Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-01-17

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

All times shown according to UTC.

Time Nick Message
23:03 mgagne michchap_: it can if it's declared in the right order.
23:03 mgagne michchap_: afaik
23:03 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-horizon: Release 3.0.0  https://review.openstack.org/67317
23:04 mgagne michchap_: either way, I don't like this check. =)
23:04 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-keystone: Release 3.0.0  https://review.openstack.org/67318
23:05 hogepodge horizon and keystone tested. Keystone has a bug that's also in master. It's not a show stopper, and the patch will just clarify a failure rather than prevent it.
23:05 EmilienM michchap_: let me some time, I'm on something atm
23:05 michchap_ mgagne: I wouldn't mind something that ensures that neutron is in the catalog, but I think the type of user that will make the mistake of not declaring the neutron class can't really be saved by checks like this.
23:05 hogepodge If I can get horizon and kestone quickly reviewed, that would be great. Just module file changes.
23:06 hogepodge Swift, heat, and ceilometer still need to be tested.
23:06 mgagne michchap_: our modules aren't structured that way anyway...
23:07 mgagne michchap_: not sure why this check was introduced in the first place
23:08 michchap_ mgagne: EmilienM likes to have as much validation as possible. I'm inclined to agree, but in this case it's probably not feasibly
23:08 mgagne michchap_: well, I don't feel we should go that far
23:09 michchap_ hogepodge: honestly I don't have a problem with you merging your own patches for the forge releases.
23:09 mgagne hogepodge: checking
23:09 hogepodge Good to hear. I don't either, but problems have been caught in the past.
23:09 EmilienM mgagne: michchap_ i do agree with you guys.
23:10 michchap_ hogepodge: if you're just bumping the version on things, like these ones.
23:10 EmilienM please stop me when i want to go too far in testing
23:10 hogepodge Yup. That's all it is. I just needed an rc to test against.
23:10 EmilienM my job is to install openstack, so i like when i'm sure that i have no config error :-)
23:10 michchap_ EmilienM: cool, I will approve that patch then.
23:10 mgagne EmilienM: raw puppet modules might not be the right tool for such validation tbh
23:11 michchap_ EmilienM: mgagne: We have introspection tools that can do some of it
23:11 michchap_ but they aren't very mature
23:31 mgagne hogepodge: will rc3 be tagged as well?
23:31 mgagne hogepodge: if any =)
23:37 hogepodge They should be tagged.
23:37 hogepodge I think I got them all, but could have missed one.
23:38 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-nova: Fix configuring libvirt before installation  https://review.openstack.org/67330
23:51 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-openstack: Remove duplicated parameters for keystone  https://review.openstack.org/67335
23:52 mgagne EmilienM: Would it apply to havana? https://bugs.launchpad.net/puppet-neutron/+bug/1263952
23:55 mgagne hogepodge: what will/should be the next version? 3.0.1? 3.1.0?
23:55 mgagne hogepodge: depends?
23:59 unixpro1970 joined #puppet-openstack
00:04 hogepodge Anything that adds or breaks an API for Havana would be 3.1.0
00:04 hogepodge Anything that fixes a bug would be 3.0.1
00:04 hogepodge Icehouse will change entirely to be 2014.1.0, barring any strong opposition.
00:05 hogepodge We'll abandon semver proper, but OpenStack doesn't use it, and we don't release enough to really need it.
00:05 hogepodge Matching the OS release numbering makes more sense.
00:05 mgagne hogepodge: sure
00:05 mgagne hogepodge: doing a lot of cleanup in bugs =)
00:06 thumpba joined #puppet-openstack
00:06 mgagne hogepodge: marking stuff as released
00:06 hogepodge I need to get started on the puppet-openstack replacement sometime soon. Been trying to figure out the best procedure to do that.
00:21 xarses hogepodge: something sleek =)
00:21 xarses fuel uses a heavily modified version of puppet-openstack
00:22 mgagne xarses: orly? =)
00:39 hogepodge By the end of Icehouse I want to have pacemaker support in it. And sleek.
00:39 hogepodge If fuel open source?
00:40 mgagne hogepodge: afaik, it's hosted on stackforge now
00:40 hogepodge Guess it is.
00:40 hogepodge Interesting. It's quite the fork
00:41 mgagne https://github.com/stackforge/fuel-library/tree/master/deployment/puppet
00:43 mgagne hogepodge: looks a lot cleaner than the last time I checked =)
00:43 michchap joined #puppet-openstack
00:44 mgagne hogepodge: like a year ago
00:44 hogepodge All of the modules have improved in the last year.
00:44 mgagne hogepodge: sure
00:44 mgagne hogepodge: but nothing compared to the chaos I saw =)
00:48 hogepodge :-)
00:50 hogepodge good night.
