Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-11-25

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

All times shown according to UTC.

Time Nick Message
00:16 openstackgerrit Colleen Murphy proposed stackforge/puppet-keystone: Use openstackclient for keystone_tenant  https://review.openstack.org/134844
00:31 vinsh joined #puppet-openstack
00:52 arnaud_orange joined #puppet-openstack
01:11 topshare joined #puppet-openstack
01:17 rmoe joined #puppet-openstack
01:30 aimon joined #puppet-openstack
01:45 openstackgerrit Colleen Murphy proposed stackforge/puppet-openstacklib: Swap aviator for openstackclient  https://review.openstack.org/134843
01:45 openstackgerrit Colleen Murphy proposed stackforge/puppet-keystone: Use openstackclient for keystone_tenant  https://review.openstack.org/134844
01:52 openstackgerrit Colleen Murphy proposed stackforge/puppet-ceilometer: Clean up stray copy-paste error  https://review.openstack.org/136942
01:52 xarses joined #puppet-openstack
01:57 gildub joined #puppet-openstack
02:08 xingchao joined #puppet-openstack
02:18 sputnik13 joined #puppet-openstack
02:52 crinkle jenkins finally reported back for https://review.openstack.org/#/c/136852/
02:53 gildub joined #puppet-openstack
03:02 mmaglana joined #puppet-openstack
03:28 JayJ joined #puppet-openstack
03:39 xarses joined #puppet-openstack
04:27 sputnik13 joined #puppet-openstack
04:40 sputnik13 joined #puppet-openstack
04:54 subscope_ joined #puppet-openstack
05:53 gildub joined #puppet-openstack
05:59 fedexo joined #puppet-openstack
06:21 xingchao_ joined #puppet-openstack
06:32 starmer joined #puppet-openstack
06:49 sputnik13 joined #puppet-openstack
07:06 saju_m joined #puppet-openstack
08:03 dgurtner joined #puppet-openstack
08:09 stamak joined #puppet-openstack
08:16 qba73 joined #puppet-openstack
08:27 tfz joined #puppet-openstack
08:30 tfz joined #puppet-openstack
08:36 mmagr joined #puppet-openstack
09:14 derekh joined #puppet-openstack
09:52 xingchao joined #puppet-openstack
10:05 topshare joined #puppet-openstack
10:05 topshare joined #puppet-openstack
10:10 saju_m joined #puppet-openstack
10:13 saju_m joined #puppet-openstack
10:14 saju_m joined #puppet-openstack
10:58 ikke-t joined #puppet-openstack
10:59 openstackgerrit Christian Berendt proposed stackforge/puppet-horizon: Add support for the configuration of OPENSTACK_CINDER_FEATURES  https://review.openstack.org/136762
10:59 gildub joined #puppet-openstack
11:01 ikke-t joined #puppet-openstack
11:11 openstackgerrit Christian Berendt proposed stackforge/puppet-horizon: Add support for the configuration of OPENSTACK_CINDER_FEATURES  https://review.openstack.org/136762
11:26 stamak joined #puppet-openstack
11:29 rcallawa joined #puppet-openstack
12:01 saju_m joined #puppet-openstack
12:21 openstackgerrit Bogdan Dobrelya proposed stackforge/puppet-openstack_extras: Add pacemaker provider for HA services  https://review.openstack.org/136708
12:31 ikkeT joined #puppet-openstack
12:46 EmilienM good morning!
13:04 morazi joined #puppet-openstack
13:08 EmilienM bogdando: oh boy, that's a huge patch !
