| *** sdake_ has joined #openstack-ansible | 00:00 | |
| *** sdake has quit IRC | 00:02 | |
| *** weezS has quit IRC | 00:02 | |
| *** sdake has joined #openstack-ansible | 00:09 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Fixes RabbitMQ guest user creation https://review.openstack.org/198794 | 00:09 |
|---|---|---|
| *** TheIntern has quit IRC | 00:12 | |
| *** sdake_ has quit IRC | 00:13 | |
| *** galstrom_zzz is now known as galstrom | 00:48 | |
| *** fawadkhaliq has joined #openstack-ansible | 00:53 | |
| *** fawadkhaliq has quit IRC | 00:56 | |
| *** daneyon has joined #openstack-ansible | 01:06 | |
| *** annashen has joined #openstack-ansible | 01:17 | |
| *** annashen has quit IRC | 01:22 | |
| *** JRobinson__ is now known as JRobinson__afk | 01:32 | |
| *** JRobinson__afk is now known as JRobinson__ | 01:44 | |
| *** galstrom is now known as galstrom_zzz | 02:06 | |
| *** daneyon has quit IRC | 02:08 | |
| *** daneyon has joined #openstack-ansible | 02:08 | |
| *** annashen has joined #openstack-ansible | 02:18 | |
| *** annashen has quit IRC | 02:22 | |
| *** weezS has joined #openstack-ansible | 02:25 | |
| *** weezS has joined #openstack-ansible | 02:26 | |
| *** daneyon has quit IRC | 02:45 | |
| *** annashen has joined #openstack-ansible | 03:19 | |
| *** annashen has quit IRC | 03:23 | |
| *** weezS has quit IRC | 03:54 | |
| *** weezS has joined #openstack-ansible | 03:56 | |
| *** JRobinson__ is now known as JRobinson__afk | 04:07 | |
| *** annashen has joined #openstack-ansible | 04:19 | |
| *** annashen has quit IRC | 04:24 | |
| *** weezS has quit IRC | 04:30 | |
| *** shausy has joined #openstack-ansible | 04:45 | |
| *** JRobinson__afk is now known as JRobinson__ | 04:53 | |
| openstackgerrit | Miguel Grinberg proposed stackforge/os-ansible-deployment: [WIP] SSL support for haproxy https://review.openstack.org/198957 | 04:58 |
| openstackgerrit | Miguel Grinberg proposed stackforge/os-ansible-deployment: [WIP] SSL support for haproxy https://review.openstack.org/198957 | 05:00 |
| *** jwagner is now known as jwagner_away | 05:07 | |
| *** annashen has joined #openstack-ansible | 05:21 | |
| *** annashen has quit IRC | 05:25 | |
| *** ig0r_ has joined #openstack-ansible | 05:52 | |
| *** ig0r__ has quit IRC | 05:55 | |
| *** annashen has joined #openstack-ansible | 06:21 | |
| *** annashen has quit IRC | 06:26 | |
| *** radek__ has joined #openstack-ansible | 06:28 | |
| *** shausy has quit IRC | 07:05 | |
| *** shausy has joined #openstack-ansible | 07:06 | |
| *** vdo has joined #openstack-ansible | 07:17 | |
| *** annashen has joined #openstack-ansible | 07:22 | |
| *** annashen has quit IRC | 07:27 | |
| *** git-harry has quit IRC | 07:56 | |
| *** git-harry has joined #openstack-ansible | 07:56 | |
| *** JRobinson__ has quit IRC | 08:09 | |
| *** annashen has joined #openstack-ansible | 08:23 | |
| *** annashen has quit IRC | 08:28 | |
| *** shausy has quit IRC | 09:03 | |
| *** shausy has joined #openstack-ansible | 09:03 | |
| *** annashen has joined #openstack-ansible | 09:24 | |
| *** annashen has quit IRC | 09:29 | |
| *** alexisc has joined #openstack-ansible | 09:35 | |
| *** alexisc has left #openstack-ansible | 09:37 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Fix scripts/run-upgrade.sh https://review.openstack.org/198311 | 09:57 |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: Set default inventory file for inventory-manage.py https://review.openstack.org/199040 | 10:19 |
| *** annashen has joined #openstack-ansible | 10:25 | |
| *** annashen has quit IRC | 10:30 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: Set default inventory file for inventory-manage.py https://review.openstack.org/199040 | 10:38 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Fix scripts/run-upgrade.sh https://review.openstack.org/198311 | 10:41 |
| openstackgerrit | Darren Birkett proposed stackforge/os-ansible-deployment: Added in keystone reserved port https://review.openstack.org/196702 | 11:05 |
| *** annashen has joined #openstack-ansible | 11:26 | |
| *** annashen has quit IRC | 11:30 | |
| *** markvoelker has quit IRC | 11:59 | |
| *** markvoelker has joined #openstack-ansible | 11:59 | |
| *** jaypipes has joined #openstack-ansible | 12:02 | |
| evrardjp | hello everyone | 12:03 |
| *** annashen has joined #openstack-ansible | 12:27 | |
| *** tlian has joined #openstack-ansible | 12:30 | |
| *** annashen has quit IRC | 12:31 | |
| *** markvoelker has quit IRC | 12:31 | |
| openstackgerrit | Hugh Saunders proposed stackforge/os-ansible-deployment: Upgrade the Keystone library to use v3 https://review.openstack.org/196943 | 12:34 |
| odyssey4me | howdy evrardjp | 12:39 |
| *** markvoelker has joined #openstack-ansible | 12:40 | |
| evrardjp | svg: FYI ceph-ansible from leseb isn't only for deploying ceph, it also handles creating the keys for openstack (cf ceph-mon role) | 12:43 |
| evrardjp | hello odyssey4me! nice weather in UK? | 12:49 |
| svg | evrardjp: so are you saing i should re-use that role to do the key providing? | 12:58 |
| evrardjp | I just mean there will be overlapping if we do the key provisioning in osad | 13:03 |
| evrardjp | (for the future) | 13:03 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Limit swift_vars / swift_proxy_vars https://review.openstack.org/199094 | 13:22 |
| *** annashen has joined #openstack-ansible | 13:27 | |
| *** annashen has quit IRC | 13:32 | |
| *** KLevenstein has joined #openstack-ansible | 13:39 | |
| *** ccrouch has joined #openstack-ansible | 13:42 | |
| *** ccrouch has left #openstack-ansible | 13:45 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Add Keystone SSL key/cert generation & distribution https://review.openstack.org/194474 | 13:49 |
| *** TheIntern has joined #openstack-ansible | 13:50 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Limit swift_proxy_vars to swift_proxy container https://review.openstack.org/199094 | 13:52 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Limit swift_proxy_vars to swift_proxy container https://review.openstack.org/199094 | 13:53 |
| *** shausy has quit IRC | 13:54 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:57 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Keystone SSL cert/key distribution and configuration https://review.openstack.org/194474 | 14:00 |
