| *** eil397_ has joined #openstack-ansible | 00:06 | |
| *** eil397__ has quit IRC | 00:09 | |
| *** eil397 has quit IRC | 00:10 | |
| *** v1k0d3n has quit IRC | 00:16 | |
| *** eil397_ has quit IRC | 00:17 | |
| *** eil397_ has joined #openstack-ansible | 00:19 | |
| *** sdake has quit IRC | 00:23 | |
| *** cloudtrainme has quit IRC | 00:24 | |
| *** hwcomcn has joined #openstack-ansible | 00:24 | |
| *** cloudtrainme has joined #openstack-ansible | 00:26 | |
| *** spotz is now known as spotz_zzz | 00:30 | |
| *** cloudtrainme has quit IRC | 00:31 | |
| *** jamesden_ has quit IRC | 00:38 | |
| *** thorst has joined #openstack-ansible | 00:48 | |
| *** jamielennox is now known as jamielennox|away | 00:49 | |
| *** thorst has quit IRC | 00:54 | |
| *** thorst has joined #openstack-ansible | 00:54 | |
| *** hwcomcn has quit IRC | 00:55 | |
| *** thorst has quit IRC | 00:59 | |
| *** elo has quit IRC | 00:59 | |
| *** elo has joined #openstack-ansible | 01:11 | |
| *** sdake has joined #openstack-ansible | 01:13 | |
| *** sdake has quit IRC | 01:13 | |
| *** sdake has joined #openstack-ansible | 01:13 | |
| *** sdake_ has joined #openstack-ansible | 01:15 | |
| *** eil397_ has quit IRC | 01:17 | |
| *** sdake has quit IRC | 01:18 | |
| *** asettle has joined #openstack-ansible | 01:22 | |
| *** rstarmer has joined #openstack-ansible | 01:23 | |
| *** asettle has quit IRC | 01:27 | |
| *** jamielennox|away is now known as jamielennox | 01:34 | |
| *** winggundamth has quit IRC | 01:39 | |
| *** eil397_ has joined #openstack-ansible | 01:41 | |
| *** winggundamth has joined #openstack-ansible | 01:46 | |
| *** thorst has joined #openstack-ansible | 02:04 | |
| *** thorst has quit IRC | 02:05 | |
| *** thorst has joined #openstack-ansible | 02:05 | |
| *** thorst has quit IRC | 02:10 | |
| *** abelur has joined #openstack-ansible | 02:17 | |
| *** woodard has quit IRC | 02:17 | |
| *** sdake_ has quit IRC | 02:19 | |
| *** elopez has joined #openstack-ansible | 02:25 | |
| *** elopez is now known as Guest78543 | 02:25 | |
| *** Guest78543 has quit IRC | 02:26 | |
| *** eil397_ has quit IRC | 02:26 | |
| *** jamielennox is now known as jamielennox|away | 02:32 | |
| *** elo has quit IRC | 02:56 | |
| *** elo has joined #openstack-ansible | 02:59 | |
| *** spotz_zzz is now known as spotz | 03:01 | |
| *** rstarmer has quit IRC | 03:06 | |
| *** zerda2 has joined #openstack-ansible | 03:14 | |
| *** eil397_ has joined #openstack-ansible | 03:35 | |
| *** Guest78543 has joined #openstack-ansible | 03:36 | |
| *** Guest78543 has quit IRC | 03:38 | |
| *** eil397_ has quit IRC | 03:39 | |
| *** xar- has quit IRC | 03:53 | |
| *** xar- has joined #openstack-ansible | 03:56 | |
| openstackgerrit | Adam Reznechek proposed openstack/openstack-ansible-os_nova: Update PowerVM compute driver location https://review.openstack.org/348720 | 03:58 |
|---|---|---|
| openstackgerrit | Adam Reznechek proposed openstack/openstack-ansible-os_nova: Update PowerVM compute driver path https://review.openstack.org/348720 | 04:04 |
| openstackgerrit | Adam Reznechek proposed openstack/openstack-ansible: Add discovery and build for multiple CPU architectures https://review.openstack.org/346829 | 04:28 |
| *** eil397_ has joined #openstack-ansible | 04:41 | |
| *** McMurlock1 has joined #openstack-ansible | 04:42 | |
| *** McMurlock1 has quit IRC | 04:44 | |
| *** elopez has joined #openstack-ansible | 04:46 | |
| *** elopez is now known as Guest31932 | 04:47 | |
| *** hybridpollo has quit IRC | 04:48 | |
| *** elo has quit IRC | 04:56 | |
| *** elo has joined #openstack-ansible | 05:00 | |
| *** eil397 has joined #openstack-ansible | 05:00 | |
| *** janki has joined #openstack-ansible | 05:21 | |
| *** shausy has joined #openstack-ansible | 05:31 | |
| *** wadeholler has quit IRC | 05:42 | |
| *** wadeholler has joined #openstack-ansible | 05:43 | |
| openstackgerrit | Darren Chan proposed openstack/openstack-ansible: [docs] Remove duplicated content https://review.openstack.org/348737 | 05:43 |
| openstackgerrit | Darren Chan proposed openstack/openstack-ansible: [docs] Remove duplicated content https://review.openstack.org/348737 | 05:45 |
| *** david-lyle has joined #openstack-ansible | 06:07 | |
| *** shausy has quit IRC | 06:20 | |
| *** david-lyle has quit IRC | 06:45 | |
| *** pcaruana has joined #openstack-ansible | 06:45 | |
| *** eil397_ has quit IRC | 06:50 | |
| *** eil397 has quit IRC | 06:51 | |
| *** rstarmer has joined #openstack-ansible | 06:56 | |
| *** rstarmer has quit IRC | 06:56 | |
| odyssey4me | o/ | 07:00 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-plugins: Update py_pkgs to set options group membership and itemise global pins https://review.openstack.org/347930 | 07:07 |
| odyssey4me | csmart ping? I hear you want help getting OSA to work behind a proxy? | 07:07 |
| csmart | odyssey4me: hey, yeah I've been testing OSA behinda proxy, I think I have it mostly nutted out now.. the missing piece of the puzzle was havig /etc/environment populated with proxy settings before running run-playbooks (which also populates /etc/environment) | 07:10 |
| csmart | (playbooks was failing to download the pip installer) | 07:10 |
| odyssey4me | csmart yes, the host must already work before running any scripts | 07:10 |
| csmart | just spinning up a fresh vm to test it now | 07:10 |
| odyssey4me | so you need to, at least, export HTTP_PROXY, http_proxy, etc on the host before running the scripts | 07:10 |
| odyssey4me | it may be worth adding some documentation to that effect to make it clear | 07:11 |
| *** spotz is now known as spotz_zzz | 07:12 | |
| csmart | rgr, I'm just making sure that I understand it properly and have it working reliably, then I'll update the doco :-) | 07:13 |
| csmart | odyssey4me: yeah however the thing appears to be that even with proxies exported, run-playbooks.sh will fail if /etc/environment is not populated also | 07:14 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Only build wheels for services which are deployed https://review.openstack.org/344912 | 07:17 |
| odyssey4me | csmart ah, that's probably due to a subshell | 07:18 |
| odyssey4me | I wonder if we shouldn't actually do away with the environment building and simply have the containers pull in the environment file, as it does for the DNS config and apt config | 07:20 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-plugins: Update py_pkgs to set options group membership and itemise global pins https://review.openstack.org/347930 | 07:26 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Only clone git repositories for services which are deployed https://review.openstack.org/345642 | 07:27 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Only clone git repositories for services which are deployed https://review.openstack.org/345642 | 07:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Only clone git repositories for services which are deployed https://review.openstack.org/345642 | 07:29 |
| odyssey4me | jmccrory you around? | 07:30 |
| mancdaz | odyssey4me https://review.openstack.org/#/c/347195/7 | 07:32 |
| odyssey4me | mancdaz yep, can you add a patch afterwards to update the manual steps documentation? | 07:32 |
| mancdaz | odyssey4me yeah, my last comment essentially says 'can I do this in a separate patch' | 07:33 |
| csmart | odyssey4me: ok, yeah I'll play with this a bit and then I'll poke around some more | 07:33 |
| odyssey4me | mancdaz yeah, just spotted that - thanks | 07:33 |
| mancdaz | odyssey4me you're tagging today anyway? So I will prepare the patch this morning | 07:33 |
| odyssey4me | mancdaz git-harry so we have a bit of a conundrum regarding the override of upper-constraints | 07:34 |
| mancdaz | oh? | 07:34 |
| odyssey4me | global-requirements-pins.txt is only applied to *requirements* | 07:34 |
| odyssey4me | unfortunately if your requirement is different to the constraint, pip makes the constraint win | 07:35 |
| odyssey4me | so if you have a requirements file with 'elasticsearch<2.1.0' and a constraints file with 'elasticsearch<=2.1.0' then the constraint will win | 07:35 |
| odyssey4me | jmccrory tried to make the g-r-p file be applied to u-c too https://review.openstack.org/348502 - but it's not working because requirements aren't exactly constraints | 07:36 |
| odyssey4me | the initial idea though, was to have a var which you would override - either through group vars or user_vars | 07:36 |
| odyssey4me | would that be suitable? | 07:37 |
| *** mikelk has joined #openstack-ansible | 07:37 | |
| mancdaz | I think as long as there is some mechanism to be able to guarantee a version it should be fine | 07:38 |
| odyssey4me | this was the initial idea: https://review.openstack.org/#/c/348502/1 | 07:38 |
| mancdaz | though for this case maybe we need to change the elasticsearch requirement in rpco so that we can move forward | 07:38 |
| odyssey4me | mancdaz updating elasticsearch is not trivial, it involves data migrations and such | 07:39 |
| odyssey4me | the formatting of the data changes with versions | 07:39 |
| mancdaz | 2.0 > 2.1? | 07:39 |
| odyssey4me | that said, I don't know how hard the 2.0 - 2.1 update would be - worth investigating | 07:39 |
| git-harry | Yeah, we just need a mechanism that allows us to force a version on the assumption we know what we're doing. | 07:40 |
| odyssey4me | that said, doing that today before tagging would not be wise | 07:40 |
| mancdaz | if it's anything more than completely trivial, we probably don't want to do it | 07:40 |
| git-harry | I think it was pinned for a reason, it's certainly not me preferred option | 07:40 |
| odyssey4me | yeah - I want to put a band aid on it now to help your guys get a reliable upgrade out the door | 07:40 |
| odyssey4me | ok, so our option right now that has minimal change is to add a var to the repo build process - this is the most backportable option as it keeps things stable and doesn't change the process much | 07:41 |
| odyssey4me | the only other option is to revert to a broken state, which realistically is not an option | 07:41 |
| odyssey4me | unless someone has any other ideas? | 07:42 |
| *** afred312_ has joined #openstack-ansible | 07:42 | |
| git-harry | Accept an argument could be made that it is a feature if people are relying on it. | 07:42 |
| *** wadeholl_ has joined #openstack-ansible | 07:42 | |
| odyssey4me | git-harry sure, except that the brokenness was that the build was not repeatable (ie the package list for the tag today is not the same as the list for the tag tomorrow)... that certainly outweigh's the inability to override u-c | 07:43 |
| git-harry | I'm not sure it does for us :P | 07:44 |
| odyssey4me | technically being able to override u-c is a new feature | 07:44 |
| *** wadeholler has quit IRC | 07:44 | |