13:15 bogdando EmilienM, just a  new revision :)
13:15 EmilienM bogdando: I would suggest you WIP it while it's not open for review
13:16 EmilienM and when you think it's ready, drop the -1 and people will review it
13:17 rharrison joined #puppet-openstack
13:17 JayJ joined #puppet-openstack
13:22 JayJ joined #puppet-openstack
13:27 JayJ joined #puppet-openstack
13:28 openstackgerrit Dmitry Ilyin proposed stackforge/puppet-openstack_extras: Add pacemaker provider for HA services  https://review.openstack.org/136708
13:31 rcallawa joined #puppet-openstack
13:32 dgurtner joined #puppet-openstack
13:32 JayJ joined #puppet-openstack
13:38 topshare joined #puppet-openstack
14:08 richm joined #puppet-openstack
14:30 openstackgerrit Bogdan Dobrelya proposed stackforge/puppet-openstack_extras: Add pacemaker provider for HA services  https://review.openstack.org/136708
14:52 openstackgerrit Bogdan Dobrelya proposed stackforge/puppet-openstack_extras: Add pacemaker provider for HA services  https://review.openstack.org/136708
14:52 bogdando EmilienM, looks like it is ready for review now. I added docs, and Dmitry fixed the failing test case
14:53 EmilienM you guys work fast, nice
14:53 EmilienM bogdando: I'll review it asap, thx
14:53 guimaluf crinkle, EmilienM, bodepd could you please tell me how is your puppet developemnt workflow and how you manage your development? do you use vagrant? how do you make a review? do you checkout and then run inside a vm or test in some scenario? I really want to help on puppet-openstack development!
14:54 EmilienM guimaluf: I have servers and use to validate each patch in this env
14:54 guimaluf EmilienM, do you use r10k to manage each environemtn?
14:54 EmilienM guimaluf: yes
14:55 guimaluf my impression is that sometimes things are really manual, and this take time...
14:56 guimaluf I'm really curious about that, cause I don't understand how do you contribute to so many modules and don't get crazy managing the environment
14:58 guimaluf EmilienM, if you do a minor fix and want to test again you git commit/push to your working repo, r10k deploy, and puppet agent on your nodes? all this manual?
15:04 EmilienM crinkle: mgagne: do you think you could have an eye on  https://review.openstack.org/#/c/135840/ ?
15:06 rcallawa_ joined #puppet-openstack
15:15 openstackgerrit Merged stackforge/puppet-ceilometer: Clean up stray copy-paste error  https://review.openstack.org/136942
15:16 openstackgerrit Merged stackforge/puppet-horizon: Update horzion::wsgi::apache for new apache module  https://review.openstack.org/136852
15:17 openstackgerrit Merged stackforge/puppet-openstacklib: Release 5.0.0 - Juno  https://review.openstack.org/136674
15:36 sputnik13 joined #puppet-openstack
15:39 pradk joined #puppet-openstack
15:42 vinsh joined #puppet-openstack
15:47 ericpeterson joined #puppet-openstack
16:04 rcallawa joined #puppet-openstack
16:23 mdorman joined #puppet-openstack
16:40 sputnik13 joined #puppet-openstack
16:49 mdorman joined #puppet-openstack
16:54 arnaud_orange left #puppet-openstack
16:57 crinkle EmilienM: hmm, I would rather avoid using inheritance since the style guide recommends against it https://docs.puppetlabs.com/guides/​style_guide.html#class-inheritance
16:57 crinkle I think I would rather pass the parameters in explicitly and use ensure_resource on the class itself
16:58 crinkle but I think mgagne is against ensure_resource so I would wait for his opinion
16:59 crinkle guimaluf: I use vmware fusion in vagrant with r10k, with the modules in a synced folder
16:59 mgagne crinkle: I still don't understand why we need ensure_resource in this case and even less inheritance. I though I made my point clear last week.
17:00 crinkle mgagne: maybe you could reiterate for me
17:00 mgagne crinkle: https://gist.github.com/mg​agne/6ed00808815b0af0fe0e
17:01 mgagne crinkle: I don't see why you could get duplicated resources if include is used.
17:02 crinkle in my mind include wasn't used, which would necessitate ensure_resource, but I'm fine with your solution
17:02 mgagne crinkle: I'm a bit confused before I thought we had an agreement on the implementation details last week.
