Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-07-03

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

All times shown according to UTC.

Time Nick Message
00:17 aldavud joined #puppet-openstack
00:31 aimon joined #puppet-openstack
00:35 openstackgerrit Colleen Murphy proposed a change to stackforge/puppet-openstack-specs: Common OpenStack Database Resource  https://review.openstack.org/104289
00:39 openstackgerrit Spencer Krum proposed a change to stackforge/puppet-nova: Initial rspec-beaker testing scaffolding  https://review.openstack.org/102020
01:08 aldavud joined #puppet-openstack
01:11 gildub joined #puppet-openstack
01:12 StarBeast joined #puppet-openstack
01:17 djm68 joined #puppet-openstack
01:19 xarses joined #puppet-openstack
01:21 StarBeast joined #puppet-openstack
02:22 StarBeast joined #puppet-openstack
02:30 openstackgerrit Clayton O'Neill proposed a change to stackforge/puppet-cinder: Add region support for type creation/setting  https://review.openstack.org/104420
03:07 sputnik13 joined #puppet-openstack
03:08 djm68 joined #puppet-openstack
03:09 sputnik1_ joined #puppet-openstack
03:10 sputnik13 joined #puppet-openstack
03:11 sputnik13 joined #puppet-openstack
03:21 djm68 joined #puppet-openstack
03:23 StarBeast joined #puppet-openstack
04:23 StarBeast joined #puppet-openstack
05:07 hogepodge joined #puppet-openstack
05:24 StarBeast joined #puppet-openstack
06:05 ianw joined #puppet-openstack
06:25 StarBeast joined #puppet-openstack
06:29 tfz joined #puppet-openstack
06:39 aldavud joined #puppet-openstack
06:56 giorgiodinapoli joined #puppet-openstack
07:04 artem_panchenko joined #puppet-openstack
07:07 EmilienM good morning
07:11 openstackgerrit Gilles Dubreuil proposed a change to stackforge/puppet-vswitch: Refacfored a more suitable ovs_redhat provider  https://review.openstack.org/94504
07:16 aldavud joined #puppet-openstack
07:26 StarBeast joined #puppet-openstack
07:35 mmaglana joined #puppet-openstack
07:55 StarBeast joined #puppet-openstack
08:05 derekh_ joined #puppet-openstack
08:11 aldavud joined #puppet-openstack
09:07 mmagr joined #puppet-openstack
09:14 sanjayu joined #puppet-openstack
09:23 dgollub joined #puppet-openstack
09:36 mmaglana joined #puppet-openstack
09:43 qba73 joined #puppet-openstack
10:28 marun joined #puppet-openstack
10:53 gildub joined #puppet-openstack
10:59 gildub hogepodge, EmilienM, can you please merge https://review.openstack.org/#/c/100377/ ?
11:00 derekh_ joined #puppet-openstack
11:01 comptona joined #puppet-openstack
11:14 comptona joined #puppet-openstack
11:15 dgollub joined #puppet-openstack
11:32 aldavud gildub: I think this is now a duplicate due to https://review.openstack.org/#/c/104274/
11:35 gildub aldavud, that's excellent! Thank you.
11:43 openstackgerrit Gilles Dubreuil proposed a change to stackforge/puppet-vswitch: Refacfored a more suitable ovs_redhat provider  https://review.openstack.org/94504
11:44 comptona joined #puppet-openstack
11:51 openstackgerrit Gilles Dubreuil proposed a change to stackforge/puppet-vswitch: Refacfored a more suitable ovs_redhat provider  https://review.openstack.org/94504
11:55 dprince joined #puppet-openstack
11:58 openstackgerrit Gilles Dubreuil proposed a change to stackforge/puppet-vswitch: Refacfored a more suitable ovs_redhat provider  https://review.openstack.org/94504
12:07 comptona joined #puppet-openstack
12:10 gildub joined #puppet-openstack
12:20 comptona joined #puppet-openstack
12:27 EmilienM sbadia: https://github.com/spotify/puppetexplorer
12:28 sbadia nice!