00:51 mgagne thanks, you too
01:15 badiane_ka joined #puppet-openstack
01:24 bodepd joined #puppet-openstack
01:34 openstackgerrit A change was merged to stackforge/puppet-horizon: Release 3.0.0  https://review.openstack.org/67317
01:35 openstackgerrit A change was merged to stackforge/puppet-keystone: Release 3.0.0  https://review.openstack.org/67318
01:37 openstackgerrit A change was merged to stackforge/puppet-neutron: Remove check for correct core_plugin when using ml2.  https://review.openstack.org/63068
01:46 xingchao joined #puppet-openstack
02:10 openstackgerrit Michael Chapman proposed a change to stackforge/puppet_openstack_builder: Syntax errors in heat mappings  https://review.openstack.org/67357
02:14 tnoor joined #puppet-openstack
02:47 ilbot3 joined #puppet-openstack
02:47 Topic for #puppet-openstack is now Place to collaborate on Puppet/OpenStack tools: logs at http://irclog.perlgeek.de/puppet-openstack/today
05:15 EmilienM mgagne: no, https://bugs.launchpad.net/puppet-neutron/+bug/1263952 don't apply to havana.
05:21 openstackgerrit A change was merged to stackforge/puppet_openstack_builder: Syntax errors in heat mappings  https://review.openstack.org/67357
05:23 pzimmer left #puppet-openstack
05:32 tnoor joined #puppet-openstack
05:57 bauzas joined #puppet-openstack
06:46 openstackgerrit A change was merged to stackforge/puppet-nova: Fix configuring libvirt before installation  https://review.openstack.org/67330
08:08 bauzas joined #puppet-openstack
08:48 mmagr joined #puppet-openstack
08:52 piliman974 joined #puppet-openstack
08:53 piliman974 joined #puppet-openstack
09:07 bauzas joined #puppet-openstack
11:05 derekh joined #puppet-openstack
12:33 EmilienM good morning
13:37 morazi joined #puppet-openstack
13:55 dprince joined #puppet-openstack
14:00 fvollero EmilienM: Bonjour :)
14:06 EmilienM fvollero: hey
14:06 morazi joined #puppet-openstack
14:10 fvollero EmilienM: how things going? :)
14:11 prad joined #puppet-openstack
14:11 EmilienM fvollero: I'm doing the tempest/neutron sprint in Montreal, quite busy!
14:12 fvollero EmilienM: and 'cool' i would say :)
14:13 fvollero EmilienM: I am going to update the puppet-ceilometer adding the new agent service :)
14:13 EmilienM fvollero: which one?
14:13 fvollero EmilienM: ceilometer-agent-notification
14:13 fvollero EmilienM: Or it's already there?
14:14 EmilienM fvollero: don't think so, i was not aware about this agent
14:14 fvollero EmilienM: I discovered it recently but not had time to do it before
14:27 EmilienM fvollero: i just had a talk with sileht who is ceilo-dev
14:28 EmilienM fvollero: so yeah, we have now this new agent which use rpc only
14:28 EmilienM fvollero: and collector also have now rpc/udp options.
14:28 EmilienM fvollero: if you produce patches, please me and sileht  as reviewers :-)
14:31 dvanders joined #puppet-openstack
14:34 fvollero EmilienM: i will add obviously also eglynn
14:34 fvollero EmilienM: :)
14:57 marun joined #puppet-openstack
14:57 dmsimard joined #puppet-openstack
14:58 finnx joined #puppet-openstack
15:08 dmsimard fc__: Are you okay with the e-mail I sent you ?
15:10 fc__ dmsimard: I didn't have time to look at it more, if it can make it work why not, the only problem I'm seeing is that it will be complicated to have swift+ceilometer installed with other ceilometer components …
15:11 dmsimard fc__: Works for me. I wouldn't personally install swift-proxy on something like a nova compute node but I see what you mean
15:13 fc__ dmsimard: but I would install swift-proxy on the same node as ceilometer-collector or ceilometer-api
15:14 fc__ dmsimard: in fact, I'm not sure moving all logs to another place, changing directories ownerships, etc.… would be cleaner than applying a 3 lines patch
15:15 dmsimard fc__: I'll try and make something as backwards compatible and as configurable as possible
15:30 bcrochet joined #puppet-openstack
15:33 xarses joined #puppet-openstack
15:36 marun joined #puppet-openstack
15:43 bcrochet joined #puppet-openstack
16:05 marun joined #puppet-openstack
16:32 marun joined #puppet-openstack
17:07 sputnik13 joined #puppet-openstack
17:13 otherwiseguy joined #puppet-openstack
17:30 tnoor joined #puppet-openstack
17:33 angdraug joined #puppet-openstack
17:46 pzimmer joined #puppet-openstack
17:47 thumpba joined #puppet-openstack
17:49 thumpba joined #puppet-openstack
18:00 xarses joined #puppet-openstack
18:02 dmsimard Any reason why you guys thinks internal and admin urls are hardcoded as http ? https://github.com/stackforge/puppet-swift/blob/master/manifests/keystone/auth.pp#L80-L82
18:05 dmsimard I would probably implement internal_protocol as well as admin_protocol to be in line with the public protocol that can be configured.