17:03 mgagne crinkle: and now we have to revisit it again for no clear reason (well, I found none in the commit message)
17:04 * mgagne goes to lunch
17:04 crinkle mgagne: I was just commenting on EmilienM's review, forgive me for forgetting the details of the outcome of last week's discussion
17:04 mgagne crinkle: sure, I think we both agree that inheritance is not required
17:04 crinkle yes
17:04 openstackgerrit Merged stackforge/puppet-horizon: Release 5.0.0 - Juno  https://review.openstack.org/136667
17:05 mgagne crinkle: even less ensure_resource which I feel is a sign of bad implementation
17:05 mdorman joined #puppet-openstack
17:07 EmilienM crinkle: I'll produce a new patchset
17:09 mmaglana joined #puppet-openstack
17:10 EmilienM mgagne: sorry for my ignorance. I think it's clear for me now.
17:18 xarses joined #puppet-openstack
17:19 mmaglana joined #puppet-openstack
17:21 mdorman_ joined #puppet-openstack
17:32 openstackgerrit Emilien Macchi proposed stackforge/puppet-nova: Configure database parameters on the right nodes  https://review.openstack.org/135840
17:33 openstackgerrit Emilien Macchi proposed stackforge/puppet-nova: Configure database parameters on the right nodes  https://review.openstack.org/135840
17:39 mmaglana joined #puppet-openstack
17:45 aimon joined #puppet-openstack
17:46 rmoe joined #puppet-openstack
17:52 aimon joined #puppet-openstack
17:57 patcable q: is there a reason that the behavior of $rabbit_host and $rabbit_hosts is not consistent between the modules? if $rabbit_hosts is defined, rabbit_host should be absent, but the modules dont seem to handle that well
17:58 patcable i'm working on patching and getting all the stuff in place so i can contribute (sigh) but i wanted to make sure i wasnt Doing Things Wrong™
17:58 mgagne why should someone use keystone::resource::service_identity instead of keystone::endpoint? https://github.com/stackforge/puppet-key​stone/blob/master/manifests/endpoint.pp
17:59 mgagne patcable: it used to be that way to support service without support for rabbit_hosts
17:59 mgagne patcable: we could "always" use rabbit_hosts even though you only have one server but some services didn't support it at that time
17:59 mgagne patcable: and rabbit_hosts is supposed to override rabbit_host if defined
18:00 mgagne crinkle: ping
18:00 patcable mgagne: in our experience, we've had to explicitly ensure rabbit_host was not set if rabbit_hosts was set
18:00 crinkle mgagne: pong
18:00 patcable mgagne: or, ensure => absent for that parameter
18:01 mgagne patcable: ok, situation might have changed since because I had to read the code to make sure rabbit_hosts was used instead of rabbit_host if defined
18:01 mgagne crinkle: why should someone use keystone::resource::service_identity instead of keystone::endpoint? https://github.com/stackforge/puppet-key​stone/blob/master/manifests/endpoint.pp
18:01 mgagne crinkle: it makes as much sense as deprecating nova::auth::keystone and asking people to use keystone::resource::service_identity instead
18:03 EmilienM mgagne: I think it was the idea
18:03 EmilienM mgagne: but my patches are for now to have a feature parity
18:04 EmilienM I could add a warning to say we deprecate
18:04 mgagne EmilienM: I think keystone::resource::service_identity is a defined resource and you can't control it through Hiera
18:05 EmilienM true
18:05 mgagne I would like the deprecation warning to be removed
18:05 EmilienM it's up to the composition layer
18:05 EmilienM mgagne: did I add a warning?
18:05 mgagne I'm continuing work on listing deprecated parameters
18:05 mgagne EmilienM: yes
18:05 mgagne https://github.com/stackforge/puppet-keyst​one/blob/master/manifests/endpoint.pp#L89
18:13 EmilienM mgagne: just wantes to highlight sbadia +2 the patches, if you think it's not correct, don't hesitate to -2
18:13 mgagne EmilienM: link?
