| *** markvoelker has quit IRC | 00:02 | |
| *** BjoernT has quit IRC | 00:12 | |
| *** sdake_ has joined #openstack-ansible | 00:14 | |
| *** sdake has quit IRC | 00:17 | |
| *** ccrouch-bbl is now known as ccrouch | 00:31 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 00:48 | |
| *** jmccrory has joined #openstack-ansible | 01:18 | |
| *** jmccrory has quit IRC | 01:18 | |
| *** ccrouch is now known as ccrouch-afk | 01:26 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 02:14 | |
| *** devlaps has quit IRC | 02:54 | |
| *** sdake has joined #openstack-ansible | 03:00 | |
| *** sdake_ has quit IRC | 03:04 | |
| *** sdake has quit IRC | 03:59 | |
| *** sdake has joined #openstack-ansible | 04:02 | |
| *** stevemar has joined #openstack-ansible | 04:08 | |
| *** sdake has quit IRC | 04:10 | |
| *** stevemar2 has joined #openstack-ansible | 04:32 | |
| *** stevemar has quit IRC | 04:35 | |
| *** markvoelker has joined #openstack-ansible | 04:39 | |
| *** markvoelker has quit IRC | 04:43 | |
| *** stevemar2 is now known as stevemar | 04:53 | |
| *** markvoelker has joined #openstack-ansible | 05:39 | |
| *** markvoelker has quit IRC | 05:43 | |
| *** ishant has joined #openstack-ansible | 06:36 | |
| *** markvoelker has joined #openstack-ansible | 06:40 | |
| *** markvoelker has quit IRC | 06:44 | |
| *** markvoelker has joined #openstack-ansible | 07:40 | |
| *** markvoelker has quit IRC | 07:45 | |
| *** stevemar has quit IRC | 07:59 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: Update cinder services to kilo https://review.openstack.org/168066 | 08:05 |
|---|---|---|
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Heat config changes for kilo https://review.openstack.org/167947 | 08:36 |
| *** markvoelker has joined #openstack-ansible | 08:41 | |
| *** markvoelker has quit IRC | 08:46 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Bump requirements in preparation for kilo https://review.openstack.org/166849 | 09:26 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Bump requirements in preparation for kilo https://review.openstack.org/166849 | 09:26 |
| *** markvoelker has joined #openstack-ansible | 09:42 | |
| *** markvoelker has quit IRC | 09:46 | |
| *** svg has joined #openstack-ansible | 09:47 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Add neutron neutron_num_sync_threads variable https://review.openstack.org/168287 | 10:41 |
| *** markvoelker has joined #openstack-ansible | 10:43 | |
| *** markvoelker has quit IRC | 10:48 | |
| svg | Is it possible to use the ansible rpc setup in combination with a ceph backend? Or perhaps possible but not recommended? If so why? thx | 11:09 |
| andymccr | svg: via the os-ansible-deployment plays, no - but it will setup the openstack bits, you could then add ceph as a backend in the same way as you normally would. I don't see why you wouldn't/couldn't except that it wouldn't be managed as part of the plays, and potentially the cinder configuration will be overwritten when you re-run the plays so that could be problematic for supporting it. | 11:16 |
| svg | so that would be the cinder config on e.g. the compute nodes? even if no cinder hosts are configured in the user config? | 11:17 |
| andymccr | svg: you could specify no cinder hosts, that'd work | 11:18 |
| andymccr | that way they wouldnt be managed by ansible (if you are looking to use ceph as a cinder backend that is) | 11:18 |
| svg | yes, no need for cinder, only ceph here | 11:18 |
| andymccr | ahh ok | 11:18 |
| andymccr | so the only issues are any shared config, e.g. config that needs to be adjusted to reference ceph that is also managed by ansible | 11:18 |
| svg | something else: It's not clear to me, how the scripts calculate how many containers of the different components get deployed, in only see one gets to configure a number of hosts as infra_hosts | 11:19 |