| odyssey4me | svg it may just be worth going with any minor changes you've made and it can be iterated later when others have a chance to work with it | 14:01 |
| *** Mudpuppy has joined #openstack-ansible | 14:01 | |
| odyssey4me | evrardjp it's warm, humid but overcast... :p | 14:02 |
| *** ccrouch has joined #openstack-ansible | 14:02 | |
| cloudnull | good morning | 14:03 |
| *** sdake has quit IRC | 14:14 | |
| odyssey4me | o/ cloudnull | 14:14 |
| *** sdake has joined #openstack-ansible | 14:14 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Updated juno for new dev work https://review.openstack.org/199119 | 14:16 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Set default inventory file for inventory-manage.py https://review.openstack.org/199040 | 14:17 |
| sigmavirus24 | hughsaunders: from what I udnerstood from the bug that mattt linked yesterday, as long as we use auth_plugin for Nova's neturonclient config, it will support keystone v3 | 14:20 |
| *** Bjoern_ has joined #openstack-ansible | 14:22 | |
| cloudnull | 11.0.4 is official out the door | 14:23 |
| cloudnull | if you have a hold on an item for inclusion into 11.1.0 please revise / update them | 14:23 |
| cloudnull | IE https://review.openstack.org/#/q/status:open+project:stackforge/os-ansible-deployment+branch:kilo,n,z | 14:23 |
| b3rnard0 | thanks cloudnull | 14:23 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Updated kilo for new dev work https://review.openstack.org/199124 | 14:23 |
| hughsaunders | sigmavirus24: ok, will try. I just looked at the imports and only saw v2 | 14:25 |
| sigmavirus24 | The auth_plugin in keystoneclient is what will handle v3 | 14:26 |
| hughsaunders | ok | 14:26 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Updated master for new dev work https://review.openstack.org/199126 | 14:26 |
| evrardjp | if it's official, congrats everyone! | 14:26 |
| cloudnull | thanks evrardjp :) | 14:27 |
| cloudnull | can we get someone to review https://review.openstack.org/#/c/173067/ | 14:27 |
| cloudnull | ceilometer implementation . | 14:28 |
| *** annashen has joined #openstack-ansible | 14:28 | |
| *** toddnni_ has joined #openstack-ansible | 14:29 | |
| *** toddnni has quit IRC | 14:29 | |
| *** toddnni_ is now known as toddnni | 14:29 | |
| *** weezS has joined #openstack-ansible | 14:31 | |
| openstackgerrit | Miguel Alejandro Cantu proposed stackforge/os-ansible-deployment: Implement Ceilometer https://review.openstack.org/173067 | 14:32 |
| *** annashen has quit IRC | 14:33 | |
| *** jwagner_away is now known as jwagner | 14:33 | |
| *** alextricity has joined #openstack-ansible | 14:34 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Keystone SSL cert/key distribution and configuration https://review.openstack.org/194474 | 14:38 |
| odyssey4me | cloudnull any chance you can take a look at https://review.openstack.org/194474 as an updated review based on miguelgrinberg's earlier work - this definitely works, and seems like a much better solution to me based on previous review comments | 14:43 |
| * cloudnull looking now | 14:50 | |
| d34dh0r53 | how does the idea of breaking user_secrets.yml into a secrets.d directory sound? cray-cray? | 15:10 |
| d34dh0r53 | probably needs to be done with user_group_vars.yml as well | 15:11 |
| mattt | d34dh0r53: make sure you keep scripts/run-upgrade.sh updated if you do that :) | 15:11 |
| mattt | d34dh0r53: (we're still dealing w/ the env.d changes that went into master/kilo) | 15:11 |
| d34dh0r53 | mattt: yeah | 15:11 |
| hughsaunders | d34dh0r53: why? | 15:12 |
| odyssey4me | d34dh0r53 if we had a conf.d type thingy for user_vars then we wouldn't need something else for the secrets - we can just use the same folder? | 15:12 |
| d34dh0r53 | well, I'm looking at the ceilometer patch and we have conf.d and env.d where you can easily drop small files, then you have to go an edit the monolithic user_secrets and user_group_vars, which kind of defeats the purpose of the *.d directories | 15:13 |
| odyssey4me | d34dh0r53 if we could do away with using the command line overriding and have more predictable inventory-based overrides then I think it'd make it easier to understand precedence | 15:13 |
| odyssey4me | quite honestly I'd prefer the stuff just to go into conf.d | 15:13 |
| d34dh0r53 | not just for precedence but for ease of adding optional features | 15:13 |
| d34dh0r53 | yeah, that would work | 15:14 |
| odyssey4me | d34dh0r53 it'd be useful to perhaps then augment the dynamic inventory to output warnings about var duplications | 15:15 |
| d34dh0r53 | yeah | 15:15 |
| d34dh0r53 | very useful | 15:15 |
| odyssey4me | d34dh0r53 git-harry has been working his way through variable-related weirdness and has recently commented about all this not being great\ | 15:16 |
| d34dh0r53 | odyssey4me: I can see that, it's confusing at best | 15:16 |
| alextricity | d34dh0r53: Do we still *require* a logging host? | 15:19 |
| odyssey4me | hughsaunders sigmavirus24 can it be this easy? http://adam.younglogic.com/2015/03/convince-nova-to-use-the-v3-version-of-the-api/ | 15:19 |
| d34dh0r53 | RAX does, yes | 15:19 |
| odyssey4me | alextricity require, not really - but it's expected that most environments will want at least the centralised rsyslog | 15:20 |
| alextricity | Thanks! | 15:21 |
| openstackgerrit | Darren Birkett proposed stackforge/os-ansible-deployment: Set default inventory file for inventory-manage.py https://review.openstack.org/199149 | 15:24 |
| sigmavirus24 | odyssey4me: maybe but the only problem I've been seeing with nova is its interaction with neutron | 15:25 |
| odyssey4me | hughsaunders sigmavirus24 also, almost every play/role is set to use the internalURL as the default endpoint - in your tests is that endpoint v3? | 15:25 |
| sigmavirus24 | odyssey4me: yes | 15:26 |
| odyssey4me | hughsaunders sigmavirus24 here's another resource worth working through: http://adam.younglogic.com/2015/05/rdo-v3-only/ | 15:26 |
| *** annashen has joined #openstack-ansible | 15:29 | |
| sigmavirus24 | odyssey4me: do we want it to be v3 only though? | 15:29 |
| odyssey4me | sigmavirus24 essentially I think that we should configure all services to use v3, but still leave the v2 endpoint available for end-users | 15:30 |