18:14 EmilienM mgagne: https://review.openstack.org/#/q/branch:maste​r+topic:bp/common-openstack-identity-resource,n,z
18:14 mgagne EmilienM: those are ok because they don't deprecate anything afaik
18:14 mgagne EmilienM: except the one in keystone
18:15 EmilienM mgagne: there is no patch in p-keystone
18:15 EmilienM in review I meant
18:15 mgagne EmilienM: could you please follow the link I posted?
18:15 sbadia :)
18:15 sbadia (hi)
18:15 mgagne EmilienM: I said they look ok because *unlike* keystone, they don't deprecate stuff
18:16 EmilienM mgagne: I like your good mood :)
18:17 mgagne EmilienM: I find it hard sometimes to get my point across, might be a language barrier just like when you can lost meaning if you double translate text
18:20 openstackgerrit Emilien Macchi proposed stackforge/puppet-sahara: Refactorise Keystone resources management  https://review.openstack.org/136394
18:21 openstackgerrit Mathieu Gagné proposed stackforge/puppet-keystone: Undeprecate keystone::endpoint class  https://review.openstack.org/137164
18:24 openstackgerrit Merged stackforge/puppet-tuskar: Refactorise Keystone resources management  https://review.openstack.org/136405
18:26 cwolferh joined #puppet-openstack
18:26 mgagne EmilienM: I'll blame my mood on the after OpenStack summit crash =)
18:27 mgagne EmilienM: I also thought about having a OpenStack Puppet code sprint in Montreal to we can clear as much technical debts and reviews as possible
18:30 openstackgerrit Merged stackforge/puppet-keystone: Undeprecate keystone::endpoint class  https://review.openstack.org/137164
18:33 openstackgerrit Colleen Murphy proposed stackforge/puppet-openstack-specs: Use OpenStackClient in Module Resources  https://review.openstack.org/136186
18:36 sputnik13 joined #puppet-openstack
18:47 sputnik13 joined #puppet-openstack
18:47 openstackgerrit Mathieu Gagné proposed stackforge/puppet-cinder: Remove deprecated parameters  https://review.openstack.org/137170
18:51 xarses joined #puppet-openstack
18:53 openstackgerrit Mathieu Gagné proposed stackforge/puppet-glance: Remove deprecated parameters  https://review.openstack.org/137172
18:59 openstackgerrit Mathieu Gagné proposed stackforge/puppet-horizon: Remove deprecated parameters  https://review.openstack.org/137174
19:06 mschmitt joined #puppet-openstack
19:07 xarses joined #puppet-openstack
19:11 mmaglana joined #puppet-openstack
19:18 openstackgerrit Mathieu Gagné proposed stackforge/puppet-keystone: Remove deprecated parameters  https://review.openstack.org/137178
19:26 mmaglana_ joined #puppet-openstack
19:30 EmilienM mgagne: that's an awesome idea
19:33 openstackgerrit Mathieu Gagné proposed stackforge/puppet-neutron: Remove deprecated parameters  https://review.openstack.org/137180
19:33 EmilienM mgagne: you should shoot an email to the puppet/operators mailing list maybe but I'm not sure
19:34 mgagne EmilienM: I haven't figured out the details yet
19:34 mgagne EmilienM: we should come up with a list of point pains we wish to address or features to implement
19:35 mgagne EmilienM: we wish would also be to have a sprint in french =)
19:35 EmilienM mgagne: we need to define if we will eat poutine or not
19:35 mgagne EmilienM: haha, we should sure
19:35 EmilienM the productivity would depend of this choice.
19:35 mgagne EmilienM: poutine au poulet :D
19:35 EmilienM gosh
19:35 mgagne EmilienM: it's gorgeous
19:37 cwolferh joined #puppet-openstack
19:37 mgagne http://s3.amazonaws.com/foodspotting-ec2/​reviews/2207633/thumb_600.jpg?1344696763 :D
19:37 mgagne unfortunately, pictures always make it less appealing :-/
19:37 EmilienM oh my god
19:37 EmilienM how can you be productive all the afternoon after this?