| odyssey4me | but that's semantics | 07:44 |
| *** afred312 has quit IRC | 07:45 | |
| evrardjp | good morning | 07:45 |
| mancdaz | odyssey4me so that review does not work right now? | 07:46 |
| odyssey4me | no | 07:46 |
| odyssey4me | pip fails to handle it | 07:46 |
| odyssey4me | quite honestly the g-r-p is a requirement override, not a constraint override, so I really don't think it's appropriate to mix them anyway | 07:47 |
| git-harry | I agree, I don't think the purpose of the file should be changed in Liberty | 07:48 |
| mancdaz | git-harry the elasticsearch pip package was only introduced in liberty rpco, because it was a requirement of elasticsearch-curator. I think we actually use the elasticsearch apt package | 07:49 |
| git-harry | mancdaz: Yes, just looking at it. | 07:49 |
| mancdaz | since we don't pin elasticsearch-curator, and we are likely now getting a later version than we were at the time, we could probably bump the required version of elasticsearch | 07:50 |
| mancdaz | oh no wait we do now | 07:51 |
| mancdaz | so elasticsearch-curator==3.4.1 (which is what we get with elasticsearch-curator<3.5.0') requires elasticsearch>=1.8.0,<2.4.0 | 07:54 |
| *** admin0 has joined #openstack-ansible | 07:54 | |
| git-harry | mancdaz: https://github.com/elastic/curator/blob/v3.4.1/requirements.txt#L2 | 07:54 |
| git-harry | Where are you seeing that, the repo says different | 07:54 |
| odyssey4me | hmm, so if the elasticsearch python package is only there for the libs it provides then it should not affect the data | 07:54 |
| git-harry | Yes, but there it can't see a version of elasticsearch-curator that supports elasticsearch===2.1.0 | 07:55 |
| mancdaz | git-harry in venv if I install elasticsearch==2.1.0, and then elasticsearch-curator==3.4.1 | 07:55 |
| mancdaz | Requirement already satisfied (use --upgrade to upgrade): elasticsearch>=1.8.0,<2.4.0 in /Users/darren.birkett/.virtualenvs/esc/lib/python2.7/site-packages (from elasticsearch-curator<3.5.0) | 07:56 |
| mancdaz | pip freeze | grep elas | 07:56 |
| mancdaz | elasticsearch==2.1.0 | 07:56 |
| mancdaz | elasticsearch-curator==3.4.1 | 07:56 |
| git-harry | https://github.com/elastic/curator/blob/v3.4.1/setup.py#L25 | 07:57 |
| git-harry | The specify different versions in different places | 07:57 |
| mancdaz | awesome | 07:57 |
| mancdaz | https://github.com/elastic/curator/commit/30c26b41a9540547e89b3bba12f5ca36577a5b50 | 07:59 |
| mancdaz | seems like they forgot to update requirements.txt when they did this | 07:59 |
| git-harry | Yup | 07:59 |
| git-harry | The previous one had updated both | 08:00 |
| mancdaz | lol https://github.com/elastic/curator/issues/524 | 08:01 |
| mancdaz | 'probably' | 08:01 |
| *** eil397_ has joined #openstack-ansible | 08:05 | |
| *** ameobapox has quit IRC | 08:06 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Provide ability to override upper-constraints https://review.openstack.org/348775 | 08:09 |
| *** eil397_ has quit IRC | 08:09 | |
| odyssey4me | the test in ^ will fail right now - I'm doing a build to validate that it does what I want, then will figure out the test | 08:10 |
| *** electrofelix has joined #openstack-ansible | 08:13 | |
| *** maddy__ has joined #openstack-ansible | 08:15 | |
| maddy__ | good noon | 08:19 |
| odyssey4me | o/ maddy__ | 08:24 |
| csmart | odyssey4me: quick one, when are openstack-ansible/etc/openstack_deploy/user*.yml used? | 08:24 |
| odyssey4me | csmart those are examples/samples - not used for production... some of the files are used when building the AIO | 08:25 |
| odyssey4me | csmart https://github.com/openstack/openstack-ansible/blob/master/tests/roles/bootstrap-host/tasks/prepare_aio_config.yml#L28-L60 | 08:25 |
| odyssey4me | only the secrets file would typically be used directly for production | 08:26 |
| odyssey4me | the user_variables.yml file has a blurb in the beginning which explains its purpose | 08:26 |
| csmart | odyssey4me: kk thanks, makes sense | 08:28 |
| maddy__ | iam facing issue with AIO, In thee path `/etc/ansible/roles/openstack_hosts/tasks`, file `openstack_update_hosts_file.yml` failing with fillowing error whe in run `run-playbooks.sh` | 08:30 |
| maddy__ | fatal: [192.168.18.163] => {'msg': "AnsibleUndefinedVariable: One or more undefined variables: 'dict object' has no attribute 'ansible_ssh_host'", 'failed': True} | 08:30 |
| csmart | odyssey4me: ok, so /etc/openstack_deploy/user_variables.yml comes from openstack-ansible/tests/roles/bootstrap-host/templates/user_variables.aio.yml.j2 which seems to be run as a part of bootstrap-aio.ymlf | 08:31 |
| * csmart thinks he's slowly piecing things together | 08:33 | |
| maddy__ | (continuing on my question....) The issue come from openstack-host-hostfile-setup.sh.j2 ? | 08:34 |
| odyssey4me | maddy__ it sounds to me like you've modified the configuration of the AIO, so it's failing | 08:41 |
| odyssey4me | the AIO is a development/testing tool and is only made to work with a very specific configuration | 08:41 |
| odyssey4me | maddy__ so the best is to revert your changes one by one to see what change you made which breaks it | 08:42 |
| odyssey4me | csmart in an AIO, yes - for production it's expected that a user_variables.yml file would be put together by the deployer | 08:43 |
| odyssey4me | csmart that file is entirely optional | 08:43 |
| maddy__ | so i need to revert from `openstack_user_config.yml`, `user_variables.yml`, `user_conf_files.yml`? | 08:43 |
| csmart | odyssey4me: rgr, thanks | 08:43 |
| odyssey4me | the user_* files in /etc/openstack_deploy are added as a -e CLI option to the ansible-playbook command through the convenience bash script 'openstack-ansible' | 08:44 |
| odyssey4me | maddy__ I would recommend starting with a fresh VM again and executing the AIO build exactly as it shows in the guide with no other modifications. | 08:44 |
| *** shausy has joined #openstack-ansible | 08:45 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Provide ability to override upper-constraints https://review.openstack.org/348775 | 08:48 |
| odyssey4me | evrardjp git-harry mancdaz ^ reviews please | 08:48 |
| odyssey4me | adding a reno now | 08:49 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Provide ability to override upper-constraints https://review.openstack.org/348775 | 08:52 |
| evrardjp | seems nice | 08:56 |
| evrardjp | I'm surprise that upper_constraints: "{{ upper_constraints | xxx }}" works though | 08:57 |
| evrardjp | it used to not work | 08:57 |
| odyssey4me | evrardjp tested with 1.9.4 and it seems to wor | 08:59 |
| odyssey4me | *work | 08:59 |
| *** DrifterZA has joined #openstack-ansible | 08:59 | |
| odyssey4me | bear in mind, though, that it's a set fact - not something in group_vars or extra vars | 08:59 |
| odyssey4me | I was also surprised, actually :) | 09:00 |
| DrifterZA | morning | 09:00 |
| evrardjp | yup that's the only way it works | 09:00 |
| odyssey4me | o/ DrifterZA | 09:00 |
| evrardjp | because it just "applies after parsing" | 09:00 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Better control of mariadb restarts in k>l upgrades https://review.openstack.org/347195 | 09:10 |
| maddy__ | odyssey4me.. Changing IP's and ignoring (commenting) some services on config files, will affect anything ? | 09:13 |
| evrardjp | odyssey4me: we have frequent networking issues on xenial: https://review.openstack.org/348775 just failed on xenial due to a failure of pip_install :/ | 09:13 |
| evrardjp | it's listed as a problem "Failed to validate SSL certificate" but I don't trust pip error reporting | 09:14 |
| evrardjp | I think it's gonna worth an effort at some point | 09:15 |
| odyssey4me | maddy__ the IP's set and the configs set are specifically for the AIO's deployment | 09:16 |
| odyssey4me | maddy__ to change the services deployed, you need to adjust the openstack_user_config / conf.d entries, as those define what services are deployed | 09:17 |
| odyssey4me | evrardjp yeah, mhayden has been working that out - https://review.openstack.org/347372 resolved it for CentOS and it looks like Xenial may need similar treatment | 09:17 |
| odyssey4me | evrardjp as I recall there was a comment that the issues appear to be specific to some providers | 09:18 |
| odyssey4me | maddy__ I recommend that you read the install guide thoroughly before modifying the configs | 09:18 |
| evrardjp | interesting | 09:19 |
| *** jiteka has joined #openstack-ansible | 09:30 | |
| odyssey4me | evrardjp https://review.openstack.org/#/c/344844/ | 09:30 |
| odyssey4me | although it seems that trusty's python (2.7.6) shouldn't have native support for SNI either | 09:31 |
| *** permalac has joined #openstack-ansible | 09:33 | |
| evrardjp | interesting | 09:34 |
| evrardjp | where did you find the information about the sni in trusty? | 09:34 |
| evrardjp | please don't answer google | 09:34 |
| evrardjp | I can google | 09:34 |
| evrardjp | :p | 09:34 |
| odyssey4me | yep, google | 09:35 |
| evrardjp | :p | 09:35 |
| odyssey4me | https://github.com/SiCKRAGETV/issues/wiki/SSL-Errors | 09:35 |
| odyssey4me | http://stackoverflow.com/questions/28656418/python-ssl-via-sni-proxy | 09:35 |
| odyssey4me | http://stackoverflow.com/questions/29572161/how-to-securely-use-pip-with-ssl-on-ubuntu-trusty | 09:36 |
| evrardjp | I thought pyopenssl and cryptography were enough to be honest | 09:36 |
| odyssey4me | hmm, so the trick here is that we need the SSL working properly to get pip in the first place | 09:37 |
| evrardjp | I'll spin up a xenial and check the installed packages and pip versions | 09:42 |
| evrardjp | it's gonna help me in the long run | 09:42 |
| evrardjp | will quickly write something with requests, now that requests supports sni | 09:42 |
| * evrardjp wrotes again in its todo-list | 09:42 | |
| evrardjp | odyssey4me: I +W on the 344844 | 09:45 |
| evrardjp | if it's not a fine fix, we'll fix that later | 09:45 |
| evrardjp | it's master anyway | 09:45 |
| evrardjp | :p | 09:45 |
| *** tlbr has quit IRC | 09:52 | |
| *** tlbr has joined #openstack-ansible | 09:53 | |
| *** tlbr has quit IRC | 09:53 | |
| *** tlbr has joined #openstack-ansible | 09:54 | |
| *** tlbr has quit IRC | 09:54 | |
| *** tlbr has joined #openstack-ansible | 09:56 | |
| *** tlbr is now known as Guest45466 | 09:56 | |
| *** eil397_ has joined #openstack-ansible | 09:58 | |
| *** Guest45466 has quit IRC | 10:00 | |
| *** eil397_ has quit IRC | 10:03 | |
| automagically | super, very, incredibly early morning | 10:04 |
| DrifterZA | heya | 10:04 |
| DrifterZA | Travis I made that role a galaxy role now | 10:04 |
| DrifterZA | https://galaxy.ansible.com/drifterza/openstack-ansible-os_mongo/ | 10:04 |
| DrifterZA | if you want to peruse it | 10:05 |
| automagically | Saw your message, thx. Will take a look | 10:05 |