| odyssey4me | we need horizon to use v3, and federation needs v3 - and v3's been around for ages... it's time to move :) | 15:30 |
| odyssey4me | horizon's websso for federation will not work without v3 | 15:32 |
| *** annashen has quit IRC | 15:34 | |
| odyssey4me | sigmavirus24 shall I prepare a separate patch related to allow 'insecure' comms to keystone (for self-signed certs and all that)? | 15:35 |
| sigmavirus24 | odyssey4me: maybe | 15:45 |
| sigmavirus24 | I'm still trying to figure out why this is broken | 15:45 |
| sigmavirus24 | because nothing can ever be simple in openstack | 15:46 |
| odyssey4me | sigmavirus24 I'm going to go ahead with that - I think it deserves a separate patch. | 15:46 |
| sigmavirus24 | I hadn't forgotten about it | 15:46 |
| sigmavirus24 | Just wanted to fix this crap first before piling more on =P | 15:46 |
| odyssey4me | you can add the option to the module inside the module patch, but I'll ensure that the services and CLI are configured to work properly | 15:47 |
| *** sdake_ has joined #openstack-ansible | 15:48 | |
| *** sdake has quit IRC | 15:48 | |
| openstackgerrit | Andy McCrae proposed stackforge/os-ansible-deployment-specs: Multi-region swift https://review.openstack.org/198322 | 15:50 |
| *** daneyon has joined #openstack-ansible | 16:00 | |
| b3rnard0 | bug triage? | 16:02 |
| cloudnull | bug triage time cloudnull, mattt, andymccr, d34dh0r53, hughsaunders, b3rnard0, palendae, Sam-I-Am, odyssey4me, serverascode, rromans, mancdaz, dolphm, _shaps_, BjoernT, claco, echiu, dstanek, jwagner, ayoung | 16:03 |
| dstanek | o/ | 16:04 |
| odyssey4me | o/ | 16:04 |
| palendae | p/ | 16:04 |
| cloudnull | first up https://bugs.launchpad.net/openstack-ansible/+bug/1472295 | 16:05 |
| openstack | Launchpad bug 1472295 in openstack-ansible "Juno: cinder and glance client have endpoint selection issues" [Undecided,New] | 16:05 |
| rromans | . | 16:06 |
| cloudnull | based on the issue it looks like the clients have once again broken the ability to select endpoint types. | 16:06 |
| cloudnull | key word, once again ... | 16:06 |
| sigmavirus24 | o/ | 16:07 |
| sigmavirus24 | hm | 16:08 |
| cloudnull | oh wait maybe not. Bjoern_ you around ? | 16:08 |
| Bjoern_ | yes | 16:08 |
| sigmavirus24 | Bjoern_: seems to be indicating that this is an issue with f5's | 16:08 |
| *** Bjoern_ is now known as BjoernT | 16:08 | |
| cloudnull | ^ that | 16:08 |
| BjoernT | we did try setting x-forwarded-proto to https and the keystone middleware does not seem to pick this up | 16:08 |
| sigmavirus24 | I suspect <ip> is the internalURL | 16:08 |
| BjoernT | only once I changed public_endpoint it did fix this issue. | 16:09 |
| sigmavirus24 | And so when the client is made to that we can't see it | 16:09 |
| sigmavirus24 | BjoernT: public endpoint in the keystone catalog? | 16:09 |
| BjoernT | interestingly this issue is not present on older glance/cinder clients | 16:09 |
| BjoernT | public endpoint was changed to https correctly | 16:09 |
| *** alextricity has quit IRC | 16:09 | |
| sigmavirus24 | That is interesting | 16:09 |
| BjoernT | as I said the glance client didn't even download the catalog and just tried http://public_vip | 16:10 |
| *** alextricity has joined #openstack-ansible | 16:10 | |
| BjoernT | and it fixed after we changed linkes: href | 16:10 |
| BjoernT | public_endpoint is set to None by default | 16:10 |
| odyssey4me | Perhaps I have the wrong end of the stick here? Do you mean 'public_endpoint' in a conf file or in your keystone endpoint-list? | 16:11 |
| sigmavirus24 | odyssey4me: yeah that's where I'm confused as well | 16:12 |
| cloudnull | its the keystone.conf setting that we used to set by default in icehouse i supose. | 16:12 |
| cloudnull | https://github.com/stackforge/os-ansible-deployment/blob/juno/rpc_deployment/roles/keystone_common/templates/keystone.conf.j2#L10-L12 | 16:12 |
| cloudnull | ^ BjoernT you set that right ? | 16:12 |
| BjoernT | None is supposedly correct, taking the ip from the original request but I guess there is a bug in keystone middleware.To fix it for now i used the keystone_public_endpoint setting | 16:13 |
| BjoernT | cloudnull: correcy | 16:14 |
| BjoernT | correct | 16:14 |
| sigmavirus24 | dolphm: dstanek ^ | 16:15 |
| cloudnull | so is this an issue? because it has the ability to be set dstanek, lbragstad, dstanek ^ | 16:15 |
| cloudnull | sigmavirus24: :) | 16:15 |
| *** fawadkhaliq has joined #openstack-ansible | 16:15 | |
| dstanek | sigmavirus24: sorry, was trouble shooting something. is the question wether or not setting public_endpoint to None is valid? | 16:16 |
| sigmavirus24 | dstanek: the question is whether or not this is a bug in keystonemiddleware I believe | 16:16 |
| odyssey4me | BjoernT is this only in Juno? have you tried replicating it in kilo? | 16:17 |
| BjoernT | same issue for kilo | 16:17 |
| BjoernT | since we added the https fix in kilo we might set public_endpoint automatically once the protocol is https | 16:18 |
| BjoernT | or we need to expose this issue in the release notes once the public endpoints are using https | 16:19 |
| dstanek | sigmavirus24: not sure where the bug is there; the returned tokens link should be secure right? | 16:20 |
| *** yaya has joined #openstack-ansible | 16:20 | |
| odyssey4me | so yeah, I'm still trying to understand - is the issue that the wrong endpoint address is exposed to the client? | 16:21 |
| odyssey4me | ie the internalURL instead of the publicURL? | 16:21 |
| BjoernT | I think the issues is as simple as public_endpoint = None does not account for the protocol https vs http. Otherwise everything is correct. It does not return the internal endpoint | 16:23 |
| cloudnull | i think its that the ssl termination is happening on an LB and by setting the keystone_public_endpoint within config its able to redirect the clients back to the lb otherwise its failing to get the service catalog ? | 16:23 |
| cloudnull | otherwise im confused. or not well caffeinated | 16:23 |
| BjoernT | since older glance and cinder clients just ignored the href from the links response header, the issue did happen since it was using OS_AUTH_URL and then downloading the service catalog | 16:24 |