19:37 mgagne I don't know tbh lol
19:38 mgagne There is also one with smoked meat at La belle province
19:38 mgagne http://www.followmefoodie.com/wp-content​/uploads/2011/11/Montreal-Poutine-7.jpg
19:39 EmilienM mgagne: is it healthy? :P
19:39 mgagne if you don't know the place already, there is "La banquise" which has a menu mainly composed of different kinds of poutine
19:39 mgagne EmilienM: poutine is always healthy :D
19:40 EmilienM mgagne: let's to the Puppet sprint there
19:40 mgagne EmilienM: we won't get anything done :P
19:40 mgagne EmilienM: place is open 24h and very crowded
19:41 EmilienM mgagne: maybe but we will enjoy the food :D
19:41 mgagne EmilienM: :P
19:41 spredzy La banquiiiiiiiise. When ?
19:41 mgagne EmilienM: only *large* poutine I finished at La banquise was "La T-Rex" which is composed of ground beef, pepperoni, bacon and hot-dog sausage. It was years ago and I never managed to do it again =(
19:42 * spredzy sorry for disrupting the convo
19:42 EmilienM lol
19:42 EmilienM spredzy: please don't disturb us with puppet things
19:43 mgagne haha
19:45 tfz joined #puppet-openstack
19:46 richm come for the puppet - stay for the poutine
19:46 EmilienM richm: ahah, please go ahead !
19:49 mmaglana joined #puppet-openstack
20:02 mmaglana joined #puppet-openstack
20:02 xarses joined #puppet-openstack
20:06 openstackgerrit Mathieu Gagné proposed stackforge/puppet-nova: Remove deprecated parameters  https://review.openstack.org/137186
20:07 openstackgerrit Mathieu Gagné proposed stackforge/puppet-swift: Remove deprecated class  https://review.openstack.org/137187
20:11 openstackgerrit Mathieu Gagné proposed stackforge/puppet-cinder: Remove deprecated parameters  https://review.openstack.org/137170
20:12 mgagne alright, that should make our manifests much cleaner I guess
20:13 mgagne especially this one: https://review.openstack.org/#/c/135840/
20:18 EmilienM mgagne: in your composition layer, you only define neutron::plugin for neutron server right?
20:18 mgagne EmilienM: yes
20:18 EmilienM mgagne: not for agents, like l3 or dhcp, etc, right?
20:18 mgagne EmilienM: I don't
20:18 EmilienM ok cool
20:18 mgagne EmilienM: the agent is enough
20:19 mgagne EmilienM: I once saw commit to the RPC part of the agent to make it compatible with ml2
20:19 EmilienM indeed
20:20 mgagne EmilienM: so this tells me that the agent doesn't really care if it's the ovs plugin or ml2 as the RPC calls are standardized in some way
20:31 mgagne EmilienM: In *::keystone::auth how about we use Keystone::resource::service_identity[$auth_name] to establish notification relationship with service?
20:32 mgagne EmilienM: L95 and L98 https://review.openstack.org/#/c/​135842/1/manifests/keystone/auth.pp
20:34 mgagne EmilienM: would that work? because if you set $configure_endpoint to false, current relationship will fail
20:34 mgagne EmilienM: not that it was better before
20:34 EmilienM mgagne: yeah I did not change the behavior
20:35 EmilienM mgagne: I think we should add " if $configure_endpoint" before the dependency
20:51 social joined #puppet-openstack
21:03 gildub joined #puppet-openstack
21:04 mmaglana joined #puppet-openstack
21:07 imcsk8 joined #puppet-openstack
22:13 mdorman joined #puppet-openstack
22:18 JayJ joined #puppet-openstack
22:18 openstackgerrit Patrick Cable proposed stackforge/puppet-cinder: fixed behavior for rabbit_hosts variable  https://review.openstack.org/137227
22:25 mdorman_ joined #puppet-openstack
22:26 patcable woo, i was able to successfully do a review! baby steps.
22:36 EmilienM patcable: \o/, nice
22:37 patcable my openstack foundation account appeared to be created, but it wasnt actually created
22:37 patcable which was problematic
22:38 rcallawa joined #puppet-openstack
22:49 rcallawa joined #puppet-openstack
22:58 EmilienM mgagne: do you use Pacemaker in your OpenStack infra?