18:09 xarses dmsimard: I thought we where slowly changing these to single var URI, but yes, that is fugly, 92-94 too
18:13 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-ceilometer: Release candidate 3.0.0-rc3  https://review.openstack.org/67542
18:14 dmsimard xarses: Got to work with that is there for the time being :)
18:14 dmsimard xarses: Single var URI, you mean with the protocol and the port in it ?
18:14 dmsimard xarses: like public_address => 'http://proxy:80' for instance
18:15 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-heat: Release candidate 3.0.0-rc3  https://review.openstack.org/67543
18:17 hogepodge joined #puppet-openstack
18:22 marun joined #puppet-openstack
18:23 hogepodge The gate is down, so you might want to hold off on sending gerrit reviews until it's back up. Otherwise you're going to have to recheck no error.
18:24 hogepodge Infra is working on it.
18:26 xarses dmsimard: protocol, host, port; yes
18:30 hogepodge joined #puppet-openstack
18:31 hogepodge ping fc__
18:35 marun joined #puppet-openstack
18:43 dmsimard1 joined #puppet-openstack
19:05 marun joined #puppet-openstack
19:29 marun joined #puppet-openstack
19:37 finnx left #puppet-openstack
19:37 bauzas joined #puppet-openstack
19:39 hogepodge joined #puppet-openstack
19:41 marun joined #puppet-openstack
19:52 sputnik13 joined #puppet-openstack
20:00 marun joined #puppet-openstack
20:03 sputnik13 joined #puppet-openstack
20:04 pzimmer joined #puppet-openstack
20:15 marun joined #puppet-openstack
20:18 dmsimard joined #puppet-openstack
20:19 sputnik13 joined #puppet-openstack
20:26 prad_ joined #puppet-openstack
20:29 francois1 joined #puppet-openstack
20:29 dmsimard xarses: Going back in time here but
20:29 dmsimard xarses: I do believe that it's best to use one variable for the entire URL
20:29 xarses woot time travel
20:30 dmsimard xarses: Except it breaks backwards compatibility and is not in line with other modules (e.g, nova::keystone::auth: https://github.com/stackforge/puppet-nova/blob/master/manifests/keystone/auth.pp)
20:31 dmsimard The ::keystone::auth classes seem largely similar across openstack modules now that I look at it in fact
20:32 dmsimard Maybe bodepd or hogepodge can chime in
20:32 xarses dmsimard:  all i remember is that we talked about it back in nov/dec
20:32 hogepodge We came to a conclusion.
20:32 hogepodge Can't remember what it was.
20:32 dmsimard xarses: In the meantime though, I need to be able to have a http admin/internal URL so i'll introduce a change that is backwards compatible :)
20:32 xarses and that i thought the majority was to start moving to url
20:32 xarses erm uri
20:33 xarses i think it started with some changes to horizon or keystone its self
20:34 xarses I dono, I just wan't to keep up with what our standard is supposed to be
20:34 xarses maybe we should start a hacking guide
20:34 dmsimard What's the policy when introducing a backwards incompatible change ? Needs to be depreciated first, etc etc ?
20:35 xarses dmsimard: exactly why we need a hacking guide !
20:35 dmsimard xarses: ehhhh
20:38 openstackgerrit David Moreau Simard proposed a change to stackforge/puppet-swift: Allow configuration of admin and internal protocols for keystone endpoint  https://review.openstack.org/67571
20:38 bauzas joined #puppet-openstack
20:38 thumpba_ joined #puppet-openstack
20:40 EmilienM dmsimard: nice one^
20:40 dmsimard EmilienM: :D
20:41 dmsimard EmilienM: I'm attacking swift::proxy::ceilometer now.
20:41 marun joined #puppet-openstack
20:42 EmilienM dmsimard: take care
20:43 EmilienM dmsimard: it's a warfield!
20:43 dmsimard EmilienM: Well, my attack stopped violently - I'm not going to be able to fix it - puppet-swift and puppet-ceilometer would be fighting
20:44 dmsimard ::ceilometer enforces permissions on /etc/ceilometer, if I try to set anything in swift::proxy::ceilometer, ::ceilometer will just reset the permissions back to what they were :(
20:44 EmilienM i do not like permissions stuffs in puppet. IMHO, that's packaging area.