12:28 sbadia thanks
12:42 tfhartmann joined #puppet-openstack
12:53 csschwe joined #puppet-openstack
13:10 mjblack joined #puppet-openstack
13:14 morazi joined #puppet-openstack
13:29 comptona joined #puppet-openstack
13:47 openstackgerrit Lukas Bezdicka proposed a change to stackforge/puppet-nova: Remove ability to manage the nova uid/gid.  https://review.openstack.org/104550
13:54 openstackgerrit Lukas Bezdicka proposed a change to stackforge/puppet-nova: Remove ability to manage the nova uid/gid.  https://review.openstack.org/104550
14:07 daneyon joined #puppet-openstack
14:08 daneyon joined #puppet-openstack
14:16 meylor joined #puppet-openstack
14:21 prad joined #puppet-openstack
14:30 hogepodge_ joined #puppet-openstack
14:30 hogepodge_ left #puppet-openstack
14:32 prad joined #puppet-openstack
14:43 comptona joined #puppet-openstack
14:53 comptona joined #puppet-openstack
15:14 comptona joined #puppet-openstack
15:18 openstackgerrit Lukas Bezdicka proposed a change to stackforge/puppet-nova: Remove ability to manage the nova uid/gid.  https://review.openstack.org/104583
15:26 social could please anyone look at this puppet-nova pull? ^^
15:32 comptona joined #puppet-openstack
15:41 openstackgerrit Lukas Bezdicka proposed a change to stackforge/puppet-nova: Remove ability to manage the nova uid/gid.  https://review.openstack.org/104583
15:44 openstackgerrit Colleen Murphy proposed a change to stackforge/puppet-openstack-specs: Common OpenStack Database Resource  https://review.openstack.org/104289
15:47 comptona joined #puppet-openstack
16:02 mmaglana joined #puppet-openstack
16:05 xarses joined #puppet-openstack
16:18 kaiserpathos joined #puppet-openstack
16:19 EmilienM hogepodge: puppet-trove seems ready to land in Stackforge: https://review.openstack.org/#/c/104244/
16:19 EmilienM we are currently performing functional testing, but the code is here, and I think we can continue the work in gerrit.
16:22 xarses_ joined #puppet-openstack
16:30 aimon joined #puppet-openstack
16:35 rickerc joined #puppet-openstack
16:35 aldavud joined #puppet-openstack
16:56 hogepodge Thanks EmilienM. I’ll take a look
16:57 hogepodge Ah, Precise failure. :-/
17:11 aimon joined #puppet-openstack
17:45 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-neutron: Link from ovs config file to plugin.ini on RedHat  https://review.openstack.org/104624
17:46 hogepodge EmilienM https://review.openstack.org/#/c/104624/1 This bug was the source of my headaches before. :-)
17:47 hogepodge I love the cherry pick button.
17:52 hogepodge mgagne A couple of suggestions https://review.openstack.org/#/c/104278/1
17:55 mgagne hogepodge: replied
18:13 sputnik13 joined #puppet-openstack
18:38 aldavud_ joined #puppet-openstack
19:51 zeus joined #puppet-openstack
19:55 morazi joined #puppet-openstack
19:55 kaiserpathos joined #puppet-openstack
20:00 aldavud joined #puppet-openstack
20:24 uberj question about openvswitch (seems there are people in this channel that know about it): When I do ovs-vsctl show and I see something like:
20:24 uberj Bridge br-int
20:24 uberj Port br-int
20:24 uberj what is that saying?
20:24 uberj the self reference doesn't make sense to me
20:27 Quack joined #puppet-openstack
20:34 daneyon joined #puppet-openstack
20:38 openstackgerrit Spencer Krum proposed a change to stackforge/puppet-nova: Initial rspec-beaker testing scaffolding  https://review.openstack.org/102020
20:39 EmilienM uberj: that means you have a br-int which is your integration bridge in Neutron
20:41 uberj so are there really two different objects we are calling 'br-int'? One bridge and one interface?
20:42 daneyon_ joined #puppet-openstack
20:51 openstackgerrit Spencer Krum proposed a change to stackforge/puppet-nova: Initial rspec-beaker testing scaffolding  https://review.openstack.org/102020
20:53 sputnik13 joined #puppet-openstack
20:56 aldavud joined #puppet-openstack
20:56 openstackgerrit Russell Bryant proposed a change to stackforge/puppet-nova: Update incorrect comment about nova-network status  https://review.openstack.org/104685
20:58 mgagne hogepodge: ping
20:58 hogepodge hi mgagne
20:59 mgagne hogepodge: regarding the author field, I still have questions =)
20:59 mgagne hogepodge: some modules were not originally authored by Puppet Labs (puppet-designate, ceilometer, etc.)
20:59 mgagne hogepodge: what should I put?
21:00 mgagne hogepodge: if we need to discriminate for each modules, I would rather prefer to just put "Stackforge Contributors" everywhere and be done
21:01 hogepodge The original authors.