| DrifterZA | cool, let me know if theres anything I missed. | 10:06 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server: Fix get_url SNI issues in CentOS 7 https://review.openstack.org/344844 | 10:08 |
| *** tlbr_ has joined #openstack-ansible | 10:13 | |
| automagically | evrardjp: Mind reviewing https://review.openstack.org/#/c/345568/ https://review.openstack.org/#/c/348011/ and https://review.openstack.org/#/c/348036/ | 10:14 |
| *** tlbr_ has quit IRC | 10:17 | |
| *** sdake has joined #openstack-ansible | 10:25 | |
| evrardjp | automagically: will do | 10:30 |
| automagically | thx | 10:31 |
| evrardjp | that was the only thing to add rally support? I'll update the etherpad too then | 10:33 |
| automagically | Yep, no deb packages installed, so pretty simple | 10:34 |
| openstackgerrit | Merged openstack/openstack-ansible-os_zaqar: Removing unused tasks and templates for upstart https://review.openstack.org/348036 | 10:41 |
| odyssey4me | hmm, evrardjp automagically it would seem to me that the SNI issue is going to be a problem across all platforms soon enough | 10:46 |
| automagically | Yep | 10:46 |
| odyssey4me | as people adjust the web services provided to use it | 10:46 |
| odyssey4me | so we'd better get some sort of solution in that works across all branches | 10:46 |
| odyssey4me | sooner is better | 10:47 |
| automagically | change for openstack_hosts perhaps? | 10:47 |
| automagically | I know we just took that role out of alot of the role testing, but for these types of cross-cutting concerns, it seems like we could cover them there | 10:47 |
| automagically | and/or in lxc_container_create | 10:47 |
| openstackgerrit | Merged openstack/openstack-ansible-os_rally: Implement Ubuntu 16.04 support https://review.openstack.org/345568 | 10:48 |
| evrardjp | automagically: what about using neutron_openvswitch_agent_ini_overrides for https://review.openstack.org/#/c/348011 ? You have to use config overrides anyway... | 10:52 |
| evrardjp | (I mean for the rest of the vars) | 10:52 |
| automagically | evrardjp: Simply because all of the other qos config is already there in the role | 10:53 |
| automagically | We document that all a deployer need do is add qos to the plugin list and voila… | 10:53 |
| automagically | And I like the simplicity of that approach | 10:54 |
| odyssey4me | automagically well, it seems that the packages are needed on the deployment host - so bootstrap-ansible is one place | 10:56 |
| odyssey4me | then it need to be in the containers too | 10:56 |
| odyssey4me | then I would guess on lxc hosts too | 10:56 |
| automagically | forgot about the deployment host | 10:56 |
| evrardjp | automagically: I suspect Newton and Mitaka don't have a config change for qos too | 11:02 |
| automagically | Not sure I understand evrardjp | 11:03 |
| automagically | I based the role changes on http://docs.openstack.org/mitaka/networking-guide/adv-config-qos.html#configuration | 11:03 |
| evrardjp | this is mitaka ^ | 11:04 |
| evrardjp | your change is in master | 11:05 |
| evrardjp | I hope this won't change in master :D | 11:05 |
| evrardjp | I mean I hope the doc is still valid | 11:05 |
| automagically | Some of my colleagues wrote the OVS agent code for QOS DSCP marks, and the doc updates and role updates are based on my conversations with them and testing | 11:06 |
| automagically | That said, ultimately, the same config should not need to be in ml2_conf.ini and openvswitch_agent.ini, but that is the current state of the neutron codebase | 11:07 |
| evrardjp | I see | 11:07 |
| evrardjp | we can change that later if the need arises | 11:07 |
| *** sguduru has joined #openstack-ansible | 11:13 | |
| automagically | Gah, bitten by https://bugs.launchpad.net/cloud-images/+bug/1565985 | 11:14 |
| openstack | Launchpad bug 1565985 in cloud-images "vagrant vb ubuntu/xenial64 cannot mount synced folders" [Undecided,In progress] - Assigned to Dan Watkins (daniel-thewatkins) | 11:14 |
| automagically | So much for being able to run Xenial tests locally | 11:14 |
| evrardjp | automagically: for qos, what do you think of changing the /etc/neutron/plugins/ml2/ml2_conf.ini to add extensions = qos in [agent] ? | 11:16 |
| evrardjp | this way we keep the "all things included" view | 11:17 |
| automagically | evrardjp: Let me test with it, and perhaps take a swing through the neutron source | 11:18 |
| automagically | Added to my lengthy todo list | 11:18 |
| evrardjp | I know the feeling | 11:18 |
| *** asettle has joined #openstack-ansible | 11:20 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron: Configure qos extension in openvswitch_agent.ini https://review.openstack.org/348011 | 11:20 |
| *** eil397_ has joined #openstack-ansible | 11:22 | |
| *** eil397_ has quit IRC | 11:26 | |
| *** sdake has quit IRC | 11:27 | |
| *** zerda2 has quit IRC | 11:35 | |
| odyssey4me | hmm, the implementation of openssl-dbg makes no sense to solve this - it's a debug extension | 11:35 |
| *** Guest5363 has joined #openstack-ansible | 11:36 | |
| *** Guest5363 has quit IRC | 11:36 | |
| odyssey4me | I suppose it actually makes sense to have the packages installed before trying to download pip? | 11:39 |
| *** spotz_zzz is now known as spotz | 11:40 | |
| *** thorst has joined #openstack-ansible | 11:40 | |
| *** retreved has joined #openstack-ansible | 11:46 | |
| mhayden | morning | 11:51 |
| mhayden | evrardjp / odyssey4me: my SSL SNI fix for CentOS 7 appears to work, but it makes Xenial unhappy | 11:51 |
| mhayden | have we seen Xenial fail cert validation? | 11:52 |
| *** sguduru has quit IRC | 11:52 | |
| odyssey4me | mhayden yes | 11:53 |
| mhayden | do you have a link to the error? | 11:53 |
| mhayden | i'm curious to see if it's the same as centos 7 | 11:53 |
| odyssey4me | I think automagically commented that it appears to be common to some providers, which probably means that the anycast address is pushing that traffic to a particular set of nodes. | 11:53 |
| mhayden | because xenial has a version of python that *should* have SNI built in | 11:53 |
| odyssey4me | hmm, are you sure about that? | 11:53 |
| odyssey4me | bear in mind that we're talking python2 here - not python3 | 11:54 |
| mhayden | it came in in 2.7.9 | 11:54 |
| odyssey4me | ansible only uses python2 | 11:54 |
| mhayden | http://distrowatch.com/table.php?distribution=ubuntu | 11:54 |
| mhayden | 15.10 had 2.7.9 | 11:54 |
| mhayden | so we know xenial is >= 2.7.9 | 11:55 |
| * mhayden should have gotten a bigger coffee | 11:55 | |
| odyssey4me | hmm fair point | 11:55 |
| mhayden | this is where i was messing around with it -> https://review.openstack.org/#/c/344844/ | 11:55 |
| mhayden | however, xenial did randomly fail once on SSL validation, but then did okay on the recheck | 11:56 |
| odyssey4me | it would appear, though that we need to ensure that http://packages.ubuntu.com/xenial/python-ndg-httpsclient / http://packages.ubuntu.com/trusty/python-pyasn1-modules / http://packages.ubuntu.com/trusty/python-openssl are installed | 11:56 |
| odyssey4me | note the first one is xenial only | 11:56 |
| mhayden | automagically left a helpful comment in that one | 11:56 |
| mhayden | odyssey4me: i'm not sure if i can do a conditional pip install based on OS in test-requirements.txt or tox.ini | 11:57 |
| odyssey4me | mhayden also note https://pkgs.org/centos-7/epel-x86_64/python-ndg_httpsclient-0.3.2-1.el7.noarch.rpm.html | 11:57 |
| *** spotz is now known as spotz_zzz | 11:57 | |
| mhayden | weird, i did look for that | 11:58 |
| odyssey4me | mhayden no, I think we need to actually ensure that we install it as part of the deployment - not just tests | 11:58 |
| mhayden | odyssey4me: want me to change up the OSA commit to have this stuff in other-requirements.txt ? | 11:58 |
| mhayden | or well, make a new commit | 11:58 |
| odyssey4me | I was thinking that perhaps it makes sense to add it to the pip_install role? | 11:58 |
| mhayden | you know what i mean :P | 11:58 |
| mhayden | well, i guess the bigger question is -- do we want to install these from .deb/.rpm or from pip? | 11:58 |
| mhayden | ndg-httpsclient isn't in g-r | 11:59 |
| mhayden | but the other two are | 11:59 |
| odyssey4me | yeah, so that all depends on what uses it | 11:59 |
| odyssey4me | it looks like ansible uses httplib2, which is what ndg-httpsclient is needed for | 11:59 |
| mhayden | ansible will be the only thing using it, honestly | 11:59 |
| odyssey4me | but we've also seen SNI errors when trying to download get-pip.py | 12:00 |
| mhayden | or on centos we install a modern python 2.7.x via SCE | 12:00 |
| odyssey4me | ah, but that's also ansible | 12:00 |
| mhayden | SCE = Software Collections | 12:00 |
| mhayden | yeah the issue is ansible's get_url | 12:00 |
| mhayden | it doesn't use python-requests in 1.9.x AFAIK | 12:00 |
| odyssey4me | ok, what do we think is the best way to solve this in a way that's suitable for backports? | 12:00 |
| mhayden | python-requests handles the stuff just fine when i open a python shell and test | 12:00 |
| mhayden | hmm | 12:00 |
| mhayden | perhaps pkgs is the best way to tackle it | 12:01 |
| odyssey4me | bootstrap-ansible, clearly | 12:01 |
| *** weshay is now known as weshay_mtg | 12:01 | |
| odyssey4me | and yes, packages I think | 12:01 |
| odyssey4me | all our service stuff goes into venvs, so the system packages are fair game | 12:01 |
| mhayden | so i should get the packages into other-requirements.txt so that testing works but then also add them to bootstrap-ansible? | 12:01 |
| odyssey4me | we'd need it wherever we use get_url, right? | 12:02 |
| mhayden | yessir | 12:02 |
| odyssey4me | which would mean pip_install, rabbitmq-server, galera-server | 12:02 |
| mhayden | makes sense | 12:02 |
| mhayden | i assume i should update those roles so that those pkgs are installed when those roles are installed, correct? | 12:02 |
| odyssey4me | I don't think it needs to be in other-requirements.txt for the most part, because we install stuff into containers | 12:03 |
| odyssey4me | so the container needs it, not the host | 12:03 |
| mhayden | well wait, the #1 place this is needed is on the deployment host, since that's where get_url runs | 12:03 |
| odyssey4me | oh, is that the *only* place? | 12:03 |
| mhayden | AFAIK, yes | 12:03 |
| odyssey4me | ok, then let's try this | 12:04 |
| mhayden | because on the rabbitmq issues, the problem was that the deployment host couldn't do get_url correctly | 12:04 |
| mhayden | the containers were unaffected | 12:04 |
| odyssey4me | bootstrap-ansible needs to install the packages (not the wheels) | 12:04 |
| odyssey4me | it needs to do it for every platform we support | 12:04 |
| mhayden | agreed | 12:04 |
| odyssey4me | then in every other-requirements.txt file we need the packages deployed | 12:04 |
| odyssey4me | and run_tests | 12:04 |
