|afk joined #puppet-openstack
05:54 tvb|afk joined #puppet-openstack
06:16 tvb|afk_ joined #puppet-openstack
06:31 tvb joined #puppet-openstack
07:31 mmagr joined #puppet-openstack
08:16 derekh joined #puppet-openstack
08:47 qba73 joined #puppet-openstack
09:26 michchap_ joined #puppet-openstack
09:27 tvb joined #puppet-openstack
09:41 britthouser joined #puppet-openstack
10:28 tvb joined #puppet-openstack
10:28 tvb joined #puppet-openstack
11:23 tvb joined #puppet-openstack
12:04 britthouser joined #puppet-openstack
12:09 dprince joined #puppet-openstack
12:09 dmsimard joined #puppet-openstack
12:11 morazi joined #puppet-openstack
13:26 mjblack joined #puppet-openstack
13:31 rcrit bodepd, any chance you can take a look at https://review.openstack.org/#/c/43139/ ? Just needs one more ack to move along
13:34 michchap joined #puppet-openstack
13:46 otherwiseguy joined #puppet-openstack
14:21 mjblack joined #puppet-openstack
14:25 mjblack joined #puppet-openstack
14:26 mjblack joined #puppet-openstack
14:39 mjblack mgagne: when you did havana testing did you do it with 12.04 or 12.10?
14:40 mgagne 12.04
14:40 mgagne mjblack: I don't have much time this week to test havana
14:41 mjblack mgagne: thats ok, just was seeing if someone else had it working on precise, atleast package installation perspective
14:41 ryanycoleman joined #puppet-openstack
15:15 ari_ joined #puppet-openstack
15:23 tvb joined #puppet-openstack
15:23 tvb joined #puppet-openstack
15:25 badiane_ka joined #puppet-openstack
15:41 otherwiseguy joined #puppet-openstack
15:56 prad joined #puppet-openstack
15:58 ryanycoleman joined #puppet-openstack
16:03 hogepodge joined #puppet-openstack
16:09 prad left #puppet-openstack
16:12 hogepodge joined #puppet-openstack
16:39 marun joined #puppet-openstack
16:41 mjblack joined #puppet-openstack
16:41 tvb joined #puppet-openstack
16:41 mjblack joined #puppet-openstack
17:02 hogepodge joined #puppet-openstack
17:30 tvb joined #puppet-openstack
17:33 bodepd rcrit: sorry for the delay
17:35 hogepodge ping EmilienM
17:36 rcrit bodepd, no worries, thanks for looking
17:37 hogepodge or anyone with experience with the ceilometer module
17:38 hogepodge ^ mgagne,dprince
17:39 dprince hogepodge: I'm using it, if that counts as experience :)
17:40 hogepodge :-)
17:40 hogepodge There are a bunch of patches against master that haven't been cherry-picked back to grizzly. I'm wonder if that's because of incompatibility or just oversight.
17:41 dprince hogepodge: probably oversight. The Ceilo modules are quite new... and I doubt they'd work without the most recent patches.
17:42 dprince hogepodge: Ideally we'd have tests setup (which use stable branch packages) where we can test these for assurance.
17:42 hogepodge yeah. I have a test structure set up internally, but it doesn't touch ceilometer. I could add it in, though, and start verifying.
17:44 bodepd hogepodge: I wouldn't worry about it until havana
17:44 bodepd hogepodge: I think that's when it becomes core
17:45 hogepodge bodepd: Sounds reasonable to me. I was going to move it to the bottom of the stack anyway.
17:46 hogepodge bodepd: My test env for internal PL is set up, and I'm just starting to diff the master and stable/grizzly logs to see what needs to be cherry picked back.
17:46 openstackgerrit A change was merged to stackforge/puppet-openstack: Fix typo with vncproxy_host for ::nova::compute  https://review.openstack.org/48512
17:47 hogepodge bodepd: it looks like a lot of changes made it back, but a few slipped through the cracks
17:50 hogepodge bodepd: Once stable/grizzly is released I'm going to turn to stable/havana. I'm shooting for as close to Oct 17 as I can get.