20:45 dmsimard https://github.com/stackforge/puppet-ceilometer/blob/master/manifests/init.pp#L96-L107
20:45 dmsimard Apparently puppet takes care of creating the user and group
20:46 dmsimard Hence why puppet handles the permissions
20:47 EmilienM dmsimard: again, it's my opinion which should be wrong due to my short experience in puppet.
20:50 dmsimard EmilienM: No, no, you're right, the packaging should be doing things better
20:50 dmsimard EmilienM: The puppet modules do a lot of things that should be done by packaging :)
20:58 EmilienM dmsimard: pretty sure that https://review.openstack.org/#/c/67571/ would deserve a backport
20:59 dmsimard EmilienM: Yes, yes, it would. Should I create them right now ? I was waiting until that one went through
20:59 EmilienM dmsimard: it could wait
21:00 dmsimard EmilienM: It's just so I don't have to modify the code at three places if changes are required ;)
21:01 marun joined #puppet-openstack
21:14 marun joined #puppet-openstack
21:26 sputnik13 joined #puppet-openstack
21:27 marun joined #puppet-openstack
21:28 openstackgerrit A change was merged to stackforge/puppet-ceilometer: Release candidate 3.0.0-rc3  https://review.openstack.org/67542
21:29 openstackgerrit A change was merged to stackforge/puppet-heat: Release candidate 3.0.0-rc3  https://review.openstack.org/67543
21:36 dmsimard Ugh, I'll have to take another look monday. This is giving me headaches.
21:38 marun joined #puppet-openstack
21:38 dmsimard EmilienM: If the change gets merged (protocols) i'll backport them to havana and grizzly when I get the e-mail notification
21:38 dmsimard EmilienM: (I see you're asking for reviewers so it looks like you're kind of anxious about this)
21:39 EmilienM dmsimard: i'm anxious about getting a review quickly. Not about the patch though. no worries
21:39 EmilienM dmsimard: i usually add one or 2 core, to merge it asap
21:39 EmilienM dmsimard: if i was anxious, i would not +2 a patch :-)
21:39 tnoor joined #puppet-openstack
21:54 derekh joined #puppet-openstack
21:55 marun joined #puppet-openstack
21:57 sputnik13 havana dashboard (horizon) does not show password panel under user settings...  anyone familiar with this?
21:58 xarses joined #puppet-openstack
21:59 thumpba_ joined #puppet-openstack
22:00 EmilienM i have a dummy puppet question
22:00 EmilienM hogepodge: i'm sure you could help :)
22:00 hogepodge :-)
22:01 EmilienM let's say, i have 3 class: controller, network and compute. I want to ensure network and compute have neutron::agents::ovs, but also want to allow a user to declare network & compute on same node. What is the best way ? use ensure_resource with neutron::agents::ovs class?
22:01 thumpba_ joined #puppet-openstack
22:05 EmilienM hogepodge: ^
22:06 hogepodge Heh. It's a problem I was facing. I didn't solve it.
22:06 EmilienM hogepodge: so you you did not solve it, I'm lost \o/
22:07 hogepodge No, I've called in reinforcements
22:08 EmilienM hogepodge: does the ensure_resource could solve it?
22:08 Hunner That would solve it... usually
22:08 EmilienM Hunner: hey!
22:08 Hunner Unless you have different parameters, etc
22:08 EmilienM Hunner: no, always the sames
22:09 EmilienM Hunner: my concern is also the puppet version, i try to ensure i can run it on > 2.7
22:09 Hunner >= 2.7?
22:09 Hunner Does ensure_resource not work on 2.7? (sorry, it's been a while)
22:10 hogepodge I might do it by disabling the agent in one service if you want to colocate.
22:10 hogepodge Pass a flag to your wrapper class. But Hunner is far more experienced in these things than I am.
22:10 Hunner I mean, you could do something janky like have a middle-class declare neutron::agents::ovs with $calling_class::<whatever parameter name> to access the supposed values, and both network & compute can include the middle class
22:11 hogepodge so if enabled { agent }
22:13 thumpba_ joined #puppet-openstack
22:13 hogepodge my language is a bit vague. Basically add a flag for colocation so you can enforce only declaring the class once.
22:13 EmilienM Hunner: hogepodge: I'll try, thank you guys
22:13 Hunner So there is the middle-class option, `if defined(Class['network']) or defined(Class['compute'])` option (eww)...
22:13 Hunner Or the ensure_resource() option
22:14 EmilienM Hunner: i'll try both
22:24 thumpba_ joined #puppet-openstack
22:29 marun joined #puppet-openstack
22:55 xarses joined #puppet-openstack

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