| mhayden | ah, so every repo needs it i guess | 12:04 |
| odyssey4me | other-reqs is used by zuul - run_tests is used by humans/vagrant | 12:05 |
| *** tlbr_ has joined #openstack-ansible | 12:05 | |
| odyssey4me | it might be a good time to adjust run_tests so that it uses the more standard OS detection method that csmart implemented in the integrated repo | 12:05 |
| *** hwcomcn has joined #openstack-ansible | 12:05 | |
| odyssey4me | and yeah, while not *every* repo needs it - I'd prefer to keep things relatively standard | 12:06 |
| odyssey4me | it makes doing future updates easier | 12:06 |
| mhayden | i tend to agree there | 12:06 |
| mhayden | i shall break out my for loops | 12:06 |
| *** hwcomcn has quit IRC | 12:07 | |
| *** hwcomcn has joined #openstack-ansible | 12:07 | |
| odyssey4me | automagically mhayden thoughts on https://review.openstack.org/348775 ? | 12:08 |
| odyssey4me | i nmy view, it's the best way of providing the facility in an easily backportable manner | 12:09 |
| * mhayden looks | 12:09 | |
| odyssey4me | the other option being explored involved changes to py_pkgs which started getting quite hairy | 12:09 |
| *** hwcomcn has quit IRC | 12:09 | |
| * mhayden prefers less hair | 12:09 | |
| *** tlbr_ has quit IRC | 12:09 | |
| *** hwcomcn has joined #openstack-ansible | 12:10 | |
| openstackgerrit | Merged openstack/openstack-ansible: Better control of mariadb restarts in k>l upgrades https://review.openstack.org/347195 | 12:10 |
| odyssey4me | mhayden must like b3rnard0 then :p | 12:10 |
| odyssey4me | git-harry mancdaz ^ w00t! | 12:10 |
| mancdaz | yay! | 12:11 |
| *** hwcomcn has quit IRC | 12:11 | |
| odyssey4me | mhayden regardless of whether we're actually seeing the SNI errors today for Trusty/Xenial, it would appear that seeing the errors is inevitable as more and more services use the changed certs | 12:11 |
| odyssey4me | so let's get the fix in early across all our supported platforms | 12:12 |
| odyssey4me | it helps a tag cut today work tomorrow :p | 12:12 |
| *** hwcomcn has joined #openstack-ansible | 12:12 | |
| git-harry | partay | 12:12 |
| asettle | git-harry: dude did you just say partay? | 12:12 |
| asettle | Wat | 12:12 |
| *** hwcomcn has quit IRC | 12:13 | |
| git-harry | asettle: I'm hip | 12:13 |
| *** hwcomcn has joined #openstack-ansible | 12:14 | |
| asettle | I'm actually speechless. | 12:14 |
| asettle | What else can you do git-harry ? | 12:14 |
| * git-harry wishes he'd worked out how to do that a couple of months ago | 12:14 | |
| asettle | HA | 12:14 |
| asettle | FUNNY | 12:15 |
| asettle | I don't have to speak to make your life difficult | 12:15 |
| * asettle puts fingerprints all over git-harry 's monitor | 12:15 | |
| * git-harry doesn't want to play any more | 12:15 | |
| asettle | I am victorious! | 12:16 |
| openstackgerrit | Merged openstack/openstack-ansible: [docs] Remove duplicated content https://review.openstack.org/348737 | 12:21 |
| *** woodard has joined #openstack-ansible | 12:22 | |
| *** woodard has quit IRC | 12:22 | |
| *** woodard has joined #openstack-ansible | 12:22 | |
| mhayden | odyssey4me: since i'm switching those python modules to packages, i should remove them from tox.ini and test-requirements.txt, right? | 12:29 |
| odyssey4me | yes | 12:31 |
| mhayden | okay | 12:31 |
| odyssey4me | considering that they're only needed for ansible | 12:31 |
| *** weshay_mtg is now known as weshay | 12:31 | |
| odyssey4me | the only exception would be in the integrated repo, where ansible is installed into a venv | 12:31 |
| mhayden | and those pkgs should be in other-requirements.txt/run_tests.sh in every IRR, correct? | 12:31 |
| odyssey4me | in that case we need those installed into the venv | 12:32 |
| odyssey4me | yes, every IRR | 12:32 |
| odyssey4me | let me work out the integrated repo bit - you work on the rest... happy witht hat? | 12:32 |
| mhayden | ah, i was just about to propose the integrated patch :P | 12:32 |
| odyssey4me | oh, then go ahead | 12:32 |
| *** v1k0d3n has joined #openstack-ansible | 12:33 | |
| *** Guest41883 has joined #openstack-ansible | 12:35 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Add SNI support via OS packages https://review.openstack.org/348860 | 12:36 |
| mhayden | odyssey4me: let me know if i'm off base ^^ | 12:36 |
| *** psilvad has joined #openstack-ansible | 12:36 | |
| mhayden | i *think* the gate has epel release already, but i'm not sure | 12:37 |
| odyssey4me | hmm, for the integrated repo we actually should have a other-requirements.txt file that's as empty as possible... the bootstrap script must install the things | 12:37 |
| odyssey4me | we must ensure that our deployment process preps the deployment host - no other system | 12:38 |
| mhayden | so should we just put these packages into bootstrap-ansible.sh ? | 12:38 |
| odyssey4me | yes, I think so | 12:38 |
| mhayden | okay, i'll fixy | 12:39 |
| odyssey4me | I'm not even sure why those other packages are there for the integrated repo, but a fix for that can follow on. | 12:39 |
| odyssey4me | the other-requirements.txt change is great for the IRR's. | 12:39 |
| *** Guest41883 has quit IRC | 12:40 | |
| mhayden | okay, i'll propose those in a sec | 12:40 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Add SNI support via OS packages https://review.openstack.org/348860 | 12:40 |
| mhayden | odyssey4me: ah, my capitalization fix still went in -- i'll remove that | 12:40 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Add SNI support via OS packages https://review.openstack.org/348860 | 12:41 |
| odyssey4me | is there any chance that we could perhaps take those long lines and break them somehow? | 12:41 |
| odyssey4me | it's getting out of hand | 12:41 |
| mhayden | i was just about to ask that :P | 12:41 |
| odyssey4me | not sure if a line break will work there? | 12:42 |
| odyssey4me | I guess it should | 12:42 |
| mhayden | should be able to backslash it | 12:42 |
| odyssey4me | yep | 12:42 |
| *** maddy__ has quit IRC | 12:42 | |
| odyssey4me | probably best to also put apt-get update && DEBIAN_FRONTEND=noninteractive into seperate lines | 12:42 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build: Provide ability to override upper-constraints https://review.openstack.org/348775 | 12:42 |
| odyssey4me | git-harry mancdaz ^ | 12:43 |
| odyssey4me | I'll do the backport requests now | 12:43 |
| mhayden | yeah backslashes work fine in the case | 12:45 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Add SNI support via OS packages https://review.openstack.org/348860 | 12:45 |
| git-harry | odyssey4me: great, thanks odyssey4me | 12:45 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Provide ability to override upper-constraints https://review.openstack.org/348865 | 12:46 |
| *** tlbr_ has joined #openstack-ansible | 12:48 | |
| *** abelur is now known as abelur|afk | 12:48 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Provide ability to override upper-constraints https://review.openstack.org/348869 | 12:50 |
| mhayden | odyssey4me: does this patch look reasonable for the IRR's? https://gist.github.com/major/b840c1c6c118528c3b528f021c004c5b | 12:50 |
| mhayden | i haven't done one of these IRR blasts before, so i'd like to not do something silly | 12:50 |
| mhayden | or sillier than i normally do | 12:50 |
| odyssey4me | mhayden :) I use this strategy | 12:51 |
| odyssey4me | 1 - make a patch for one repo, verify that it works and ideally get it merged so that the pattern is accepted | 12:51 |
| odyssey4me | 2 - figure out the automation to blast the same patch to all the repositories (except specs and sometimes also leaving out ops, plugins) | 12:52 |
| odyssey4me | also skip the integrated repo, obviously | 12:52 |
| odyssey4me | 3 - test the automated method against a single repo | 12:52 |
| *** tlbr_ has quit IRC | 12:52 | |
| odyssey4me | if it works, skip that repo too and blast | 12:52 |
| mhayden | alrighty | 12:52 |
| odyssey4me | LGTM though - bang it against one and let's see it pass and merge | 12:53 |
| mhayden | alrightysir | 12:53 |
| odyssey4me | mhayden boo: http://logs.openstack.org/60/348860/4/check/gate-openstack-ansible-dsvm-commit/c13cdc8/console.html#_2016-07-29_12_55_40_519786 | 12:57 |
| odyssey4me | that package is *only* available on xenial | 12:57 |
| mhayden | hmm | 12:57 |
| mhayden | okay, i'll break that out | 12:58 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Remove SSL protocol/cipher from AIO user_variables.yml https://review.openstack.org/348877 | 13:06 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Add SNI support via OS packages https://review.openstack.org/348860 | 13:07 |
| odyssey4me | backport review request: https://review.openstack.org/348865 | 13:08 |
| odyssey4me | :/ mhayden I think you got the wrong end of the stick | 13:09 |
| odyssey4me | for ubuntu 16.04 only python-ndg-httpsclient was the special case | 13:09 |
| odyssey4me | the others must also go onto 14.04 | 13:09 |
| mhayden | ah yes | 13:10 |
| odyssey4me | also, you're missing python-openssl | 13:11 |
| mhayden | odyssey4me: for ubuntu? | 13:11 |
| odyssey4me | yep | 13:12 |
| mhayden | i'll add it | 13:12 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible: Add SNI support via OS packages https://review.openstack.org/348860 | 13:13 |
| *** messy has joined #openstack-ansible | 13:18 | |
| *** tlbr has joined #openstack-ansible | 13:19 | |
| *** tlbr is now known as Guest28658 | 13:20 | |
| mrhillsman | g'morning | 13:20 |
| *** Guest28658 has quit IRC | 13:24 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Provide ability to override upper-constraints https://review.openstack.org/348869 | 13:26 |
| *** michaelgugino has joined #openstack-ansible | 13:26 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Provide ability to override upper-constraints https://review.openstack.org/348869 | 13:27 |
| odyssey4me | git-harry that was a git failure on my part, thanks for picking it up! | 13:27 |
| *** timl0101 has joined #openstack-ansible | 13:27 | |
| cloudnull | mornings | 13:29 |
| michaelgugino | morning | 13:32 |
| *** deadnull has joined #openstack-ansible | 13:33 | |
| *** KLevenstein has joined #openstack-ansible | 13:34 | |
| *** TxGirlGeek has joined #openstack-ansible | 13:35 | |
| *** janki has quit IRC | 13:35 | |
| *** eil397_ has joined #openstack-ansible | 13:37 | |
| automagically | o/ michaelgugino and cloudnull | 13:37 |
| michaelgugino | howdy | 13:38 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_container_create: Allow a deployer to disable container restarts https://review.openstack.org/348885 | 13:39 |
| odyssey4me | mancdaz ^ forward port for mitaka | 13:39 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-pip_install: Add SNI support via OS packages https://review.openstack.org/348887 | 13:40 |