22:59 mgagne EmilienM: I don't think so. The only part that could use it is the database and I think we use something else
22:59 EmilienM mgagne: in a/p ? you don't run galera?
22:59 mgagne EmilienM: we don't use galera
23:00 mgagne EmilienM: I think pacemaker is used though
23:00 mgagne it's installed on the database server (I didn't install/configure it)
23:00 EmilienM we run galera with one MASTER and 2 backups in real-only and we configured HAproxy to have the readonly binding for Nova database
23:00 mgagne EmilienM: sure, we just didn't find the time to test and install galera
23:00 EmilienM mgagne: ok, so you consume your infra to install OpenStack
23:01 mgagne EmilienM: we used a time proven recipe due to time constraint
23:01 EmilienM ok
23:01 EmilienM makes sense
23:04 mgagne EmilienM: come to think of it, we still have to get ride of nova::generic_service like we said at the summit
23:04 mgagne EmilienM: bug is open, just need someone to do it =)
23:04 EmilienM oh that one
23:05 mgagne EmilienM: https://bugs.launchpad.net​/puppet-nova/+bug/1392074
23:06 EmilienM that one...
23:06 mgagne EmilienM: I don't mind working on it though
23:06 mgagne EmilienM: just need to find a free slot in the schedule
23:07 mgagne EmilienM: debugging intermittent issues in our internal gate :-/
23:07 mgagne EmilienM: sometimes, I get errors in the sql migration of glance or cinder
23:07 EmilienM oh
23:07 mgagne EmilienM: like table exists already
23:07 EmilienM I spent all the day to debug a MTU issue in Neutron
23:07 mgagne haha
23:08 mgagne EmilienM: spent 2 days debugging and trying to upgrade our vagrant version. All this so the console output is prefixed with hostname of the provisioned box
23:09 mgagne EmilienM: because we provision 30 boxes at the same time and all output is mixed up. You don't know which box throws which output
23:10 EmilienM mgagne: you use vagrant for CI or in prod also?
23:10 mgagne EmilienM: CI only
23:10 EmilienM oh ok
23:10 mgagne EmilienM: we have an extensive vagrant setup
23:10 mgagne EmilienM: especially the openstack project which is in fact a replica of our prod setup
23:11 EmilienM mgagne: I guess it works well with Linux Bridge :P
23:11 EmilienM (b/c with OVS it's tricky tricky)
23:11 mgagne EmilienM: we use a fork of vagrant-openstack to spawn vagrant boxes in an internal openstack cloud
23:12 mgagne EmilienM: instead of VirtualBox because otherwise devs would have to get 256GB laptop
23:13 EmilienM oh nice
23:13 mgagne EmilienM: so for instance, Vagrantfile rapidly grew too big for any sane human being. So I found a way to define boxes in a yaml file instead and Vagrantfile reads it.
23:13 mgagne EmilienM: + I managed to define dependencies between boxes which an external script can consume to provision boxes in the right order
23:14 EmilienM oh dear
23:14 mgagne EmilienM: + tag support for boxes
23:14 EmilienM mgagne: you almost did an OpenStack product \o/
23:14 mgagne haha
23:15 mgagne EmilienM: example of ONE box definition: https://gist.github.com/mg​agne/8c0ffecdd4a400f1c84c
23:16 mgagne EmilienM: yaml file has over 400 lines
23:16 EmilienM mgagne: it's cool
23:16 EmilienM mgagne: guess you can define puppet params inside
23:17 mgagne EmilienM: so far, it's part of the original Vagrantfile but I guess it shouldn't but hard to allow it to be configured in the yaml file
23:20 sputnik13 joined #puppet-openstack
23:22 mgagne EmilienM: http://paste.openstack.org/show/138437/
23:23 mgagne same with Nova this time :-/
23:24 mgagne EmilienM: haha, found it, thanks Vagrant 1.6
23:28 mgagne EmilienM: do you know if db sync is transactional?

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