| BjoernT | at some point we might decide if we go back to 10.1.1 versions for cinder and glance | 16:24 |
| BjoernT | or set public_endpoint in keystone.conf correctly | 16:24 |
| odyssey4me | BjoernT it would seem to me that public_endpoint was built for this exact purpose | 16:25 |
| odyssey4me | if you're front-ending your keystone with something that keystone doesn't know about, then use that setting | 16:25 |
| dstanek | if you don't set the public_endpoint it tries to figure it out: http://git.openstack.org/cgit/openstack/keystone/tree/keystone/common/wsgi.py#n806 | 16:25 |
| odyssey4me | otherwise it'll return values from the service catalogue, which it seems you don't want | 16:25 |
| BjoernT | Yeah, but it seems it going to be deprecated in favor of the service catalog | 16:25 |
| odyssey4me | there's no deprecation notice: http://docs.openstack.org/kilo/config-reference/content/section_keystone.conf.html | 16:26 |
| BjoernT | either way we have to look how to deal with this issue and maybe filing a upstream bug | 16:26 |
| odyssey4me | it looks like it's behaving as designed to me, unless I'm missing something? | 16:27 |
| dstanek | BjoernT: if you properly set the value does it work? | 16:27 |
| sigmavirus24 | dstanek: yes | 16:27 |
| BjoernT | yes | 16:27 |
| BjoernT | it does | 16:27 |
| BjoernT | the href is changed inside the links | 16:27 |
| BjoernT | response | 16:27 |
| odyssey4me | the issue is that the old clients had bad habits :p | 16:27 |
| dstanek | BjoernT: i don't think it's really an upstream bug, but we could probably be smarter about how we guess | 16:28 |
| *** annashen has joined #openstack-ansible | 16:28 | |
| sigmavirus24 | dstanek: don't guess, just know. When is Keystone landing telepathy? | 16:29 |
| cloudnull | ^ lol | 16:29 |
| * sigmavirus24 is not helping | 16:29 | |
| * sigmavirus24 will shut up | 16:29 | |
| palendae | sigmavirus24: Right after Neutron does | 16:29 |
| d34dh0r53 | haha | 16:29 |
| d34dh0r53 | apt-get install libtelepathy | 16:30 |
| sigmavirus24 | palendae: neutron already does what you expect it to (break) | 16:30 |
| dstanek | sigmavirus24: next cycle. this cycle we are working on making it access your bank account | 16:30 |
| d34dh0r53 | lol | 16:31 |
| sigmavirus24 | dstanek: sounds fair | 16:31 |
| sigmavirus24 | dstanek: so keystone is trying to replace ceilometer? Thank the $DEITIES | 16:31 |
| cloudnull | ok so idk if this is an issue ? | 16:32 |
| cloudnull | it seems that theres a setting that can be used to fix the problem | 16:32 |
| cloudnull | and that the problem is unique to the environment . | 16:32 |
| sigmavirus24 | Yeah, do we provide a variable to set that config option? | 16:33 |
| cloudnull | https://github.com/stackforge/os-ansible-deployment/blob/juno/rpc_deployment/roles/keystone_common/templates/keystone.conf.j2#L10-L12 | 16:33 |
| odyssey4me | it seems to me that there is already a setting to provide the requested behaviour - the only thing that may be worth doing is a upgrade/release note to indicate that behaviour has changed between version | 16:33 |
| sigmavirus24 | If so there's a work around. I'm curious what differs between those versions of the client | 16:33 |
| odyssey4me | or keystonemiddleware, or whatever | 16:33 |
| sigmavirus24 | yeah | 16:33 |
| sigmavirus24 | I wonder if not setting it is equivalent to using None | 16:34 |
| sigmavirus24 | but that's besides the point | 16:34 |
| sigmavirus24 | I've added the clients as being affected | 16:35 |
| sigmavirus24 | I think this is something they should see and check for in current versions | 16:35 |
| *** sdake_ is now known as sdae | 16:35 | |
| sigmavirus24 | It's interesting to say the least | 16:35 |
| cloudnull | ok marked incomplete for now . | 16:36 |
| cloudnull | next https://bugs.launchpad.net/openstack-ansible/+bug/1472038 | 16:36 |
| openstack | Launchpad bug 1472038 in openstack-ansible "Sort dynamic inventory files in juno/icehouse" [Undecided,New] | 16:36 |
| *** yaya has quit IRC | 16:36 | |
| cloudnull | this is a simple fix, in kilo we sort the inventory in juno we didnt. we should do that. | 16:36 |
| cloudnull | TheIntern: ^ low hanging fruit | 16:37 |
| TheIntern | alrighty | 16:37 |
| TheIntern | How should they be sorted? | 16:38 |
| sigmavirus24 | reverse chronologically | 16:38 |
| TheIntern | will do | 16:39 |
| cloudnull | basically backport these lines into juno https://github.com/stackforge/os-ansible-deployment/blob/master/playbooks/inventory/dynamic_inventory.py#L960-L965 | 16:39 |
| *** sdae is now known as sdake | 16:39 | |
| sigmavirus24 | cloudnull: ruins all my fun | 16:40 |
| cloudnull | lol | 16:41 |
| cloudnull | next https://bugs.launchpad.net/openstack-ansible/+bug/1471926 | 16:42 |
| openstack | Launchpad bug 1471926 in openstack-ansible "Add rpc tunables to nova.conf" [Undecided,New] | 16:42 |
| cloudnull | these should be simple enough too. | 16:42 |
| palendae | That has a complement in https://github.com/rcbops/rpc-openstack/issues/220 | 16:42 |
| cloudnull | however in juno the vars are different than kilo with the new oslo messaging section . | 16:43 |
| cloudnull | next https://bugs.launchpad.net/openstack-ansible/+bug/1471746 | 16:45 |
| openstack | Launchpad bug 1471746 in openstack-ansible "Use of groups in roles" [Undecided,New] | 16:45 |
| cloudnull | git-harry: svg: ^ | 16:45 |
| cloudnull | idk agree that we are using an anti pattern , but i do agree that we can clean up some of the roles and the usage of groups and that we should document the groups that are required. | 16:46 |
| odyssey4me | it does seem that we're diverging from a standard use of roles, but other than breaking roles up I have yet to find an alternative way of doing some of the things we're doing | 16:48 |
| sigmavirus24 | I think if gregdek and other ansible folk could weigh in on this, that'd be great too | 16:49 |
| odyssey4me | for now perhaps doing the documenting thing is a good idea, but ultimately we probably need a re-look at how we do variables, roles, playbooks, etc and try to bring it more into line with better practises | 16:49 |
| sigmavirus24 | They have endorsed us as the way to deploy openstack with ansible, right? So they must not find it particularly objectionable that we're doing things this way | 16:49 |