| *** jamesdenton has joined #openstack-ansible | 13:40 | |
| mhayden | odyssey4me: let me know if i got the *right* end of the stick here ;) ^^ https://review.openstack.org/#/c/348887/ | 13:41 |
| odyssey4me | mancdaz I think that https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/roles/lxc_container_create/tasks/container_create.yml#L99-L111 will still result in unexpected restarts. I can prep a patch to fix that - do you agree with my assessment? | 13:41 |
| *** catintheroof has quit IRC | 13:42 | |
| cloudnull | hows it going today ? | 13:42 |
| odyssey4me | cloudnull ^ see my msg to mancdaz | 13:42 |
| odyssey4me | do yuo agree? | 13:43 |
| *** eil397_ has quit IRC | 13:43 | |
| odyssey4me | also, morning michaelgugino cloudnull | 13:43 |
| adreznec | Hey, does anyone know if there's a plan to support Ansible 2.0/2.1 in roles this cycle? Was writing up a change that passed locally because OSA bootstraps ansible 2.1, but roles in the gate are still using 1.9.4 | 13:43 |
| cloudnull | which message? | 13:44 |
| adreznec | If not I'll rewrite my change to be 1.9.4 compliant | 13:44 |
| cloudnull | for u-c ? | 13:44 |
| odyssey4me | cloudnull I think that https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/roles/lxc_container_create/tasks/container_create.yml#L99-L111 will still result in unexpected restarts. I can prep a patch to fix that - do you agree with my assessment? | 13:44 |
| odyssey4me | adreznec we ideally want to maintain compatibility this cycle to help people transition | 13:44 |
| adreznec | odyssey4me: Ok, sounds good, should be an easy fix | 13:44 |
| cloudnull | odyssey4me: I can see it restarting a container but that ones a little tricky. specifically those mount points must exist or else other things will die in a fire. | 13:46 |
| cloudnull | 's/that/those/g' | 13:46 |
| odyssey4me | cloudnull yeah, so I was thinking to do the container commands in one step before the config changes | 13:47 |
| odyssey4me | the config changes will have to use lineinfile | 13:47 |
| cloudnull | but if a deployer disables the restart those mount points wont be in effect | 13:47 |
| cloudnull | that is until the container restarts | 13:47 |
| cloudnull | which may create different data in the same place because of the mount. | 13:48 |
| odyssey4me | well, in theory that doesn't matter for upgrades - because the mounts are already there | 13:48 |
| cloudnull | true. | 13:48 |
| odyssey4me | and in a new deployment the restart will not be stopped and be hit by the notifier | 13:48 |
| * cloudnull just playing devils advocate | 13:48 | |
| odyssey4me | so basically it's the same level of control | 13:48 |
| odyssey4me | I guess it can stay as-is, because the control was added specifically for upgrades | 13:49 |
| cloudnull | no, if the playbook run stops the notifyer wont be triggered. | 13:49 |
| odyssey4me | and that task will not cause a restart for upgrades | 13:49 |
| cloudnull | if the deployer runs the play again the notifyer will never run | 13:49 |
| cloudnull | and the container will not have those options loaded | 13:49 |
| odyssey4me | yeah, that's an issue and I think we could solve that one using local facts | 13:49 |
| odyssey4me | evrardjp actually did a blog post on it I think | 13:50 |
| evrardjp | nope I didn't but you reminds me a good subject :d | 13:50 |
| cloudnull | we could drop a file and trigger a notice if the file exists indicating that the container should be restarted ? | 13:50 |
| odyssey4me | evrardjp I thought you did some sort of post/test for notifier trigger methods | 13:51 |
| evrardjp | that's the idea | 13:51 |
| cloudnull | and clean up that file upon restart | 13:51 |
| evrardjp | I did, but not 100% this use case | 13:51 |
| odyssey4me | cloudnull yes, exactly | 13:51 |
| evrardjp | let me give you a link | 13:51 |
| evrardjp | http://evrard.me/orchestrate-a-service-restart-during-a-maintenance-window-with-ansible.html | 13:51 |
| evrardjp | it's not high level stuff | 13:52 |
| evrardjp | just explaining an opinion | 13:52 |
| odyssey4me | yep, I think it's something that would be an excellent pattern to use to improve control and ensure that things happen even if the play fails | 13:52 |
| evrardjp | I will write something with the local facts | 13:52 |
| odyssey4me | another way is obviously to set a fact and rely on the fact cache, but it doesn't live long | 13:53 |
| odyssey4me | so I think that dropping a local fact might be a better way of handling it | 13:53 |
| evrardjp | something like "introducing checkpoint feature in ansible playbook runs" | 13:53 |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-os_nova: Implement nova-lxd driver https://review.openstack.org/346984 | 13:54 |
| *** severion has joined #openstack-ansible | 13:57 | |
| *** spotz_zzz is now known as spotz | 13:57 | |
| *** v1k0d3n_ has joined #openstack-ansible | 14:02 | |
| *** v1k0d3n_ has quit IRC | 14:02 | |
| odyssey4me | mancdaz I see a flaw in the upgrade process for liberty, unfortunately - I don't see how any upgrade tests could have worked :/ | 14:02 |
| odyssey4me | in https://github.com/openstack/openstack-ansible/blob/liberty/scripts/run-upgrade.sh#L144 the repo containers are deleted | 14:02 |
| *** v1k0d3n_ has joined #openstack-ansible | 14:02 | |
| odyssey4me | but they are never re-created | 14:03 |
| odyssey4me | that changed in your patch :/ https://review.openstack.org/#/c/347195/8/scripts/run-upgrade.sh | 14:03 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-pip_install: Add SNI support via OS packages https://review.openstack.org/348887 | 14:03 |
| *** abelur|afk has quit IRC | 14:03 | |
| *** severion has quit IRC | 14:04 | |
| odyssey4me | git-harry ^ see above (notes to mancdaz) | 14:05 |
| *** Mudpuppy has joined #openstack-ansible | 14:05 | |
| mhayden | odyssey4me: might have stumbled upon a bindep bug -> https://gist.github.com/major/dfaa0ce8d50d284d81afe5ff6735e07d | 14:07 |
| mhayden | tonyb is helping me look at that | 14:07 |
| odyssey4me | haha, whoops | 14:07 |
| *** cloudtrainme has joined #openstack-ansible | 14:09 | |
| *** hwcomcn has quit IRC | 14:16 | |
| openstackgerrit | Miguel Alex Cantu (alextricity25) proposed openstack/openstack-ansible: Added docs for removing compute host https://review.openstack.org/340458 | 14:18 |
| git-harry | odyssey4me: are you saying it was broken or that the mariadb patch has broken it? | 14:30 |
| odyssey4me | git-harry the mariadb patch has broken it | 14:30 |
| git-harry | odyssey4me: hmmmm, mancdaz has run a load of tests so this should work. Let me look. | 14:31 |
| odyssey4me | I expect that it should be an easy solution - just add another command to create them | 14:31 |
| odyssey4me | but yeah, I was confused as to how the tests had passed | 14:31 |
| *** asettle has quit IRC | 14:32 | |
| git-harry | odyssey4me: setup-hosts.yml | 14:33 |
| git-harry | that includes the create playbook | 14:33 |
| *** tlbr_ has joined #openstack-ansible | 14:33 | |
| *** tlbr_ has quit IRC | 14:34 | |
| odyssey4me | ah, it's setup-hosts !galera_all | 14:34 |
| odyssey4me | I missed the ! | 14:34 |
| odyssey4me | yeah, that takes care of it - makes sense now | 14:34 |
| odyssey4me | sorry for the false alarm | 14:34 |
| git-harry | odyssey4me: not a problem, the fear of letting a bug in is better than caffeine for keeping alert on a Friday afternoon. | 14:37 |
| odyssey4me | heh, yeah | 14:37 |
| odyssey4me | mhayden backport review please https://review.openstack.org/348865 | 14:38 |
| odyssey4me | also https://review.openstack.org/348869 | 14:39 |
| *** shausy has quit IRC | 14:40 | |
| *** Guest31932 has quit IRC | 14:42 | |
| *** david-lyle has joined #openstack-ansible | 14:42 | |
| *** david-lyle_ has joined #openstack-ansible | 14:46 | |
| ioni | can you guys merge https://git.openstack.org/cgit/openstack/openstack-ansible-os_cinder/commit/?id=042a2c2e57bd3dda20bfa7257a5f2c7c3d5db95f into stable? | 14:47 |
| ioni | if possible :D | 14:47 |
| *** david-lyle has quit IRC | 14:47 | |
| mhayden | odyssey4me: looking | 14:48 |
| *** david-lyle__ has joined #openstack-ansible | 14:48 | |
| mhayden | ioni: i'll take a look at that one shortly | 14:48 |
| *** tlbr_ has joined #openstack-ansible | 14:50 | |
| odyssey4me | ioni if you have signed the CLA, then you can propose the backport from gerrit: https://review.openstack.org/#/c/340023/ | 14:50 |
| odyssey4me | click 'cherry pick' and enter 'stable/mitaka' | 14:51 |
| odyssey4me | if you don't, let me know and I can propose it for you | 14:51 |
| odyssey4me | assuming you want it in stable/mitaka? | 14:51 |
| *** david-lyle_ has quit IRC | 14:52 | |
| *** tlbr_ has quit IRC | 14:53 | |
| odyssey4me | anyone have a moment for a second opinion on this backport? https://review.openstack.org/348865 | 14:53 |
| alextricity25 | ^^That'll be a good one for liberty too | 14:55 |
| odyssey4me | alextricity25 you're welcome to propose backports | 14:57 |
| alextricity25 | odyssey4me: Will do! | 14:57 |
| *** v1k0d3n has quit IRC | 15:03 | |
| *** v1k0d3n has joined #openstack-ansible | 15:05 | |
| *** TxGirlGeek has quit IRC | 15:06 | |
| *** TxGirlGeek has joined #openstack-ansible | 15:07 | |
| *** elopez has joined #openstack-ansible | 15:09 | |
| *** elopez is now known as Guest21368 | 15:10 | |
| *** klamath has joined #openstack-ansible | 15:10 | |
| *** klamath has quit IRC | 15:11 | |
| *** klamath has joined #openstack-ansible | 15:11 | |
| *** psilvad has quit IRC | 15:18 | |
| *** galstrom_zzz is now known as galstrom | 15:20 | |
| *** david-lyle has joined #openstack-ansible | 15:23 | |
| *** david-lyle__ has quit IRC | 15:23 | |
| *** admin0 has quit IRC | 15:24 | |
| *** weshay is now known as weshay_bbiab | 15:25 | |
| *** eil397_ has joined #openstack-ansible | 15:27 | |
| *** psilvad has joined #openstack-ansible | 15:27 | |
| *** eil397 has joined #openstack-ansible | 15:27 | |
| eil397 | good morning | 15:28 |
| *** TxGirlGeek has quit IRC | 15:29 | |
| *** Guest79767 has joined #openstack-ansible | 15:29 | |
| *** TxGirlGeek has joined #openstack-ansible | 15:29 | |
| *** Guest79767 has quit IRC | 15:34 | |
| *** sdake has joined #openstack-ansible | 15:39 | |
| *** tlbr_ has joined #openstack-ansible | 15:40 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Only build wheels for services which are deployed https://review.openstack.org/344912 | 15:41 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build: Only clone git repositories for services which are deployed https://review.openstack.org/345642 | 15:41 |
| alextricity25 | Hey guys. I don't know too much about how the gating jobs are run, but could there be anything going on that might effect the way the pip.conf is dropped in? I'm working on this change that works locally, but fails gating when trying to install pip packages on the repo server | 15:42 |