| andymccr | but of course, if thats something youre interested in, filing issues that prevent it from working/make it difficult to work could move it towards getting that going more efficiently | 11:19 |
| svg | (in the process of reviewing and selecting our options to manage openstack deployment) | 11:20 |
| andymccr | svg: we'll be exploring how we can add ceph support at some point - not sure when, but yeh currently its not "managed" but the actual OS install would work as normal if you added ceph on top manually | 11:22 |
| svg | thanks andymccr ; something else: It's not clear to me, how the scripts calculate how many containers of the different components get deployed, in only see one gets to configure a number of hosts as infra_hosts | 11:24 |
| andymccr | svg: so you can configure an affinity | 11:25 |
| andymccr | and set that to 0 if you dont want a specific container type | 11:25 |
| andymccr | or 2 or 3 if you want more of that container type on the same host | 11:26 |
| andymccr | let me find an example | 11:26 |
| svg | ok, and what happens by default? 1 of each type per physical infra_host? | 11:26 |
| svg | minimum is 3 nfra hosts, so that would mean 3 x 1 of each container type? | 11:27 |
| andymccr | svg: yeh 1 per host. | 11:27 |
| andymccr | you dont "need" 3, its just what we used as a default for resilience etc. we have an aio for gating which only uses 1 | 11:28 |
| andymccr | http://pastie.org/10057163 | 11:28 |
| andymccr | as an example i set to use swift as a standalone | 11:28 |
| svg | yes, that's what I presumed based on the aio setup | 11:29 |
| openstackgerrit | Merged stackforge/os-ansible-deployment-specs: Add master-kilofication specification https://review.openstack.org/166865 | 11:30 |
| svg | so right now I'm looking into doing a test deploy, and I have only three physical machines available right now | 11:30 |
| svg | I'm thinking of doing 2 hosts infra + logging, and one separate for computing | 11:31 |
| andymccr | that "should" work the only issue is that for db etc 2 is a bad idea for quorum, so its usually less beneficial to go with 2 than it would be to just stick with 1 or 3. | 11:31 |
| andymccr | but its just for testing so its not a real issue | 11:32 |
| svg | yes, sure, I'm aware; only to test the ease of deployment basically | 11:32 |
| svg | thank you for your advice andymccr | 11:33 |
| svg | btw, those playbooks are quite impressive; I've been an ansible user and contributor for two years, and this is refreshing advanced stuff | 11:33 |
| andymccr | thanks svg! if you have any issues deploying let us know | 11:35 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Bump requirements in preparation for kilo https://review.openstack.org/166849 | 11:36 |
| *** markvoelker has joined #openstack-ansible | 11:43 | |
| *** markvoelker has quit IRC | 11:48 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Bump requirements in preparation for kilo https://review.openstack.org/166849 | 11:53 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Bump requirements in preparation for kilo https://review.openstack.org/166849 | 12:13 |
| *** markvoelker has joined #openstack-ansible | 12:18 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Add neutron neutron_num_sync_threads variable https://review.openstack.org/168287 | 12:23 |
| *** sdake has joined #openstack-ansible | 12:51 | |
| *** galstrom_zzz is now known as galstrom | 12:56 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 12:57 | |
| svg | Is there a particular reason, why the proposed interfaces configuration is based on two separate bonds? | 12:58 |
| Sam-I-Am | thats what the rackspace dc uses | 13:04 |