| sigmavirus24 | That doesnt' mean we couldn't be doing stuff better, just that it must not be too terrible | 16:50 |
| odyssey4me | there are some real issues with using roles at the moment which make them very inflexible | 16:50 |
| cloudnull | breaking up the roles in more individual consumable parts may be a good way to do it, but then again we had that in juno; it was terribly inefficient and added complexity that need not be there. | 16:50 |
| odyssey4me | yep | 16:50 |
| palendae | IMO the big ones that need to be stand alone are Keystone and Swift | 16:50 |
| palendae | Other than that I'm not sure it's helping | 16:50 |
| odyssey4me | it's worth some thinking - we may be able to find better ways of doing things if we try | 16:51 |
| palendae | Sure | 16:51 |
| cloudnull | odyssey4me: ++ i think we need to take a good long look at inventory and how to do that better. | 16:51 |
| odyssey4me | eg: neutron & neutron-agents could split | 16:51 |
| sigmavirus24 | yeah, I think this is a good long-term goal for the organization of our inventory | 16:52 |
| odyssey4me | nova could split into some parts which are more role focused - ie those that go on controllers, and those that go on compute nodes | 16:52 |
| sigmavirus24 | We should probably focus this on the bug or ML and then turn that into a spec | 16:53 |
| odyssey4me | but these sorts of splits should not be done without a good look at how better to do the inventory, vars, etc | 16:53 |
| cloudnull | sigmavirus24: +1 | 16:53 |
| odyssey4me | +1 | 16:53 |
| sigmavirus24 | (and then turn the spec into an implementation of course) | 16:53 |
| cloudnull | what I want to avoid is the role sprawl we had before . | 16:53 |
| sigmavirus24 | Should we say that the M cycle would be best for that kind of work? | 16:53 |
| cloudnull | +1 | 16:53 |
| * sigmavirus24 doesn't think we'll get that into L | 16:53 | |
| odyssey4me | one thing I will say is that our use of groups in roles makes it harder for people who want to use the roles outside of OSAD's general framework - the roles are not as portable | 16:54 |
| odyssey4me | cloudnull sigmavirus24 agreed | 16:54 |
| prometheanfire | cloudnull: I have some of the compute/service split stuff done for nova on our side | 16:55 |
| prometheanfire | gentoo that is | 16:55 |
| cloudnull | so im marking this incomplete at this time too and we'll carry on within the ML. | 16:55 |
| palendae | odyssey4me: Out of curiosity - have you found people wanting that right now? | 16:55 |
| odyssey4me | palendae yep, ayoung specifically asked with regards to the federation work | 16:56 |
| palendae | Ok | 16:56 |
| palendae | I think Keystone's easier since it's at the root of the dependency tree (inside of openstack itself, anyway) | 16:56 |
| odyssey4me | I do think that increasing the portability and having the roles on ansible galaxy will increase the exposure and the re-use - it's like opening the door to wider use-cases | 16:57 |
| sigmavirus24 | odyssey4me: yes but I'm not sure that's exactly the goal of the project necessarily | 16:57 |
| sigmavirus24 | Anyway that's a separate discussion altogether | 16:57 |
| palendae | Yeah, sorry | 16:57 |
| odyssey4me | sigmavirus24 sure | 16:57 |
| sigmavirus24 | 1.5 min left | 16:58 |
| cloudnull | yup | 16:58 |
| cloudnull | anything we want to cover within the open issues ? | 16:58 |
| cloudnull | ok were done here . | 16:59 |
| cloudnull | thanks everyone! | 16:59 |
| odyssey4me | miguelgrinberg are you around yet? | 17:00 |
| miguelgrinberg | yep, I'm here | 17:01 |
| odyssey4me | I'm done with https://review.openstack.org/194474 - take a look there. I think the way I've done the haproxy configuration there may be of interest to you. :) | 17:01 |
| miguelgrinberg | odyssey4me: nice, so now we need a third option for the SSL terminated at haproxy | 17:03 |
| odyssey4me | miguelgrinberg yeah, if we can also have that option it'd be great | 17:03 |
| odyssey4me | I'm working on a patch that will also ensure that services can be configured to work even when the SSL certificates are self-signed. | 17:04 |
| miguelgrinberg | Yes, I think I'll have the haproxy SSL done soon. Had to go look at a problem from support with the horizon solutions catalog, but I'll get back to this later today and try to finish it. | 17:04 |
| odyssey4me | Great! sigmavirus24 and hughsaunders are still stuck trying to get the keystone module to work properly against the v3 API... something's holding it up | 17:06 |
| odyssey4me | cloudnull it would seem the the get-pip pytohn script tries to grab for rpc-repo? | 17:10 |
| odyssey4me | and for some reason http://rpc-repo.rackspace.com/os-releases/master/pip-7.1.0-py2.py3-none-any.whl appears to be missing - I'm getting master build fails | 17:10 |
| cloudnull | thats likely because im rebuilding maste.r | 17:11 |
| *** yaya has joined #openstack-ansible | 17:11 | |
| odyssey4me | I thought that if you build a local repo then all pip interaction was local | 17:11 |
| odyssey4me | ah, that makes sense :p | 17:11 |
| cloudnull | im rebuilding it with the changes coming in https://review.openstack.org/#/c/199126/ | 17:12 |
| *** yaya has quit IRC | 17:12 | |
| odyssey4me | ok, I think that's my cue to relocate - I'll be back on later to continue with this patch | 17:12 |
| cloudnull | scotch time ! | 17:13 |
| odyssey4me | cloudnull heh, good idea :) | 17:13 |
| cloudnull | ;) | 17:13 |
| cloudnull | master is rebuilt | 17:15 |
| * cloudnull lunching | 17:16 | |
| *** TheIntern has quit IRC | 17:25 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Move Cinder-volumes to "on metal" https://review.openstack.org/195181 | 17:44 |
| *** yaya has joined #openstack-ansible | 17:54 | |
| *** TheIntern has joined #openstack-ansible | 18:01 | |
| *** annashen has quit IRC | 18:03 | |
| *** annashen has joined #openstack-ansible | 18:05 | |
| *** cbaesema has joined #openstack-ansible | 18:09 | |
| cloudnull | sigmavirus24: https://review.openstack.org/#/c/195226/ can you release this change ? | 18:17 |
| cloudnull | odyssey4me: https://review.openstack.org/#/c/195397/ can you release this change ? | 18:18 |
| *** annashen has quit IRC | 18:23 | |
| *** annashen has joined #openstack-ansible | 18:24 | |