| alextricity25 | https://review.openstack.org/#/c/348533/1 | 15:42 |
| odyssey4me | hmm, so the gating job does copy a global pip conf into the containers - but a use conf will override that | 15:44 |
| odyssey4me | lemme take a look | 15:44 |
| eil397 | alextricity25: openstack ci changes and prepared pip conf | 15:44 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-pip_install: Add SNI support via OS packages https://review.openstack.org/348887 | 15:44 |
| *** tlbr_ has quit IRC | 15:45 | |
| eil397 | alextricity25: but yes. openstack ci should not affect pip configuration in containers | 15:45 |
| alextricity25 | eil397 That's for the host, right? Or does that prepare the pip conf in containers? | 15:45 |
| eil397 | ansible manages it | 15:45 |
| alextricity25 | oh you answered it before I can ask | 15:45 |
| alextricity25 | :) | 15:45 |
| alextricity25 | Right...hmm...that's interesting. I'm curious as to why the mySQL-python pip package can't be installed on the gating nodes then | 15:46 |
| eil397 | alextricity25: : ) yes. openstack-ansible should prepare and fully reposponse for pip conf . especially on repo serv er : - ) | 15:46 |
| alextricity25 | I'll do some verbosity outputs to see what's going on. | 15:46 |
| odyssey4me | alextricity25 the issue is more convoluted than all that | 15:47 |
| openstackgerrit | Merged openstack/openstack-ansible-plugins: Update py_pkgs to set options group membership and itemise global pins https://review.openstack.org/347930 | 15:47 |
| alextricity25 | odyssey4me I"m glad i'm not the only one who thinks so :) | 15:47 |
| odyssey4me | we did a pretty major overhaul of how the repo build happens - in liberty it was too complicated and unpredictable | 15:47 |
| odyssey4me | but backporting changes like this introduces quite a bit more risk because of the mesh of deps | 15:48 |
| odyssey4me | this patch will not be so simple | 15:48 |
| odyssey4me | the repo container is not meant to have a pip conf at all because it shouldn't be locked down | 15:48 |
| odyssey4me | you can see here that it's being locked down: http://logs.openstack.org/33/348533/1/check/gate-openstack-ansible-dsvm-commit/4e1cfbc/console.html#_2016-07-28_22_57_32_544048 | 15:49 |
| * mhayden overhaul | 15:49 | |
| *** deadnull is now known as _deadnull | 15:49 | |
| eil397 | alextricity25: i'm not how "pip_no_index: false" should work in list of dependecies | 15:50 |
| mhayden | kudos to tonyb for proposing a fix for the bindep underscore shenanigans! https://review.openstack.org/348901 | 15:50 |
| odyssey4me | yeah, due to the dep chain that var will have to be applied somewhere else | 15:50 |
| alextricity25 | odyssey4me Shouldn't the pip_no_index: false here bypass that?https://review.openstack.org/#/c/348533/1/playbooks/roles/repo_build/meta/main.yml | 15:50 |
| odyssey4me | lemme try something | 15:50 |
| alextricity25 | eil397: I was kindof curious about that too | 15:51 |
| *** mummer has joined #openstack-ansible | 15:52 | |
| odyssey4me | ugh, the galera_client meta dep on pip_lock_down sucks | 15:53 |
| odyssey4me | hmf | 15:53 |
| *** TxGirlGeek has quit IRC | 15:54 | |
| *** TxGirlGeek has joined #openstack-ansible | 15:54 | |
| *** eil397 has quit IRC | 15:54 | |
| *** Guest21368 has quit IRC | 15:55 | |
| michaelgugino | yes, pip lockdown sucks | 15:55 |
| michaelgugino | should only be called as part of host setup, should not be dependency of every role. | 15:55 |
| michaelgugino | or any role (I don't know all the places it's called) | 15:55 |
| *** eil397_ has quit IRC | 15:56 | |
| *** TxGirlGeek has quit IRC | 15:56 | |
| *** TxGirlGeek has joined #openstack-ansible | 15:57 | |
| *** eil397_ has joined #openstack-ansible | 15:57 | |
| odyssey4me | jmccrory automagically cloudnull d34dh0r53 stevelle mattt hughsaunders andymccr evrardjp looking for backport reviews urgently please https://review.openstack.org/348869 / https://review.openstack.org/348865 | 15:58 |
| odyssey4me | thanks automagically | 16:00 |
| *** pcaruana has quit IRC | 16:01 | |
| odyssey4me | alextricity25 so I'm wondering | 16:02 |
| cloudnull | odyssey4me: looking | 16:02 |
| cloudnull | opps. too late | 16:02 |
| odyssey4me | the purpose of having the role dep is to ensure that the built wheels have the right libs to dep on right? | 16:02 |
| alextricity25 | odyssey4me: what if we just get rid of that pip_lock_down dependency in galera_client | 16:02 |
| alextricity25 | odyssey4me. Yes | 16:02 |
| odyssey4me | alextricity25 that is not a smart move in a stable branch | 16:02 |
| *** TxGirlGeek has quit IRC | 16:02 | |
| *** cloudtrainme has quit IRC | 16:03 | |
| odyssey4me | alextricity25 ok, so why not just add the role execution to https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/repo-build.yml | 16:03 |
| odyssey4me | with that var | 16:03 |
| odyssey4me | 1st - it will only get installed on repo[0] and that's the only place it's needed | 16:04 |
| odyssey4me | 2nd - the var override in the playbook will have a higher precedence | 16:04 |
| alextricity25 | odyssey4me: Yup! That's a good idea | 16:04 |
| alextricity25 | i mean...if you guys are okay with that inconsistency | 16:04 |
| odyssey4me | what is a bit odd though, is that the wheel shouldn't have to build because the wheels have been built before on rpc-repo and infra... so I am curious about why this is an issue in the first place | 16:05 |
| alextricity25 | odyssey4me: all of the other roles seem to have galera_client as a dependency in meta/main.yml | 16:05 |
| *** TxGirlGeek has joined #openstack-ansible | 16:06 | |
| odyssey4me | alextricity25 yep, but there is no need to do that here | 16:06 |
| alextricity25 | odyssey4me: You're right. The wheels will build just fine | 16:06 |
| alextricity25 | odyssey4me: The only problem here is when someone re-builds the wheels | 16:06 |
| odyssey4me | I'm not quite sure what difference that makes, unless the build is forced to happen from source | 16:07 |
| *** TxGirlGeek has quit IRC | 16:08 | |
| odyssey4me | but meh - try the playbook route | 16:08 |
| *** TxGirlGeek has joined #openstack-ansible | 16:08 | |
| odyssey4me | it might even be better to just add some tasks to install the mariadb packages on the repo container... it is just a couple of packages | 16:09 |
| *** mikelk has quit IRC | 16:09 | |
| odyssey4me | it achieves the goal - the galera_client/pip_lock_down chain is overkill to solve this particular problem | 16:10 |
| alextricity25 | You're right. Unless there is hard dependency as to what galera_client is installing | 16:12 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-pip_install: Add SNI support via OS packages https://review.openstack.org/348887 | 16:12 |
| *** spotz is now known as spotz_zzz | 16:12 | |
| odyssey4me | alextricity25 try the playbook execution first, hopefully that works | 16:13 |
| alextricity25 | trying now | 16:13 |
| alextricity25 | odyssey4me: I thought the wheels were built from source in liberty? You mentioned something earlier that they are cloned from rpc-repo? | 16:17 |
| odyssey4me | alextricity25 the repo build role executes the build here: https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/roles/repo_build/tasks/repo_build.yml#L16 | 16:18 |
| odyssey4me | alextricity25 notice https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/roles/repo_build/tasks/repo_build.yml#L23-L26 | 16:18 |
| odyssey4me | an index, and extra index are set | 16:18 |
| odyssey4me | that is where pip looks for sources to build the wheels | 16:19 |
| *** TxGirlGeek has quit IRC | 16:19 | |
| odyssey4me | the defaults index is https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/inventory/group_vars/hosts.yml#L29 | 16:19 |
| alextricity25 | odyssey4me: Ohhh that makes sense. I see here that these values are set to pypi in liberty. What part does rpc-repo play in liberty? | 16:20 |
| odyssey4me | hmm, actually that index isn't used | 16:21 |
| odyssey4me | this is part of the nonsense we cleaned up for mitaka | 16:21 |
| alextricity25 | The reason I ask is because I've been under the impression that we dealt away with rpc-repo i liberty | 16:21 |
| odyssey4me | the index and extra index are here: https://github.com/openstack/openstack-ansible/blob/liberty/playbooks/repo-build.yml#L41-L42 | 16:21 |
| odyssey4me | so when pip does the wheel build it will first try to source a wheel that's already build | 16:22 |
| alextricity25 | odyssey4me: those damn playbook variable overrides | 16:22 |
| alextricity25 | get me every time | 16:22 |
| odyssey4me | it will consult the index and all extra indexes | 16:22 |
| odyssey4me | if it finds a built wheel on a source that's already built, it will simply download it | 16:22 |
| odyssey4me | otherwise it will download the sdist source from pypi and compile it | 16:23 |
| odyssey4me | unfortunately rpc-repo does need to be up for liberty because it's the default index... | 16:23 |
| odyssey4me | for mitaka we switched them around and also made the extra indexes optional | 16:23 |
| *** weshay_bbiab is now known as weshay | 16:24 | |
| odyssey4me | unfortunately cleaning this up in liberty is too risky - it's a stable branch and has been installed and built that way for ages | 16:24 |
| odyssey4me | so changing it now risks the stability | 16:24 |
| alextricity25 | odyssey4me: I see now. | 16:24 |
| alextricity25 | Thanks for the explanations odyssey4me :) | 16:25 |
| odyssey4me | you will notice that it's cleaned up here: https://github.com/openstack/openstack-ansible/blob/stable/mitaka/playbooks/repo-build.yml#L43 | 16:25 |
| alextricity25 | OH! odyssey4me...this is interesting | 16:26 |
| alextricity25 | Ignoring dnspython3: markers u"python_version=='3.4'" don't match your environment | 16:26 |
| odyssey4me | nope, that's no problem | 16:26 |
| odyssey4me | the requirements file has markers for python version specific packages | 16:26 |
| odyssey4me | pip knows how to deal with those appropriately | 16:26 |
| odyssey4me | the requirements for liberty come from https://github.com/openstack/requirements/blob/stable/liberty/global-requirements.txt | 16:27 |
| odyssey4me | see the markers: https://github.com/openstack/requirements/blob/stable/liberty/global-requirements.txt#L37-L38 | 16:27 |
| *** cloudtrainme has joined #openstack-ansible | 16:27 | |
| *** Guest30197 has joined #openstack-ansible | 16:28 | |
| *** Mudpuppy has quit IRC | 16:29 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build: Provide ability to override upper-constraints https://review.openstack.org/348865 | 16:29 |
| *** permalac has quit IRC | 16:31 | |
| alextricity25 | odyssey4me: Ah ..so the real problem is hereNo matching distribution found for ansible-lint<=2.3.6,>=2.0.3 (from -c /var/www/repo/os-releases/12.2.0/requirements_constraints.txt (line 24)) | 16:31 |