17:51 openstackgerrit A change was merged to stackforge/puppet-openstack: Expose ovs_enable_tunneling parameter for neutron  https://review.openstack.org/48310
17:51 openstackgerrit A change was merged to stackforge/puppet-openstack: Configurable admin for Swift proxy authtoken class  https://review.openstack.org/47834
17:51 openstackgerrit A change was merged to stackforge/puppet-openstack: Allow passing keystone_host to controller class  https://review.openstack.org/48715
17:52 openstackgerrit A change was merged to stackforge/puppet-openstack: Add charset option for db setting  https://review.openstack.org/48420
18:01 ari__ joined #puppet-openstack
18:05 ryanycoleman joined #puppet-openstack
18:19 openstackgerrit Rob Crittenden proposed a change to stackforge/puppet-openstack: Options to configure MySQL for SSL and glance, keystone and nova clients.  https://review.openstack.org/43139
18:39 bodepd hogepodge: every change doesn't have to make it back
18:39 bodepd hogepodge: I would say it's up to the individual committer to determine if they want their changes to work against grizzly
18:40 bodepd hogepodge: what's wrong with releasing what it's the repos now as stable grizzly?
18:40 hogepodge bodepd: there's nothing wrong with that.
18:40 hogepodge bodepd: assuming bug fixes made it back.
18:41 hogepodge bodepd: I can start by checking the logs to see what the actual differences are, but it they're nominal and stable/grizzly passes testing, then I see no problem with releasing as is.
18:41 mgagne hogepodge: what kind of experience
18:42 mgagne hogepodge: there was backward incompatibility introduced with how keystone auth is configured on the agents
18:42 bodepd mgagne: in grizzly?
18:42 bodepd mgagne: what commits?
18:42 mgagne havana
18:42 mgagne puppet-ceilometer
18:42 bodepd ah. gotcha
18:43 bodepd there are probably going to be some other small backwards compat breakages on havana as well
18:43 ryanycoleman joined #puppet-openstack
18:43 mgagne and it bothers me a little as I have modules/manifests relying on those parameters which got moved somewhere else
18:43 mgagne bodepd: https://review.openstack.org/#/c/45083/
18:44 bodepd I haven't been following ceilometer too much
18:47 openstackgerrit A change was merged to stackforge/puppet-openstack: Options to configure MySQL for SSL and glance, keystone and nova clients.  https://review.openstack.org/43139
18:49 mgagne bodepd: those kind of changes bothers me as people relying solely on hiera to provide configs won't be "informed" of the change, no deprecation warnings. I know that I +2 the change but looking back, I should have voiced this concern.
18:51 hogepodge bodepd: thanks for your feedback dan
18:52 hogepodge vswitch addition is still under review over at openstack-infra
18:53 hogepodge hoping it gets in soon
18:59 mgagne hogepodge: why don't we ask redhat to publish the module on forge and use it?
19:02 hogepodge mgagne: Short term that should be fine. Longer term it feels like stackforge is the place for development. There was some discussion of this on the mailing list (although it was wrt to PL vs stackforge), and the consensus seemed to be that stackforge should be the place for the main fork.
19:02 hogepodge mgagne: I set the packstack/puppet-vswitch as the upstream seed for stackforge
19:03 hogepodge mgagne: I reached out to RH, and they're on board with stackforge
19:04 hogepodge PL published a stackforge module from their fork at the last release.