| *** andymccr has quit IRC | 18:25 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Set default inventory file for inventory-manage.py https://review.openstack.org/199149 | 18:26 |
| *** annashen has quit IRC | 18:28 | |
| *** sdake has quit IRC | 18:36 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Moved user_group_vars to defaults https://review.openstack.org/199216 | 18:37 |
| *** KLevenstein_ has joined #openstack-ansible | 18:48 | |
| *** KLevenstein has quit IRC | 18:51 | |
| *** KLevenstein_ is now known as KLevenstein | 18:51 | |
| *** sdake has joined #openstack-ansible | 18:52 | |
| *** sdake_ has joined #openstack-ansible | 18:56 | |
| odyssey4me | cloudnull done :) | 18:58 |
| *** galstrom_zzz is now known as galstrom | 18:58 | |
| cloudnull | tyvm ! | 18:58 |
| cloudnull | only scotch now | 18:59 |
| Sam-I-Am | mmm scotcg | 18:59 |
| Sam-I-Am | h | 18:59 |
| *** sdake has quit IRC | 19:00 | |
| *** annashen has joined #openstack-ansible | 19:06 | |
| *** yaya has quit IRC | 19:06 | |
| sigmavirus24 | sssh no reviews. only scotch now | 19:15 |
| openstackgerrit | David Alfano proposed stackforge/os-ansible-deployment: Sort dynamic inventory files in juno https://review.openstack.org/199229 | 19:15 |
| *** BjoernT has quit IRC | 19:19 | |
| *** sacharya has joined #openstack-ansible | 19:19 | |
| *** yaya has joined #openstack-ansible | 19:20 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Implement Ceilometer https://review.openstack.org/173067 | 19:31 |
| *** mordred has joined #openstack-ansible | 19:35 | |
| odyssey4me | sigmavirus24 you around? | 19:55 |
| sigmavirus24 | odyssey4me: if having my head inside nova is around, yes :D | 19:56 |
| odyssey4me | your name is all over just about every google result on this, so I may as well shortcut my way to the answer | 19:56 |
| sigmavirus24 | lol | 19:56 |
| sigmavirus24 | that sounds bad | 19:56 |
| odyssey4me | # openstack --insecure endpoint list | 19:56 |
| odyssey4me | WARNING: keystoneclient.auth.identity.generic.base Discovering versions from the identity service failed when creating the password plugin. Attempting to determine version from URL. | 19:56 |
| odyssey4me | ERROR: openstack SSL exception connecting to https://172.29.236.100:5000/v2.0/tokens: [Errno 8] _ssl.c:510: EOF occurred in violation of protocol | 19:56 |
| odyssey4me | that's the haproxy endpoint | 19:56 |
| odyssey4me | then I switch directly to the server: | 19:56 |
| odyssey4me | and it happily works with the same command | 19:57 |
| sigmavirus24 | hm | 19:57 |
| odyssey4me | something tells me that urllib and requests are not the culprit | 19:57 |
| sigmavirus24 | My guess: Something to do with HAProxy and requests using keep-alive | 19:57 |
| sigmavirus24 | I have two ideas | 19:57 |
| sigmavirus24 | Oh poo | 19:57 |
| sigmavirus24 | openstackclient | 19:58 |
| sigmavirus24 | hm | 19:58 |
| odyssey4me | keystone client didn't work either | 19:58 |
| odyssey4me | I can use that if you like? | 19:58 |
| sigmavirus24 | No, I'm just used to telling people how to use requests | 19:58 |
| sigmavirus24 | =P | 19:58 |
| sigmavirus24 | I'm not sure how to attack this when it's behind a cli | 19:58 |
| odyssey4me | heh, I think I see the problem - the haproxy config may be whacked - hang a sec | 19:59 |
| Sam-I-Am | its always haproxy | 19:59 |
| sigmavirus24 | Sam-I-Am: yeah proxies like that are always a pain when improperly configured | 19:59 |
| sigmavirus24 | I would look at keep-alive settings | 19:59 |
| odyssey4me | but the reason I actually wanted to contact you was to ask about that environment variable that you thought might remove the need to use --insecure? | 19:59 |
| odyssey4me | what is it | 19:59 |
| sigmavirus24 | CURL_CA_BUNDLE/REQUESTS_CA_BUNDLE | 19:59 |
| sigmavirus24 | point that to a pem file and you're done | 19:59 |
| sigmavirus24 | either one works | 19:59 |
| odyssey4me | bugger - I found that one, but hoped it wasn't that - let me try it anyway | 20:00 |
| sigmavirus24 | CURL_CA_BUNDLE has the benefit of also helping curl | 20:00 |
| odyssey4me | sigmavirus24 what do you know - it bloody works | 20:03 |
| odyssey4me | and I found a bug in my Keystone SSL patch | 20:03 |
| sigmavirus24 | odyssey4me: yw | 20:03 |
| odyssey4me | sigmavirus24 no really, thank you :) | 20:05 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Keystone SSL cert/key distribution and configuration https://review.openstack.org/194474 | 20:07 |
| sigmavirus24 | odyssey4me: I wasn't being sarcastic :D | 20:09 |
| *** fawadkhaliq has quit IRC | 20:15 | |
| odyssey4me | sigmavirus24 meh, although your env variables are nice that'll mean shipping the offending cert to all hosts and containers... aint nobody got time for that :/ | 20:18 |
| odyssey4me | it'll be easier to alias the commands for now, until we have a better certificate solution | 20:18 |
| *** daneyon has quit IRC | 20:26 | |
| openstackgerrit | Tom Jose Kalapura proposed stackforge/os-ansible-deployment: Introduce logrotate in all Openstack containers. https://review.openstack.org/199273 | 20:53 |
| *** jwagner is now known as jwagner_away | 20:55 | |
| *** daneyon has joined #openstack-ansible | 21:10 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Enable all services to use Keystone 'insecurely' https://review.openstack.org/199307 | 21:19 |
| *** sdake_ is now known as sdake | 21:20 | |
| odyssey4me | sigmavirus24 hughsaunders miguelgrinberg https://review.openstack.org/199307 is useful for the federation testing, and all sorts of other testing :) | 21:23 |
| sigmavirus24 | I don't believe you | 21:23 |
| sigmavirus24 | =P | 21:23 |
| sigmavirus24 | Also why are you still working | 21:23 |
| *** JRobinson__ has joined #openstack-ansible | 21:24 | |
| odyssey4me | sigmavirus24 well, you know, stuff :p | 21:24 |
| sigmavirus24 | and things? | 21:24 |
| odyssey4me | For hughsaunders' sake, I'll be doing a test tempest run using the above patch to ensure that everything is above board :p | 21:24 |
| odyssey4me | things YES! definitely things :p | 21:25 |
| cloudnull | things and stuff is where its at! | 21:30 |
| sigmavirus24 | Now I'm getting 400s from Keystone with the neutronclient/keystoneclient auth plugin stuff | 21:31 |
| sigmavirus24 | Trying to figure out what the magic options are now | 21:31 |