| Sam-I-Am | so its the default | 13:05 |
| *** jaypipes is now known as leakypipes | 13:05 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Removed lxc templates that are no longer used (icehouse) https://review.openstack.org/168174 | 13:24 |
| openstackgerrit | Hugh Saunders proposed stackforge/os-ansible-deployment-specs: Manage resolv.conf https://review.openstack.org/168074 | 13:29 |
| *** ishant has quit IRC | 13:31 | |
| andymccr | svg: do you mean the interface sample confs? its to show how you can set out the networks using bonds on the physical hosts. we want to split management traffic and instance traffic (for example). | 13:35 |
| svg | yes, we have one big 2x10g bond for everything here | 13:35 |
| svg | so i keep just 1 bond0 | 13:36 |
| Sam-I-Am | theres no limitations or requirements, really. | 13:37 |
| andymccr | you could use 1 bond, we used vlans in that sample to split 1 bond into 2 bridges as an example. | 13:38 |
| Sam-I-Am | as long as the bridges know where to connect | 13:38 |
| svg | the used_ips example: - 172.29.236.1,172.29.236.50 are these separate ip's, or a range 1-50? | 13:40 |
| *** KLevenstein has joined #openstack-ansible | 13:40 | |
| andymccr | comma separated is a range | 13:41 |
| *** sdake_ has joined #openstack-ansible | 13:50 | |
| svg | ok, next :) what is the service net? snet it is referenced in the user config, but not in the networking docs (section 4 of the install docs) | 13:52 |
| andymccr | svg: thats rackspace specific should probably be removed from the sample conf to avoid confusion | 13:52 |
| svg | k | 13:53 |
| andymccr | im pretty sure it was just left because when we did the "de-rackspacifying" of the repo it doesnt actually mention rackspace so grep failed! | 13:53 |
| *** sdake has quit IRC | 13:54 | |
| *** Mudpuppy has joined #openstack-ansible | 13:55 | |
| svg | :) | 13:55 |
| Sam-I-Am | andymccr: wonder if it makes sense to include example files for different scenarios | 14:00 |
| *** Mudpuppy has quit IRC | 14:00 | |
| Sam-I-Am | andymccr: like $config.rax $config.aio $config.allthethings | 14:00 |
| Sam-I-Am | then in the guides we can have people copy those over as a template to start from | 14:00 |
| *** Mudpuppy has joined #openstack-ansible | 14:00 | |
| svg | andymccr: fyi, commeting the snet key yields a key error, leaving it to none, also KeyError: 'snet_netmask' | 14:15 |
| Sam-I-Am | svg: where are you seeing this? | 14:20 |
| svg | When running openstack-ansible playbooks/setup/host-setup.yml | 14:20 |
| Sam-I-Am | what version of os-ansible and whats in your user config file? | 14:21 |
| svg | at initialisation, when it's parsing the inventory | 14:21 |
| *** galstrom is now known as galstrom_zzz | 14:26 | |
| *** galstrom_zzz is now known as galstrom | 14:28 | |
| mattt | svg: you are using openstack_user_config.yml.example ? | 14:28 |
| mattt | svg: if you're commenting out snet under cidr_networks then you probably want to comment out the following also: https://github.com/stackforge/os-ansible-deployment/blob/master/etc/openstack_deploy/openstack_user_config.yml.example#L88-L97 | 14:29 |
| svg | yes based on the .example one | 14:34 |
| svg | ok, that's rpc_deploy/rpc_user_config.yml here (using v10) | 14:35 |
| svg | so is that master branch some kind of development branch? They all look like three big separate branches. | 14:36 |
| mattt | svg: eys | 14:37 |
| mattt | *yes | 14:37 |
| mattt | svg: we have a bunch of reviews in progress to get kilo into master | 14:38 |
| svg | that would be the upcoming 11 then I guess | 14:38 |
| mattt | svg: if you want to start hacking on the project, i'd recommend using master | 14:40 |
| mattt | svg: if you're looking to deploy a prod deploy, master may not be the best choice :) | 14:41 |