21:01 hogepodge eNovance, etc
21:01 hogepodge That’s what I was thinking
21:04 EmilienM hogepodge: your +1 on the patch for moving puppet-trove on stackforge would allow me to ping infra guys and say it's ok
21:04 Underyx|off joined #puppet-openstack
21:06 hogepodge Right, I didn’t realise that gate was non-voting.
21:06 hogepodge Sorry about the delay on it.
21:06 EmilienM hogepodge: no worries, there is no delay.
21:06 hogepodge EmilienM, as an eNovance guy, do you have any opinions on mgagne’s suggestion for the author field?
21:06 EmilienM hogepodge: so, tomorrow I'm starting serious testing
21:07 EmilienM hogepodge: please let me read the context
21:07 mgagne hogepodge: I'm leaving eNovance as the original author
21:07 EmilienM mgagne: we already had this discussion
21:07 EmilienM from my opinion I don't care
21:07 mgagne EmilienM: =)
21:07 EmilienM because we are all contributors
21:07 mgagne EmilienM: poor violin
21:07 EmilienM I mean, OK eNovance started the contributions
21:08 hogepodge Founder recognition. :-)
21:08 EmilienM but since we got many patches the module is part of Stackforge contributors
21:08 EmilienM so yes, Stackforge Contributors or OpenStack Foundation would make more sense
21:08 hogepodge community recognition! \o/
21:08 EmilienM one condition though
21:08 EmilienM I want my puppetlabs tee shirt each time there is a new one
21:09 mgagne go get it at the summit :P
21:09 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-ceilometer: Make Modulefile summary and description consistent  https://review.openstack.org/104277
21:09 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-cinder: Make Modulefile summary and description consistent  https://review.openstack.org/104278
21:09 EmilienM mgagne: you're right, I'm too nice, I should ask something else
21:09 hogepodge Beer cozy?
21:10 EmilienM hogepodge: you know to talk to me
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-designate: Make Modulefile summary and description consistent  https://review.openstack.org/104279
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-glance: Make Modulefile summary and description consistent  https://review.openstack.org/104280
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-heat: Make Modulefile summary and description consistent  https://review.openstack.org/104281
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-horizon: Make Modulefile summary and description consistent  https://review.openstack.org/104282
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-ironic: Make Modulefile summary and description consistent  https://review.openstack.org/104283
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-keystone: Make Modulefile summary and description consistent  https://review.openstack.org/104284
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-neutron: Make Modulefile summary and description consistent  https://review.openstack.org/104285
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-nova: Make Modulefile summary and description consistent  https://review.openstack.org/104286
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-swift: Make Modulefile summary and description consistent  https://review.openstack.org/104287
21:10 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-tempest: Make Modulefile summary and description consistent  https://review.openstack.org/104288
21:13 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-designate: Make Modulefile summary and description consistent  https://review.openstack.org/104279
21:13 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-horizon: Make Modulefile summary and description consistent  https://review.openstack.org/104282
21:14 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-keystone: Make Modulefile summary and description consistent  https://review.openstack.org/104284
21:15 mgagne that should be it I guess
21:16 EmilienM mgagne: "eNovance & Co" that sounds like weird, no?
21:16 mgagne EmilienM: I wrote that?
21:16 EmilienM but anyway, I'm +2 the patches
21:17 EmilienM no no, but you wrote eNovance & Stackforge contribs
21:17 EmilienM but I'm not a licensing guy, I'm good with it
21:17 mgagne yes, hogepodge suggested it and I like it =)
21:18 mgagne "Co" could be anyone, at least with "StackForge" we are a bit less generic imo
21:18 mgagne EmilienM: sounds less like: eNovance et la racaille :P /jk
21:19 EmilienM \o/
21:19 EmilienM mgagne: again, I really don't mind. What I care now, is I would like more contributors involved in our modules.
21:19 mgagne EmilienM: 100% agree
21:19 EmilienM or at least more contributions.
21:20 mgagne EmilienM: currently working on ML2 linuxbridge support
21:20 EmilienM it's good to participate to meetings, to talk, blabla.. but there is time to code
21:20 EmilienM mgagne: that's awesome
21:20 mgagne EmilienM: 100% agree, unfortunately, I got burn in the last months =(
21:20 EmilienM mgagne: I'm not talking about you at all.