| sigmavirus24 | I think I might have figured it out | 21:31 |
| sigmavirus24 | But goddamn would this be nice if there were real docs | 21:31 |
| Sam-I-Am | HAHAHAHAHAH | 21:32 |
| Sam-I-Am | real docs | 21:32 |
| Sam-I-Am | HAHAHAHAHAHAHA | 21:32 |
| Sam-I-Am | you know, jamie lennox is on irc | 21:32 |
| sigmavirus24 | True | 21:32 |
| sigmavirus24 | It is that time of day | 21:32 |
| sigmavirus24 | I'm pretty sure I have the magic combo down | 21:32 |
| sigmavirus24 | If this time doesn't work, I'll bug him | 21:32 |
| odyssey4me | sigmavirus24 you reckon you have it licked? | 21:33 |
| odyssey4me | (like a mad cow) | 21:33 |
| sigmavirus24 | needed to add 'project_name', 'project_domain_name', 'user_domain_name' to the neutron section | 21:33 |
| sigmavirus24 | If I guessed correctly this tempest run should pass | 21:33 |
| sigmavirus24 | and nope | 21:34 |
| sigmavirus24 | let me push what I have so I can show him | 21:34 |
| Sam-I-Am | sigmavirus24: this is new stuff for liberty, iirc? | 21:34 |
| *** alop has joined #openstack-ansible | 21:34 | |
| Sam-I-Am | since neutron/nova stuff wasnt completely updated for kilo | 21:34 |
| Sam-I-Am | it was in neutron i think, but not in nova | 21:34 |
| palendae | sigmavirus24: up up down down left right left right a b select start? | 21:35 |
| sigmavirus24 | Sam-I-Am: neutronclient can use a keystoneclient session just fine | 21:36 |
| sigmavirus24 | That's what does auth things for neutronclient | 21:36 |
| sigmavirus24 | But I can't auth against keystone through that stuff for nova right now so I'm lost | 21:36 |
| openstackgerrit | Ian Cordasco proposed stackforge/os-ansible-deployment: Upgrade the Keystone library to use v3 https://review.openstack.org/196943 | 21:37 |
| sigmavirus24 | oh | 21:37 |
| sigmavirus24 | I wonder if I have to use internalurl for auth isntead of adminurl | 21:37 |
| odyssey4me | sigmavirus24 for what? | 21:38 |
| odyssey4me | all the current settings use internalurl | 21:38 |
| sigmavirus24 | in Nova's config of neutronclient/keystoneclient | 21:38 |
| odyssey4me | in fact they use internaluri (notice the 'i' not the 'l') | 21:38 |
| Sam-I-Am | heh l vs. i.... i see that all the time in #openstack | 21:38 |
| Sam-I-Am | auth_uri and auth_url THANKS JAMIE | 21:39 |
| alop | Anyone really against making osad work on rhel? | 21:41 |
| odyssey4me | alop nope, we're just waiting for someone who's actually interested in doing so to prep a spec/review to do so | 21:42 |
| alop | Like, I think my team and I would be looking to start using it, if we can upstream chaging things from "apt" to be more os agnostic | 21:42 |
| sigmavirus24 | alop: I am only because Red Hat maintenance schedules mean I have to support projects for Python 2.6 far past when it's been end-of-lifed upstream | 21:42 |
| sigmavirus24 | alop: you'll be interested in prometheanfire's spec to get osad onto gentoo then | 21:42 |
| sigmavirus24 | it's being split into two specs | 21:42 |
| alop | alright, I'll take a look | 21:43 |
| odyssey4me | alop we currently have someone who's interested in doing os-ansible-deployment on gentoo - after some discussion he's realised that it's a two step thing: 1) prep everything to handle the different conventions, package names, etc; 2) actually make it go on another platfor | 21:43 |
| odyssey4me | *platform | 21:43 |
| alop | like, the major architecture stuff, SOLID | 21:43 |
| Sam-I-Am | i'd avoid rhel's kernel if you want to use vxlan | 21:44 |
| odyssey4me | alop definitely make contact with prometheanfire with regards to the first stage spec | 21:44 |
| alop | alright, we'll take a look | 21:44 |
| odyssey4me | sigmavirus24 what was that toolset infra's prepped for projects to specify package deps? | 21:44 |
| *** JRobinson__ is now known as JRobinson__afk | 21:45 | |
| sigmavirus24 | odyssey4me: bindep or something? | 21:45 |
| odyssey4me | alop oh yes, https://github.com/openstack-infra/bindep is a key tool to use to simplify the genericising :) | 21:45 |
| prometheanfire | hi? | 21:45 |
| odyssey4me | (just saying) | 21:45 |
| *** Mudpuppy has quit IRC | 21:45 | |
| prometheanfire | bindep is a part of what's needed | 21:45 |
| prometheanfire | definitely | 21:45 |
| alop | yeah, checking out the spec | 21:46 |
| alop | bindep is nicer than the pkg-map we use in DIB | 21:46 |
| odyssey4me | alop prometheanfire did identify even the alternative platform enablement in two stages - 1) hosts, 2) containers | 21:47 |
| openstackgerrit | Matthew Thode proposed stackforge/os-ansible-deployment-specs: Add standalone swift testing https://review.openstack.org/199316 | 21:47 |
| prometheanfire | heh | 21:47 |
| odyssey4me | start with prepping the hosts, but use ubuntu containers initially to reduce the work involved... then later move on to converting the containers | 21:47 |
| prometheanfire | containers will be a very large undertaking | 21:48 |
| sigmavirus24 | alop: assuming not having rhel containers is fine | 21:48 |
| *** sdake_ has joined #openstack-ansible | 21:48 | |
| alop | well, we can figure out the particulars, first we're just checking to make sure we wouldn't hit ideological opposition | 21:49 |
| prometheanfire | we? | 21:50 |
| odyssey4me | alop the only place we're a bit stuck on - but perhaps negotiable once you've done the first bit - is deployment of the openstack bits from source :) | 21:50 |
| alop | yeah, the team I'm on | 21:50 |
| prometheanfire | nice, more bodies :D | 21:50 |
| sigmavirus24 | prometheanfire: "bodies" | 21:52 |
| *** sdake has quit IRC | 21:52 | |
| odyssey4me | sigmavirus24 re: https://review.openstack.org/196943 I think we need to use the term 'domain_name' in keeping with 'project_name', etc | 21:54 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Updated keystone to use fernet as the default https://review.openstack.org/195226 | 21:54 |
| sigmavirus24 | odyssey4me: actually, those are two separate things | 21:54 |
| sigmavirus24 | user_domain_name, project_domain_name are different than domain_name | 21:55 |
| sigmavirus24 | domain_name is used to scope something to that domain name instead of identifying the domain name for the project/user | 21:55 |