| *** jamielennox|away is now known as jamielennox | 16:32 | |
| alextricity25 | stdout: Ignoring indexes: https://rpc-repo.rackspace.com/pools,https://pypi.python.org/simple | 16:33 |
| odyssey4me | yep, it's locked down to the repo | 16:33 |
| *** Guest30197 has quit IRC | 16:33 | |
| odyssey4me | in the pip.conf, is no-index = True or False? | 16:34 |
| odyssey4me | hang on - no index for ansible-lint? | 16:34 |
| odyssey4me | that must be on the deployment host? | 16:34 |
| odyssey4me | not the repo container | 16:34 |
| alextricity25 | No it's on the repo container: http://paste.openstack.org/show/543931/ | 16:35 |
| alextricity25 | hold on..maybe I rebuilt the repo container incorrectly | 16:35 |
| odyssey4me | ah of course - from the constraints | 16:36 |
| odyssey4me | so the issue there is that the repo container is locked down to only use itself, which will never work | 16:36 |
| alextricity25 | odyssey4me: Even with pip_no_index: false? | 16:37 |
| alextricity25 | odyssey4me: Oh you know what..I had a typo on in the repo_build play when I made the changes | 16:37 |
| alextricity25 | I had pip_lock_down: false, not pip_no_index: false | 16:37 |
| * alextricity25 facepalms | 16:38 | |
| alextricity25 | maybe it's time for break :) | 16:38 |
| odyssey4me | heh | 16:38 |
| alextricity25 | I'm off to lunch. odyssey4me Thanks for the assist :) This repo stuff can be complicated | 16:38 |
| odyssey4me | does pip expect any specific capitalisation in the file? | 16:38 |
| odyssey4me | you may have to force it to be a string | 16:38 |
| odyssey4me | ie pip_no_index: "false" | 16:39 |
| alextricity25 | all the other examples of pip_no_index are lowercase | 16:39 |
| *** tlbr_ has joined #openstack-ansible | 16:40 | |
| ngraf | could someone review https://review.openstack.org/#/c/347564/ and https://review.openstack.org/#/c/348330/? They are two related back port changes | 16:40 |
| odyssey4me | alextricity25 just by the way, I think I just realised what the real issue is... considering that the wheel most likely is actually being pulled from rpc-repo, rpc-repo needs the right libs installed so that when it compiles the wheels it does it right | 16:42 |
| *** tlbr_ has quit IRC | 16:44 | |
| *** _deadnull is now known as deadnull | 16:44 | |
| *** Oku_OS is now known as Oku_OS-away | 16:44 | |
| alextricity25 | odyssey4me: What about pipy? Doesn't pip also look at that repo because we have it defined as "repo_build_pip_extra_index"? | 16:45 |
| alextricity25 | pypi* | 16:45 |
| *** michaelgugino has quit IRC | 16:45 | |
| odyssey4me | alextricity25 yep, pypi will most often only have the source | 16:46 |
| openstackgerrit | Merged openstack/openstack-ansible: Add SNI support via OS packages https://review.openstack.org/348860 | 16:47 |
| odyssey4me | my theory only counts *if* rpc-repo has the appropriately versioned wheel pre-built | 16:47 |
| odyssey4me | so changing the repo-build play is still a good thing, but rpc-repo must also be fixed | 16:47 |
| odyssey4me | ngraf reviewed | 16:47 |
| *** wadeholl_ has quit IRC | 16:49 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat: Use keystone library for heat_stack_domain_admin https://review.openstack.org/348964 | 16:49 |
| *** spotz_zzz is now known as spotz | 16:52 | |
| *** wadeholler has joined #openstack-ansible | 16:53 | |
| *** elopez has joined #openstack-ansible | 16:53 | |
| odyssey4me | ngraf reviewed | 16:53 |
| odyssey4me | a twek on one, and we need to merge another patch to prevent a regression | 16:53 |
| odyssey4me | *tweak | 16:53 |
| *** elopez is now known as Guest10600 | 16:54 | |
| ngraf | odyssey4me: Thank you. I'll get that tweak submitted | 16:54 |
| *** TxGirlGeek has joined #openstack-ansible | 16:59 | |
| *** cloudtrainme has quit IRC | 17:01 | |
| *** Guest10600 has quit IRC | 17:02 | |
| *** jiteka has quit IRC | 17:03 | |
| *** elopez_ has joined #openstack-ansible | 17:06 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_cloudkitty: Implement Multi-OS support in os_cloudkitty https://review.openstack.org/346952 | 17:06 |
| odyssey4me | hmm, cloudnull ping | 17:07 |
| *** severion has joined #openstack-ansible | 17:12 | |
| *** Mudpuppy has joined #openstack-ansible | 17:14 | |
| *** v1k0d3n_ has quit IRC | 17:16 | |
| *** cloudtrainme has joined #openstack-ansible | 17:17 | |
| openstackgerrit | Merged openstack/openstack-ansible: Provide ability to override upper-constraints https://review.openstack.org/348869 | 17:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Add SNI support via OS/python packages https://review.openstack.org/348970 | 17:18 |
| *** adrian_otto has joined #openstack-ansible | 17:19 | |
| odyssey4me | mhayden ^ I *think* that might be required to make this all work properly. In your previous patch we didn't put the right packages into the venv... I couldn't think of a better way of ensuring that we have the right bits in place on the host and in the venv. | 17:20 |
| mhayden | odyssey4me: yeah, that's where i get confused :| | 17:22 |
| cloudnull | odyssey4me: pong | 17:22 |
| cloudnull | sorry was in a meeting. | 17:22 |
| jmccrory | odyssey4me mhayden : should those be added to the base container? | 17:24 |
| mhayden | jmccrory: it's not needed inside the containers -- just needs to be where ansible is being run | 17:25 |
| jmccrory | ah ok | 17:25 |
| *** elopez_ has quit IRC | 17:25 | |
| *** tlbr_ has joined #openstack-ansible | 17:25 | |
| *** cloudtrainme has quit IRC | 17:27 | |
| *** rstarmer has joined #openstack-ansible | 17:27 | |
| odyssey4me | jmccrory yeah, ansible uses httlib2 and not requests... silly | 17:29 |
| mhayden | odyssey4me: i think it uses requests in 2.x, right? | 17:29 |
| odyssey4me | mhayden does it? | 17:30 |
| *** tlbr_ has quit IRC | 17:30 | |
| *** severion has quit IRC | 17:33 | |
| *** itlinux has quit IRC | 17:35 | |
| *** jamielennox is now known as jamielennox|away | 17:35 | |
| *** kstev has quit IRC | 17:39 | |
| mhayden | that's what i read | 17:41 |
| mhayden | i'll go back and check | 17:41 |
| openstackgerrit | Miguel Alex Cantu (alextricity25) proposed openstack/openstack-ansible: Add the galera_client dependency to the repo build https://review.openstack.org/348533 | 17:42 |
| openstackgerrit | Miguel Alex Cantu (alextricity25) proposed openstack/openstack-ansible: Add the galera_client dependency to the repo build https://review.openstack.org/348533 | 17:43 |
| mhayden | my weekly gif about odyssey4me is complete: https://twitter.com/majorhayden/status/759081214183473154 | 17:43 |
| alextricity25 | hah | 17:44 |
| *** rstarmer has quit IRC | 17:44 | |
| *** rstarmer has joined #openstack-ansible | 17:46 | |
| odyssey4me | lol, wow dude | 17:48 |
| odyssey4me | mhayden if that's the case - then we only need to make changes to the IRR's and to the stable branches | 17:49 |
| *** sguduru has joined #openstack-ansible | 17:50 | |
| *** rstarmer has quit IRC | 17:50 | |
| *** rstarmer has joined #openstack-ansible | 17:51 | |
| mhayden | i'll go back and check | 17:52 |
| mhayden | hopefully the gif was accurate :P | 17:52 |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_aodh: Remove deprecated oslo messaging options https://review.openstack.org/348988 | 17:54 |
| *** albertcard has joined #openstack-ansible | 17:57 | |
| *** rstarmer has quit IRC | 17:57 | |
| *** rstarmer has joined #openstack-ansible | 17:57 | |
| *** TxGirlGeek has quit IRC | 17:58 | |
| *** cloudtrainme has joined #openstack-ansible | 17:58 | |
| *** elopez_ has joined #openstack-ansible | 18:01 | |
| *** elopez_ has quit IRC | 18:01 | |
| *** cloudtrainme has quit IRC | 18:07 | |
| *** cbits has joined #openstack-ansible | 18:07 | |
| stevelle | automagically: I'll get another patch set up to remove that real soon, but I'm busy on another branch on aodh right now fixing something else I found | 18:10 |
| automagically | np | 18:11 |
| automagically | I appreciate that | 18:11 |
| ioni | odyssey4me, i tried, it said Cherry pick failed: merge conflict :( | 18:16 |
| *** mummer has quit IRC | 18:16 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_container_create: Implement configurable start delay for containers https://review.openstack.org/349003 | 18:17 |
| *** rstarmer has quit IRC | 18:17 | |
| odyssey4me | ioni which patch was that? | 18:18 |
| ioni | odyssey4me, https://review.openstack.org/#/c/340023/ | 18:18 |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-os_nova: Implement nova-lxd driver https://review.openstack.org/346984 | 18:19 |
| *** cloudtrainme has joined #openstack-ansible | 18:20 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder: Fix cinder role condition when no cinder_volumes https://review.openstack.org/349004 | 18:21 |
| odyssey4me | ioni ^ | 18:22 |
| ioni | odyssey4me, thanks, care to tell me how did you do it? | 18:22 |
| ioni | is it a new commit | 18:22 |
| ioni | ? | 18:22 |
| odyssey4me | ioni via git | 18:22 |
| *** sguduru_ has joined #openstack-ansible | 18:23 | |
| odyssey4me | ioni in the repo: git checkout stable/mitaka | 18:23 |
| odyssey4me | git review -X <review ID> | 18:23 |
| odyssey4me | the cherry pick fails, so fix it up to ensure that the same changes are made | 18:23 |
| odyssey4me | git add the file | 18:23 |
| odyssey4me | git cherry-pick --continue | 18:23 |
| odyssey4me | complete the commit, and 'git review' | 18:23 |
| *** severion has joined #openstack-ansible | 18:24 | |
| *** sguduru has quit IRC | 18:24 | |
| ioni | thank you | 18:25 |
| ioni | nice to see Implement nova-lxd driver | 18:25 |
| *** v1k0d3n_ has joined #openstack-ansible | 18:25 | |
| *** mummer has joined #openstack-ansible | 18:27 | |
| ioni | i don't see a new group added like compute_hosts_lxd ? | 18:27 |
| ioni | how do you define on which compute nodes to install lxd driver? | 18:27 |
| *** rstarmer has joined #openstack-ansible | 18:28 | |
| *** electrofelix has quit IRC | 18:28 | |
| *** severion has quit IRC | 18:28 | |
| *** v1k0d3n_ has quit IRC | 18:30 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_neutron: Fix bug in RPC config that broke Rabbit SSL support https://review.openstack.org/349010 | 18:31 |
| odyssey4me | cloudnull you should probably revisit your vote on https://review.openstack.org/345642 & https://review.openstack.org/344912 | 18:32 |
| odyssey4me | they're passing now | 18:32 |
| odyssey4me | ioni it's a work in progress, but I believe that michaelgugino is presenting it at the summit | 18:33 |
| automagically | odyssey4me: Will need to make the same/similar changes as https://review.openstack.org/#/c/349010/ to many different roles, none of which currently have ssl enabled for rabbit. Any preference on how that gets submitted? | 18:33 |