21:21 EmilienM and nobody in particular
21:21 mgagne EmilienM: oh, I don't mind if it was the case =)
21:21 EmilienM "just saying"
21:23 mgagne EmilienM: I'm currently working on other personal projects =) One of them: http://imgur.com/a/ybmsy
21:25 EmilienM mgagne: gosh, what's this
21:25 mgagne EmilienM: a project called ambilight
21:25 mgagne http://en.wikipedia.org/wiki/Ambilight
21:25 mgagne I went the DIY way
21:26 EmilienM mgagne: so what, did you buy a screen and then hacked it?
21:26 mgagne EmilienM: I glued a LED strip behind the screen and drive it with both an arduino and raspberry pi
21:26 mgagne EmilienM: for non-DIY people, there was a kickstarter project: http://lightpack.tv/
21:29 EmilienM mgagne: very interesting, thanks. I did not know about this before
21:37 hogepodge new to me. looks pretty cool.
21:37 hogepodge how does it determine the edge colors?
21:38 hogepodge Ah, I see. Have to run from a supported device.
21:45 bodepd markvoelker: ping
21:45 tfhartmann joined #puppet-openstack
21:45 bodepd markvoelker: actually nm... :)
21:59 Hunner Wow, that's pretty cool
22:00 mgagne hogepodge: a daemon is taking screenshots and analyzing the edge
22:01 mgagne hogepodge: and you have to map your leds with the actual physical position around the screen
22:24 aldavud joined #puppet-openstack
22:27 mgagne EmilienM: still around?
22:30 aldavud joined #puppet-openstack
22:31 mgagne anyone remember why this parameter got deprecated ? https://github.com/stackforge/puppet-neutron/blob/master/manifests/plugins/ml2.pp#L97
22:36 mgagne I think I'll need a Neutron expert, what I'm reading in puppet-neutron doesn't make sense to me...
22:39 mtl1 joined #puppet-openstack
22:43 sputnik13 joined #puppet-openstack
22:44 hogepodge magane, it got moved
22:45 hogepodge That code didn’t work at release, and got patched later.
22:45 mgagne hogepodge: where is enable_security_group configured now?
22:45 hogepodge It’s configured by the firewall provider.
22:46 hogepodge In the agent. https://github.com/stackforge/puppet-neutron/blob/master/manifests/agents/ml2/ovs.pp
22:46 mgagne I don't see enable_security_group
22:46 hogepodge firewall driver that is
22:46 hogepodge It’s implied.
22:46 mgagne hogepodge: why is the doc referring to enable_security_group still?
22:46 mgagne http://docs.openstack.org/trunk/install-guide/install/yum/content/neutron-ml2-compute-node.html
22:47 mgagne hogepodge: and config is not deprecated: https://github.com/openstack/neutron/blob/master/neutron/agent/securitygroups_rpc.py#L30-L36
22:48 hogepodge let me look
22:48 mgagne hogepodge: from the original change, firewall_driver got moved and enable_security_group marked as deprecated (without warning)
22:48 EmilienM mgagne: yes
22:49 mgagne EmilienM: what was the original intention?
22:49 EmilienM so
22:49 EmilienM it was in havana
22:49 mgagne EmilienM: was enable_security_group supported to be deprecated?
22:49 mgagne supposed*
22:49 EmilienM yes
22:49 mgagne EmilienM: you mean enable_security_group was supposed to be deprecated but it's no longer the case?
22:50 EmilienM mgagne: since icehouse we just have to specify the firewall_driver thing, and it's in agent
22:50 EmilienM not in plugin
22:50 mgagne EmilienM: why is the doc and code still referring to enable_security_group ?
22:50 mgagne EmilienM: without any mention of deprecation?
22:50 EmilienM that's the way to go to have security group working: https://github.com/stackforge/puppet-neutron/blob/master/manifests/agents/ml2/ovs.pp#L100
22:51 mgagne EmilienM: I guess you can understand that puppet-neutron has no authority on how it should be done, neutron is the reference and so far, I can't find one
22:52 mgagne EmilienM: https://github.com/openstack/neutron/commit/f87e7d964c19cc2be33226df66f0c823af993d49
22:52 EmilienM mgagne: what I'm saying is that if you want to disable the SG api, just put the neutron.agent.firewall.NoopFirewallDriver thing in firewall_driver
22:52 mgagne EmilienM: that's not what the above commit suggests
22:53 EmilienM indeed, yes
22:53 EmilienM that stuff is so weird
22:53 EmilienM over the last months, it has change like several times...