| odyssey4me | especially with 'default' happening to be the domain ID of the default domain in Keystone v3... and 'Default' being the name... stuff gets confusing | 21:55 |
| *** radek__ has quit IRC | 21:55 | |
| sigmavirus24 | Yeah, I have something I have to run to but I'll be pestering Jamie when I get back | 21:55 |
| sigmavirus24 | The errors from keystone are woefully unhelpful | 21:55 |
| dstanek | sigmavirus24: you're welcome | 21:56 |
| *** yaya has quit IRC | 21:56 | |
| odyssey4me | hmm - in that case perhaps we need to think up a different term because 'domain' is hella confusing | 21:56 |
| sigmavirus24 | dstanek: my interpretation is correct, yes? | 21:56 |
| odyssey4me | in terms of interacting with the library | 21:56 |
| sigmavirus24 | oh maybe | 21:57 |
| sigmavirus24 | DSLs gonna DSL | 21:57 |
| dstanek | sigmavirus24: depends. some areas are better than others. what issue are you having now? | 21:57 |
| odyssey4me | if I do a command in ansible to create a project in a domain, will I be using 'domain: <the appropriate project domain name>' in the task? | 21:57 |
| sigmavirus24 | dstanek: using keystoneclient auth plugins in nova for neutronclient. If you look at the review that odyssey4me linked, I'm getting http://paste.openstack.org/show/353136/ in the nova-api-os-compute logs in the nova-api-os-compute container | 21:58 |
| *** JRobinson__afk is now known as JRobinson__ | 21:58 | |
| odyssey4me | or will it be expected to be the domain ID (ie a UUID) ? | 21:58 |
| sigmavirus24 | odyssey4me: domain name | 21:58 |
| sigmavirus24 | also domain id's are not UUIDs | 21:58 |
| odyssey4me | heh dstanek has seen enough of nova lately :p | 21:58 |
| sigmavirus24 | at least the Default's id isn't | 21:58 |
| dstanek | it looks like it's the user id or user name that's missing. | 22:00 |
| dstanek | odyssey4me: yes, nova is not kind to me | 22:00 |
| odyssey4me | sigmavirus24 the domain by the name of 'Default' has an ID of 'default', but any other domains created will have an ID of something like '65cba76f475e4f03bcaf59c253193f30' | 22:00 |
| sigmavirus24 | odyssey4me: interesting | 22:02 |
| sigmavirus24 | dstanek: yeah I can't figure out why the username is missing | 22:02 |
| sigmavirus24 | anyway | 22:03 |
| sigmavirus24 | I'll be back later | 22:03 |
| sigmavirus24 | oh also | 22:03 |
| sigmavirus24 | fun stuff | 22:03 |
| sigmavirus24 | one keystoneclient auth plugin says "username" is deprecated in favor of "user-name" | 22:03 |
| sigmavirus24 | Meanwhile others just refer to "username" | 22:03 |
| sigmavirus24 | wtf | 22:03 |
| odyssey4me | sigmavirus24 you see why I'm suggesting that our arg rather be 'domain_name' when we interact with the library? | 22:03 |
| *** sacharya has quit IRC | 22:03 | |
| sigmavirus24 | odyssey4me: which library? our keystone module? | 22:03 |
| dstanek | sigmavirus24: which one? i've never seen anything refer to user-name | 22:04 |
| odyssey4me | sigmavirus24 yep | 22:04 |
| sigmavirus24 | I have to run but I'll be back | 22:04 |
| sigmavirus24 | dstanek: v3password uses username | 22:04 |
| odyssey4me | sigmavirus24 let me be clear - the task should use 'domain_name: <blah>' explicitly and the ansible keystone library should work out the complexity in the background | 22:04 |
| *** KLevenstein_ has joined #openstack-ansible | 22:05 | |
| *** KLevenstein has quit IRC | 22:07 | |
| *** KLevenstein_ is now known as KLevenstein | 22:07 | |
| *** galstrom is now known as galstrom_zzz | 22:10 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Enable all services to use Keystone 'insecurely' https://review.openstack.org/199307 | 22:15 |
| *** weezS has quit IRC | 22:19 | |
| *** fawadkhaliq has joined #openstack-ansible | 22:22 | |
| *** daneyon has quit IRC | 22:34 | |
| *** KLevenstein has quit IRC | 22:34 | |
| *** andymccr has joined #openstack-ansible | 22:46 | |
| *** annashen has quit IRC | 22:54 | |
| openstackgerrit | Miguel Grinberg proposed stackforge/os-ansible-deployment: SSL support for haproxy https://review.openstack.org/198957 | 22:59 |
| *** fawadkhaliq has quit IRC | 23:02 | |
| *** metral has quit IRC | 23:03 | |
| *** andymccr has quit IRC | 23:09 | |
| *** andymccr has joined #openstack-ansible | 23:10 | |
| openstackgerrit | Miguel Grinberg proposed stackforge/os-ansible-deployment: SSL support for haproxy https://review.openstack.org/198957 | 23:13 |
| *** andymccr has quit IRC | 23:18 | |
| *** britthouser has quit IRC | 23:23 | |
| *** andymccr has joined #openstack-ansible | 23:24 | |
| *** andymccr has quit IRC | 23:25 | |
| *** andymccr has joined #openstack-ansible | 23:26 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Enable all services to use Keystone 'insecurely' https://review.openstack.org/199307 | 23:31 |
| dstanek | ^ best review title ever | 23:31 |
| odyssey4me | dstanek :) | 23:37 |
| dstanek | odyssey4me: if i was osad core i would +2 for the title alone | 23:43 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Enable all services to use Keystone 'insecurely' https://review.openstack.org/199307 | 23:46 |
| odyssey4me | there, I think that's the last of it :) | 23:47 |
| odyssey4me | dstanek you can always +1 with a comment :) | 23:47 |
| sigmavirus24 | odyssey4me: oh we agree them | 23:49 |
| *** TheIntern has quit IRC | 23:49 | |
| odyssey4me | sigmavirus24 good :) | 23:50 |
| sigmavirus24 | oh dstanek user-name is in generic password method | 23:50 |
| odyssey4me | sigmavirus24 I had to edit the existing Ansible libraries a bit, and the plays that use them: https://review.openstack.org/199307 | 23:50 |
| sigmavirus24 | dstanek: https://github.com/openstack/python-keystoneclient/blob/master/keystoneclient/auth/identity/generic/password.py#L29 | 23:50 |
| sigmavirus24 | I'll try out username | 23:51 |
| sigmavirus24 | should have tried that sooner | 23:51 |
| sigmavirus24 | Failing that, I'll bug Jamie | 23:51 |
| sigmavirus24 | dstanek: that was it | 23:58 |
| openstackgerrit | Ian Cordasco proposed stackforge/os-ansible-deployment: Upgrade the Keystone library to use v3 https://review.openstack.org/196943 | 23:58 |
| sigmavirus24 | now tempest passes locally | 23:58 |
| sigmavirus24 | And with that, I say good night | 23:58 |
| stevelle | nn | 23:59 |
| sigmavirus24 | or not | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!