| svg | obviously :) which is why I'm on v10.1... | 14:41 |
| svg | I like how you guys pinpoint all software versions | 14:42 |
| mattt | svg: we will be cutting a kilo branch once the dust settles, and it could be good for you to work off that | 14:42 |
| svg | well, for now, I'm testing the setup, in competition woith colleagues who try out other alternatives (miranties fuel amongst them) | 14:42 |
| svg | I'm confident ansible will be quicker :) | 14:43 |
| svg | btw, the openstack-ansible script doesn;t take $HOME into account to look for the rpc_deploy dir | 14:43 |
| svg | for now, I think an occasional patch on the juno branch might happen for me | 14:44 |
| mattt | svg: https://review.openstack.org/#/c/167010/ | 14:44 |
| svg | ah, that would be bummer, I really liked the non-priviledged install way | 14:46 |
| svg | so far only ansible needs a system wide installation | 14:46 |
| git-harry | svg: please consider reviewing the patch so add your thoughts to the discussion | 14:48 |
| *** ccrouch-afk is now known as ccrouch | 14:52 | |
| *** devlaps has joined #openstack-ansible | 15:10 | |
| alextricity | Is the containers_all host group in the inventory not in 10.1.2? | 15:17 |
| alextricity | Ah, nvm...it's all_containers | 15:17 |
| alextricity | :) | 15:17 |
| sigmavirus24 | alextricity: containers_all_containers would be a better name, no? | 15:18 |
| sigmavirus24 | ;) | 15:18 |
| alextricity | That would make a lot more sense, given the naming of the others. haha | 15:18 |
| *** stevemar has joined #openstack-ansible | 15:34 | |
| openstackgerrit | Sudarshan Acharya proposed stackforge/os-ansible-deployment: Managing policy file with default file and user variables. https://review.openstack.org/168104 | 15:47 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Update rabbitmq_package_url https://review.openstack.org/167723 | 16:00 |
| *** galstrom is now known as galstrom_zzz | 16:05 | |
| *** Mudpuppy has quit IRC | 16:05 | |
| *** Mudpuppy has joined #openstack-ansible | 16:06 | |
| *** galstrom_zzz is now known as galstrom | 16:13 | |
| *** sacharya has joined #openstack-ansible | 16:18 | |
| *** jwagner_away is now known as jwagner | 16:28 | |
| *** galstrom is now known as galstrom_zzz | 16:35 | |
| openstackgerrit | Sudarshan Acharya proposed stackforge/os-ansible-deployment: Managing policy file with default file and user variables. https://review.openstack.org/168104 | 16:37 |
| *** jwagner is now known as jwagner_away | 16:52 | |
| *** galstrom_zzz is now known as galstrom | 16:55 | |
| *** jmccrory has joined #openstack-ansible | 17:14 | |
| *** sdake_ has quit IRC | 17:28 | |
| *** sdake has joined #openstack-ansible | 17:28 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Heat config changes for kilo https://review.openstack.org/167947 | 17:41 |
| *** bluebox_ has joined #openstack-ansible | 17:46 | |
| bluebox_ | hi, newbie here, I'd like to add some enhacments to the ansible playbooks, do I need to file a blueprint/feature request (get it approved) before I can contribute? | 17:48 |
| *** galstrom is now known as galstrom_zzz | 17:53 | |
| Apsu | bluebox_: I think that mostly depends on what kind of contribution you want to make. If it's really just minor enhancements and not really new "features", design overhaul or major refactoring, you could probably just file a launchpad bug and push a gerrit review for the code change. | 18:00 |
| bluebox_ | Okay, so I'd like to make the neutron_plugins play more modular, as right now it uses ml2 by default, | 18:01 |
| bluebox_ | and a few others that it installs | 18:01 |
| Apsu | Yes indeed. ML2 and Linux Bridge agent, which we proxy the config of through the ml2_conf.ini | 18:02 |
| bluebox_ | in essence make using any other neutron plugin easy to incorporate within the main openstack-setup | 18:02 |