22:53 mgagne EmilienM: so it's a combination of firewall_driver and enable_security_group
22:53 EmilienM yeah, what I did in the past
22:54 mgagne EmilienM: enable_security_group is not deprecated, on the contrary, it got introduced in Icehouse
22:54 mgagne EmilienM: we should therefore remove the non-existant deprecation warning and allow its configuration
22:54 EmilienM mgagne: ok so let's fix this, totally agree
22:54 mgagne EmilienM: cool, one less confusion
22:55 EmilienM if you want, I can take it, but tomorrow. I was heading my bed
22:55 mgagne doing it atm
22:55 EmilienM ok
22:55 EmilienM mgagne: something we will backport.
22:55 mgagne true
22:55 mgagne we are on juno now :P
22:56 EmilienM mgagne: you should manage the "neutron.agent.firewall.NoopFirewallDriver" thing, when enable_security_group is at False
22:56 EmilienM if you want to follow the code
22:57 mgagne EmilienM: I'll try but afaik, neutron-plugin-ml2 is not installed on the compute node right?
22:57 EmilienM mgagne: no
22:57 mgagne EmilienM: I meant, neutron::plugins::ml2
22:58 EmilienM mgagne: yep, right
22:58 mgagne EmilienM: so I can't cross-check those values, they aren't managed by the same manifest
22:58 EmilienM mgagne: right
22:58 EmilienM mgagne: the SG stuff is only for agent
22:59 mgagne will continue to dig neutron::plugins::ml2, there is stuff for linuxbridge plugin in there and I don't understand why it's there
22:59 hogepodge Thanks EmilienM and mgagne.
22:59 hogepodge Neutron...
22:59 EmilienM mgagne: ok, and me I'm digging my bed
22:59 EmilienM hogepodge: heh
22:59 mgagne EmilienM: hehe, good night
22:59 hogepodge ¯\_(ツ)_/¯
22:59 EmilienM tomorrow: last day on puppet-trove, more testing on real servers...
23:00 hogepodge I upgraded our local installation from grizzly to icehouse. Got it all except for the novnc server. no idea what’s going on
23:01 mgagne hogepodge: make sure to update the websockify package
23:01 mgagne hogepodge: ubuntu (if you are with Ubuntu) has junk dependencies
23:01 hogepodge is that why I’m getting a bad rpc version?
23:01 hogepodge no, centos
23:01 mgagne hogepodge: hmmm
23:02 aldavud joined #puppet-openstack
23:02 mgagne hogepodge: I had bad rpc version too and it was because I tried to run in compatibility mode, it didn't work out for me, compute nodes were too old (not running the latest havana) for the compat mode
23:03 hogepodge compatibility mode?
23:03 mgagne hogepodge: you can run an Icehouse controller with Havana compute nodes
23:03 mgagne hogepodge: didn't work that well for me
23:03 hogepodge Oh, it shouldn’t be there.
23:04 mgagne hogepodge: ok, otherwise make sure all services have restarted
23:04 hogepodge The upgrade did some funny things, though. Variables were unset to blank rather than removed, and so on.
23:04 mgagne hogepodge: variables?
23:04 hogepodge Yup, everything is go. It’s the only puzzle piece left.
23:05 hogepodge It was silly stuff. Things that I think were set by default in grizzly and later managed explicitly.
23:05 openstackgerrit Colleen Murphy proposed a change to stackforge/puppet-openstack-specs: Common OpenStack Database Resource  https://review.openstack.org/104289
23:11 michchap joined #puppet-openstack
23:14 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-neutron: Undeprecate enable_security_group parameter  https://review.openstack.org/104716
23:14 social joined #puppet-openstack
23:18 hogepodge Good night everyone. Happy 4th to us americans! (and everyone else too, if you’re so inclined)
23:18 mgagne have a good one
23:19 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-neutron: Simplify ml2_config.ini symlink resource management  https://review.openstack.org/104717
23:20 hogepodge magane if you have a second. https://review.openstack.org/#/c/104624/
23:21 mgagne hogepodge: now it's damn confusion
23:21 mgagne hogepodge: neutron::plugins::ml2 is creating a symlink but the other way
23:21 hogepodge the patch went in a bit strange
23:21 hogepodge It’s a redhat packaging bug
23:22 mgagne hogepodge: oh, it's for ml2_conf.ini
23:22 mgagne hogepodge: strange how we have linuxbridge_config_file but not ovs_config_file
23:22 mgagne in params
23:22 hogepodge Right. neutron is looking for the plugin file. It redhat ships with the old monolithic plugin linked, but that’s deprecated in favor of ml2 in mainline.