19:10 ari__ joined #puppet-openstack
19:11 bodepd mgagne: I've kind of been working on a solution for that
19:11 bodepd mgagne: I've been using catalog diffs to validate changes during development
19:11 bodepd mgagne: which would uncover those kind of changes
19:12 mgagne bodepd: we cannot expect people to diff their catalog when upgrading :-/
19:12 bodepd mgagne: I'm just spit-balling here :)
19:12 mgagne bodepd: =)
19:12 bodepd mgagne: I'm also working on a system to pre-compile hiera-data in advance
19:12 bodepd mgagne: (it's on my TODO list anyways)
19:13 bodepd mgagne: this system could find keys that match existing classes, but that dont match class parameters
19:13 bodepd seeing the direction that hiera-2 is going in, it seems like these are the class of problem they want to evantually solve
19:14 bodepd mgagne: to answer your question. I could expect you to do catalog diffs when you rebase against upstream
19:14 bodepd mgagne: I was thinking about a jenkins plugin to publish catalogs when puppet runs
19:14 bodepd mgagne: then you could do a diff when you rebase against upstream
19:14 bodepd s/do/be able to do/ :)
19:15 bodepd anyways, I'm full of crazy ideas, and my consulting schedule is about to be wide open
19:15 * bodepd whistles
19:15 mgagne bodepd: it's not about rebasing, it's about updating to the latest version of the module when backward incompatibility is introduced. I'm mentionning an other reason why we should be careful when deprecating/moving parameters around
19:16 bodepd mgagne: I was thining of you issue as a more fundamental tooling issue introduces by hiera
19:16 bodepd mgagne: and was thinking of solutions from that perspective
19:16 mgagne bodepd: ok =)
19:16 mgagne meanwhile, we should be careful =)
19:17 bodepd mostly blathering b/c I have been thinking about this stuff b/c I've been waist-deep in hiera lately
19:17 bodepd mgagne: (how are these idioms working out for you?)
19:17 bodepd mgagne: I know it can be culturely insensitive , but I love my idioms
19:17 mgagne bodepd: we all love our own idioms =)
19:18 mgagne bodepd: working "ok", still trying to figure out if I should "trust" stackforge modules to be "stable" (parameters) or if I should instead use profiles to pass around parameters to the modules.
19:18 bodepd interesting...
19:19 bodepd I'm (as you know) going full on data
19:19 bodepd this is definitely a problem I will have to solve
19:20 mgagne bodepd: so far, we have done great regarding parameters stability but if anything changes, I don't want to play catch up
19:21 mgagne bodepd: I'm partially relying on my custom puppet-openstack module to protect me against parameter changes
19:26 bodepd mgagne: yeah.
19:26 bodepd it also makes me a hypocrite if I allow backwards compat changes and then complain about upstream :)
19:26 EmilienM hogepodge: pong (late sory)
19:27 mgagne bodepd: the problem I want to fix is avoiding the update of 3 repositories to introduce a new configuration. AND avoid having 3 different ways to add a configuration: injecting hiera values via manifests (hiera lookup) vs role vs profile vs stackforge module.
19:28 mgagne bodepd: yes, I have the same concerns
19:28 bodepd mgagne: have you had a chance to see what I've been working on?
19:28 bodepd mgagne: the docs still aren't create, but I've started to port Ciscos references architectures to it
19:28 bodepd mgagne: it's all data, so, it would not detect parameter changes
19:29 mgagne bodepd: "it would not detect parameter changes" this bothers me a lot =)
19:30 bodepd mgagne: yeah. I can fix that (but I still need to think about the performance implications)
19:30 bodepd mgagne: right now, all of the data is driven by a hiera lookup, where the only caching is the hashes from file to memory
19:30 bodepd mgagne: I've starting writing things that do a hiera pre-compile
19:30 bodepd (where the hiera lookups will just be hash key lookups)
19:30 bodepd and everything is compiled as a part of the ENC
19:31 bodepd once that works, there is no reason, I can't say:
19:31 bodepd - for all things in the class list (provided they are provided by the ENC)
19:31 bodepd - for all keys we compiled that match those classes
19:31 bodepd - verify that those params actually exist
19:32 bodepd mgagne: does that make sense?
19:32 mgagne bodepd: hold on =)
19:32 bodepd this is all weeks away before I could start hacking on something like that
19:48 ryanycoleman joined #puppet-openstack
20:05 ryanycoleman joined #puppet-openstack
20:21 otherwiseguy joined #puppet-openstack
20:24 ari__ joined #puppet-openstack
20:52 ryanycoleman joined #puppet-openstack
20:54 britthou_ joined #puppet-openstack
21:17 prad joined #puppet-openstack
21:41 ryanycoleman joined #puppet-openstack
21:55 mgagne bodepd: ping
22:01 ryanycoleman joined #puppet-openstack
22:45 ryanycoleman joined #puppet-openstack
23:07 dmsimard joined #puppet-openstack
23:25 ryanycoleman joined #puppet-openstack
23:57 bodepd mgagne: pong

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