| palendae | bluebox_: There was a recent PR for that, one sec | 18:03 |
| Apsu | So, that sounds like minor playbook changes to modularize what's there and additional variables/tasks/plays/templates to wire in new plugins. | 18:03 |
| palendae | bluebox_: https://review.openstack.org/#/c/166574/ | 18:03 |
| palendae | And contributing guidelines are at http://git.openstack.org/cgit/stackforge/os-ansible-deployment/tree/CONTRIBUTING.rst | 18:03 |
| bluebox_ | Apsu exactly, palendae what you just shared is an addition to the ml2 itself, I actually want to be able to use another plugin instead of ml2 for the networking | 18:06 |
| palendae | bluebox_: Gotcha | 18:06 |
| Apsu | bluebox_: Is the plugin you want to use not proxying its config through ml2? Because the ml2 layer is the standard mechanism in modern neutron these days, since post-havana. | 18:06 |
| Apsu | Its goal was to modularize the plugin API layer so all plugins can be managed and communicate more consistently :P | 18:07 |
| Apsu | I'm not sure I'd want to accept a change that skips ml2 unless there's a compelling reason | 18:07 |
| bluebox_ | the plugin I"d like to use is monolithic plugin which is in good standing in current neutron, plumgrid | 18:15 |
| palendae | Nice | 18:17 |
| bluebox_ | Also plumgrid doesn't require some of the agents being installed, since it provides the functionaity, so we would optionally like to include/exclude those agents as necessary | 18:17 |
| bluebox_ | hence making the whole neutron play more modular | 18:17 |
| palendae | bluebox_: I for one would like to see at least a blueprint for that | 18:18 |
| Apsu | bluebox_: Gotcha. I'd qualify that as an appropriate abdication of ml2 then :) | 18:19 |
| bluebox_ | Ok great, so I'll go ahead and file the blueprint. Do I need to wait for it to be approved before committing any changes? | 18:21 |
| palendae | bluebox_: I don't think we'll push back on code just because the BP isn't approved | 18:22 |
| palendae | We're actually specs now | 18:22 |
| *** hp-hj has joined #openstack-ansible | 18:23 | |
| palendae | http://git.openstack.org/cgit/stackforge/os-ansible-deployment-specs | 18:23 |
| bluebox_ | Apsu, it would be good to provide support for both monolithic & ml2 in the plays :) | 18:23 |
| *** galstrom_zzz is now known as galstrom | 18:23 | |
| cloudnull | bluebox_: no need to wait for BP approval to start getting code up for review. sometimes its nice to have a rough implementation to reference. | 18:24 |
| Apsu | bluebox_: Sure. Just want to make sure we're supporting the goals of ml2 going forward as we agree with them and remember how bad it was before them. But I realize there are still monolithic plugins that live in their own world and yet are good technology. | 18:24 |
| Apsu | cloudnull: Was just going to say that. +1 | 18:24 |
| cloudnull | :) | 18:24 |
| * cloudnull a helper | 18:24 | |
| *** hj-hp has quit IRC | 18:24 | |
| Apsu | .helperator() | 18:24 |
| sigmavirus24 | .get_helperator() | 18:26 |
| *** hp-hj has quit IRC | 18:26 | |
| Apsu | HelperFactoryGeneratorAtor | 18:27 |
| *** hp-hj has joined #openstack-ansible | 18:27 | |
| *** sdake_ has joined #openstack-ansible | 18:27 | |
| sigmavirus24 | GeneratorFactoryFactory | 18:27 |
| *** sdake has quit IRC | 18:28 | |
| *** sdake_ has quit IRC | 18:28 | |
| *** sdake_ has joined #openstack-ansible | 18:29 | |
| bluebox_ | Apsu yes I agree, the support for ml2 will remain, this change set should introduce a wider use case for the plays | 18:30 |
| Apsu | Yep. I look forward to seeing it | 18:31 |
| *** sdake has joined #openstack-ansible | 18:36 | |
| *** sdake_ has quit IRC | 18:39 | |
| *** galstrom is now known as galstrom_zzz | 18:45 | |
| *** sdake_ has joined #openstack-ansible | 18:52 | |