23:23 hogepodge We can fix that upstream. I’m mainly interested in the backport so that ml2 will work in the released modules.
23:23 mgagne hogepodge: so let me guess, we will now have: /etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini -> /etc/neutron/plugin.ini -> /etc/neutron/plugins/ml2/ml2_conf.ini
23:24 hogepodge Only if you set up the ovs_neutron_plugin
23:24 hogepodge Both can’t be active at once. It’s a bug if you do that
23:24 mgagne hogepodge: the other link is in neutron::plugins::ml2
23:25 mgagne hogepodge: https://github.com/stackforge/puppet-neutron/blob/master/manifests/plugins/ml2.pp#L125-L129
23:25 mgagne hogepodge: so I'm confused, does /etc/neutron/plugin.ini exist or not?
23:26 hogepodge it’s always symlinked
23:26 hogepodge (or a copy of a config)
23:26 mgagne hogepodge: why are we creating a symlink then if neutron::agents::ml2::ovs expects it to exist?
23:27 hogepodge On redhat, the link points to the monolithic. for ml2 it needs to be switched.
23:29 hogepodge Otherwise you never see the configuration and the ml2 networks don’t wonk
23:29 hogepodge work
23:29 mgagne hogepodge: ok, lets say I setup the agent only. By default, will plugin.ini point to the monolithic plugin config?
23:30 mgagne hogepodge: won't we have a circular symlink?
23:30 hogepodge Which code is creating the link?
23:30 mgagne hogepodge: yours?
23:30 hogepodge I’m confused about what you’e asking. What’s the circular?
23:31 mgagne hogepodge: I should probably get a centos running with RDO
23:31 hogepodge Out of the box: plugin.ini -> /etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini
23:31 mgagne hogepodge: right, for the agent, the proposed change is to point /etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini to plugin.ini
23:32 hogepodge After you configure ml2 with the code above: plugin.ini -> /etc/neutron/plugins/ml2/ml2_conf.ini
23:32 mgagne hogepodge: how is this not a circular symlink?
23:32 hogepodge Puppet will break the link and reassign it to the correct file.
23:32 mgagne hogepodge: no, AFAIK, you don't need neutron-plugin-ml2 with the agent
23:33 hogepodge Agent reads the neutron config.
23:33 mgagne hogepodge: there is no dependency from neutron::agents::ml2::ovs to neutron::plugins::ml2
23:33 hogepodge I’m confused about what you’re asking.
23:33 mgagne hogepodge: and it's neutron::plugins::ml2 that is setting up the symlink from plugin.ini to ml2_conf.ini
23:34 hogepodge Correct. You have to choose a core plugin. RDO configures a deprecated link in packaging. For ml2 we need to break that, so the plugin code does it.
23:34 mgagne hogepodge: the core plugin is on the node running neutron-server
23:34 mgagne hogepodge: there is no requirement for the agent to run on this particular node
23:35 mgagne hogepodge: so no requirement to have neutron::plugins::ml2 on your agent node
23:35 hogepodge I have a hard stop right now. Need to be somewhere.
23:35 mgagne hogepodge: will check tomorrow then
23:35 hogepodge We can continue this later. Can you e-mail me the details so I can write back?
23:35 mgagne hogepodge: in the review?
23:36 hogepodge Sure, the review is fine. It’s a backport from master.
23:36 mgagne really? oh god =(
23:36 hogepodge Addressing a bug.
23:36 mgagne hogepodge: will check with michchap
23:36 hogepodge he’s on vacation. :-P
23:36 mgagne michchap: damn you
23:37 hogepodge as far as I can tell it’s all kosher.
23:37 mgagne hogepodge: from my POV, it only works if you have an all-in-one setup
23:38 hogepodge because the agen’t can’t see it?
23:38 mgagne because if you don't install neutron::plugins::ml2, plugin.ini points to the monolithic config file by default (according to comments)
23:39 hogepodge Lots of code is broken in that sense. For multi node I have to send service configurations all over the place and disable the servies to get things to work
23:39 mgagne if you only install neutron::agents::ml2::ovs, plugin.ini won't be able to point to ml2_conf.ini
23:40 mgagne and the fix backported will create a circular symlink as there is no requirement to install neutron::plugins::ml2 with the agent
23:40 hogepodge I’m not 100% sure is the agent actually needs to read the config.
23:40 hogepodge No, definitely no circular symlink.
23:40 hogepodge Oh, I need to go. I’m really late.
23:40 mgagne cya!
23:57 derekh_ joined #puppet-openstack

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