| *** sdake has quit IRC | 18:55 | |
| *** ccrouch is now known as ccrouch-grabfood | 19:15 | |
| *** jmccrory has quit IRC | 19:26 | |
| *** devlaps has quit IRC | 19:36 | |
| *** leakypipes has quit IRC | 19:42 | |
| *** ccrouch-grabfood is now known as ccrouch | 19:47 | |
| *** KLevenstein has quit IRC | 19:56 | |
| *** KLevenstein_ has joined #openstack-ansible | 19:56 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Fix dedicated replication network logic https://review.openstack.org/167198 | 19:58 |
| *** jmccrory has joined #openstack-ansible | 19:58 | |
| *** jmccrory1 has joined #openstack-ansible | 20:01 | |
| *** jmccrory has quit IRC | 20:03 | |
| *** jmccrory1 has quit IRC | 20:06 | |
| *** jmccrory has joined #openstack-ansible | 20:06 | |
| *** bluebox_ has quit IRC | 20:17 | |
| sigmavirus24 | BTW, while working on getting keystone up-to-date for Kilo, I noticed a couple things that I put into a bug https://bugs.launchpad.net/openstack-ansible/+bug/1437481 I added both dolphm and dstanek as I expect they'll both have opinions | 20:18 |
| openstack | Launchpad bug 1437481 in openstack-ansible "keystone httpd conf security hardening" [Undecided,New] - Assigned to Ian Cordasco (icordasc) | 20:18 |
| *** britthouser has quit IRC | 20:39 | |
| *** sacharya has quit IRC | 20:43 | |
| *** galstrom_zzz is now known as galstrom | 20:54 | |
| *** mfisch has quit IRC | 20:56 | |
| *** mfisch has joined #openstack-ansible | 20:56 | |
| *** mfisch is now known as Guest90957 | 20:57 | |
| *** KLevenstein_ has quit IRC | 21:03 | |
| openstackgerrit | Ian Cordasco proposed stackforge/os-ansible-deployment: Keystone config changes for Kilo https://review.openstack.org/167475 | 21:19 |
| *** stevemar has quit IRC | 21:42 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: WIP: Revert cinder changes https://review.openstack.org/168497 | 21:48 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Removed lxc templates that are no longer used https://review.openstack.org/168170 | 21:54 |
| *** galstrom is now known as galstrom_zzz | 22:14 | |
| *** markvoelker has quit IRC | 22:40 | |
| *** bluebox has joined #openstack-ansible | 22:50 | |
| bluebox | hey guys, the CONTRIBUTING.rst page says that "Before creating blueprint/spec an associated issue should be raised on launchpad. " Does that mean I should file a Bug first even for feature/enhancements? | 22:54 |
| * sigmavirus24 knew that would come back to bite us | 22:55 | |
| sigmavirus24 | I'm not sure it is really necessary since you discussed it here with us | 22:55 |
| sigmavirus24 | I'll let cloudnull make the final call though when he sees all of this | 22:55 |
| bluebox | interesting :), I'll keep the bp draft on the side until cloudnull gets back to us.. | 23:02 |
| *** jmccrory has quit IRC | 23:06 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Removed lxc templates that are no longer used (juno) https://review.openstack.org/168172 | 23:16 |
| *** bluebox__ has joined #openstack-ansible | 23:31 | |
| *** bluebox has quit IRC | 23:31 | |
| cloudnull | Bluebox sigmavirus24 no need to raise an issue for a feature that we discussed in channel. Seems like a but of additional work that doesn't need to happen. | 23:39 |
| cloudnull | We should fix that in the doc | 23:40 |
| *** bluebox has joined #openstack-ansible | 23:40 | |
| *** markvoelker has joined #openstack-ansible | 23:41 | |
| cloudnull | The premise of the bp fits inline with our goals anif executed modularly, as described, should have no issues being accepted. | 23:41 |
| cloudnull | So I say submit the spec and let's start reviewing it. | 23:42 |
| *** bluebox__ has quit IRC | 23:44 | |
| *** markvoelker has quit IRC | 23:46 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!