| ioni | odyssey4me, cool, looking forward to move away from openvz :) | 18:34 |
| odyssey4me | you'd likely set the virt_type as a host var at this point, or you could create a group of hosts and apply it there if you like | 18:34 |
| *** thorst has quit IRC | 18:34 | |
| odyssey4me | automagically ah, I see you moved the section - good | 18:35 |
| odyssey4me | that should do the trick at first blush | 18:36 |
| odyssey4me | automagically I would think they just need to be submitted - with the same topic ideally | 18:37 |
| odyssey4me | not sure what you mean by 'how' ? | 18:37 |
| odyssey4me | afk for a bit while I go and fetch some dinner | 18:37 |
| automagically | Just concerned about gate results on the role tests not being particularly meaningful | 18:37 |
| automagically | That’s all | 18:37 |
| automagically | Will line em up and submit it | 18:37 |
| *** thorst has joined #openstack-ansible | 18:37 | |
| odyssey4me | ah | 18:37 |
| odyssey4me | ok, so right now ssl is disabled in group_vars so the connection is only using the transport url anyway | 18:38 |
| odyssey4me | if you do all the role bits, then the integrated repo bit and make sure it depends-on all the role patches then it won't merge until they all merge | 18:38 |
| odyssey4me | I would focus on the integrated roles first | 18:39 |
| *** deadnull has quit IRC | 18:39 | |
| automagically | Makes sense, thx. Enjoy your dinner | 18:40 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_zaqar: Test zaqar role using the zaqar tempest plugin https://review.openstack.org/348475 | 18:41 |
| *** thorst has quit IRC | 18:42 | |
| *** elopez_ has joined #openstack-ansible | 18:44 | |
| jmccrory | sigmavirus: do you know if https://github.com/sigmavirus24/openstack-ansible-searchlight is in a working state? | 18:44 |
| *** sdake has quit IRC | 18:47 | |
| *** mummer has quit IRC | 18:48 | |
| odyssey4me | jmccrory heh, when I last asked the answer was no | 18:48 |
| odyssey4me | it was built a very long time ago - back in the kilo cycle I think | 18:49 |
| odyssey4me | it may or may not be worth just starting from scratch | 18:49 |
| openstackgerrit | Nate Graf proposed openstack/openstack-ansible: Support users without projects in keystone library https://review.openstack.org/347564 | 18:49 |
| odyssey4me | I believe that nova will require searchlight for cells v2 though, so we will have to pick that up next cycle at the latest. | 18:50 |
| jmccrory | odyssey4me: hmm alright, well i'd be interested in helping get that role ready either way. think it'd be a really useful service to have | 18:51 |
| odyssey4me | excellent, if you inspect it and think it's a useful base then I can pull it into the openstack repositories next week | 18:51 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_nova: Fix bug in RPC config that broke Rabbit SSL support https://review.openstack.org/349024 | 18:52 |
| jmccrory | cool, sounds good | 18:52 |
| odyssey4me | you can, of course, fork it so long and make some progress - then I'll just suck it in from your fork | 18:52 |
| odyssey4me | ok, I'm out for the w/end | 18:53 |
| odyssey4me | cheers all - have a great w/end! | 18:53 |
| automagically | you too odyssey4me | 18:54 |
| jmccrory | see ya | 18:54 |
| prometheanfire | odyssey4me: thanks for using protection +1 | 18:56 |
| *** thorst has joined #openstack-ansible | 18:56 | |
| * mhayden pours more legos on the treadmill | 18:56 | |
| mhayden | prometheanfire: https://twitter.com/majorhayden/status/759081214183473154 | 18:56 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_glance: Fix bug in RPC config that broke Rabbit SSL support https://review.openstack.org/349028 | 18:57 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_nova: Fix bug in RPC config that broke Rabbit SSL support https://review.openstack.org/349024 | 18:58 |
| sigmavirus | jmccrory: I never finished it so the answer is most likely "no" | 18:59 |
| *** adrian_otto has quit IRC | 18:59 | |
| *** tlbr has joined #openstack-ansible | 19:00 | |
| *** tlbr is now known as Guest84124 | 19:00 | |
| openstackgerrit | Merged openstack/openstack-ansible-galera_client: Add ability to change apt/yum package state https://review.openstack.org/346335 | 19:00 |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_aodh: Update testing script for xenial https://review.openstack.org/349031 | 19:00 |
| *** thorst has quit IRC | 19:00 | |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_aodh: Remove deprecated oslo messaging options https://review.openstack.org/348988 | 19:01 |
| *** thorst has joined #openstack-ansible | 19:01 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_keystone: Fix bug in RPC config that broke Rabbit SSL support https://review.openstack.org/349032 | 19:03 |
| *** elopez_ has quit IRC | 19:03 | |
| *** thorst_ has joined #openstack-ansible | 19:03 | |
| *** Guest84124 has quit IRC | 19:04 | |
| stevelle | anyone have any clues what might be causing this with gnocchi only? I'm not seeing issues with other openstack clients http://paste.openstack.org/show/F5VQMWLxXOjMXSWEvALX/ | 19:05 |
| *** elo has quit IRC | 19:06 | |
| *** thorst has quit IRC | 19:06 | |
| *** elo has joined #openstack-ansible | 19:07 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_cinder: Fix bug in RPC config that broke Rabbit SSL support https://review.openstack.org/349034 | 19:08 |
| *** hybridpollo has joined #openstack-ansible | 19:09 | |
| mhayden | stevelle: is there a variable you can set on whether to validate the cert or not? | 19:12 |
| stevelle | mhayden: I can use --insecure on the client but that seems like a poor life choice | 19:12 |
| stevelle | it works if I do however | 19:12 |
| *** cbits has quit IRC | 19:14 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_heat: Fix bug in RPC config that broke Rabbit SSL support https://review.openstack.org/349040 | 19:15 |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_aodh: Support Vagrant testing with Xenial https://review.openstack.org/349041 | 19:17 |
| *** cloudtrainme has quit IRC | 19:17 | |
| openstackgerrit | Travis Truman (automagically) proposed openstack/openstack-ansible-os_aodh: Fix bug in RPC config that broke Rabbit SSL support https://review.openstack.org/349043 | 19:20 |
| *** TxGirlGeek has joined #openstack-ansible | 19:22 | |
| automagically | Later all, enjoy your weekend | 19:23 |
| *** rstarmer has quit IRC | 19:33 | |
| *** sdake has joined #openstack-ansible | 19:35 | |
| *** TxGirlGeek has quit IRC | 19:40 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:40 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_aodh: Test aodh role using the aodh tempest plugin https://review.openstack.org/348048 | 19:41 |
| *** tlbr has joined #openstack-ansible | 19:42 | |
| *** tlbr is now known as Guest79100 | 19:42 | |
| *** grumpycatt has quit IRC | 19:44 | |
| *** grumpycatt has joined #openstack-ansible | 19:44 | |
| mhayden | automagically: enjoy! | 19:48 |
| *** TxGirlGeek has quit IRC | 19:48 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:48 | |
| *** Guest79100 has quit IRC | 19:51 | |
| *** rstarmer has joined #openstack-ansible | 19:51 | |
| *** TxGirlGeek has quit IRC | 19:51 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:52 | |
| *** TxGirlGeek has quit IRC | 19:54 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:54 | |
| *** grumpycatt has quit IRC | 19:55 | |
| *** grumpycatt has joined #openstack-ansible | 19:56 | |
| *** grumpycatt has quit IRC | 19:57 | |
| *** grumpycatt has joined #openstack-ansible | 19:57 | |
| *** grumpycatt has quit IRC | 19:57 | |
| *** TxGirlGeek has quit IRC | 19:57 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:58 | |
| *** asettle has joined #openstack-ansible | 19:58 | |
| *** rahuls has joined #openstack-ansible | 19:59 | |
| *** sdake has quit IRC | 19:59 | |
| *** grumpycatt has joined #openstack-ansible | 20:00 | |
| *** TxGirlGeek has quit IRC | 20:00 | |
| *** TxGirlGeek has joined #openstack-ansible | 20:01 | |
| *** asettle has quit IRC | 20:03 | |
| *** TxGirlGeek has quit IRC | 20:06 | |
| *** rstarmer has quit IRC | 20:06 | |
| *** TxGirlGeek has joined #openstack-ansible | 20:06 | |
| *** adrian_otto has joined #openstack-ansible | 20:11 | |
| *** TxGirlGeek has quit IRC | 20:11 | |
| *** jthorne has joined #openstack-ansible | 20:13 | |
| *** rstarmer has joined #openstack-ansible | 20:15 | |
| *** rstarmer has quit IRC | 20:20 | |
| *** jthorne has quit IRC | 20:28 | |
| *** TxGirlGeek has joined #openstack-ansible | 20:29 | |
| *** klamath has quit IRC | 20:32 | |
| *** sguduru_ has quit IRC | 20:40 | |
| *** GMAzrael has quit IRC | 20:45 | |
| *** messy has quit IRC | 20:46 | |
| *** retreved has quit IRC | 20:53 | |
| *** rstarmer has joined #openstack-ansible | 20:54 | |
| *** TxGirlGeek has quit IRC | 20:58 | |
| *** TxGirlGeek has joined #openstack-ansible | 20:58 | |
| *** TxGirlGeek has quit IRC | 21:04 | |
| *** TxGirlGeek has joined #openstack-ansible | 21:04 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible: [TESTING] Run using Ansible "v2.1.1.0" https://review.openstack.org/346839 | 21:09 |
| *** jamesdenton has quit IRC | 21:13 | |
| *** KLevenstein has quit IRC | 21:14 | |
| *** elo has quit IRC | 21:14 | |
| *** Rodrigo_BR has quit IRC | 21:19 | |
| *** elo has joined #openstack-ansible | 21:22 | |
| *** TxGirlGeek has quit IRC | 21:22 | |
| *** TxGirlGeek has joined #openstack-ansible | 21:23 | |
| *** elo has quit IRC | 21:27 | |
| *** Guest97891 has joined #openstack-ansible | 21:32 | |
| *** johnmilton has quit IRC | 21:35 | |
| *** Guest97891 has quit IRC | 21:39 | |
| *** TxGirlGeek has quit IRC | 21:45 | |
| *** tlbr_ has joined #openstack-ansible | 21:46 | |
| *** tlbr_ has quit IRC | 21:47 | |
| *** johnmilton has joined #openstack-ansible | 21:48 | |
| *** weshay is now known as weshay_afk | 21:51 | |
| *** timl0101 has quit IRC | 21:52 | |
| *** elo has joined #openstack-ansible | 21:56 | |
| *** TxGirlGeek has joined #openstack-ansible | 21:59 | |
| *** tlbr_ has joined #openstack-ansible | 22:00 | |
| *** TxGirlGeek has quit IRC | 22:01 | |
| *** zhangjn has quit IRC | 22:02 | |
| *** zhangjn has joined #openstack-ansible | 22:04 | |
| *** tlbr_ has quit IRC | 22:04 | |
| *** Guest13477 has joined #openstack-ansible | 22:14 | |
| *** daneyon has quit IRC | 22:15 | |
| *** Guest13477 has quit IRC | 22:19 | |
| *** tlbr_ has joined #openstack-ansible | 22:24 | |
| *** tlbr_ has quit IRC | 22:29 | |
| *** jthorne has joined #openstack-ansible | 22:29 | |
| *** psilvad has quit IRC | 22:32 | |
| *** psilvad has joined #openstack-ansible | 22:32 | |
| *** jthorne has quit IRC | 22:33 | |
| *** tlbr_ has joined #openstack-ansible | 22:35 | |
| *** galstrom is now known as galstrom_zzz | 22:37 | |
| *** tlbr_ has quit IRC | 22:40 | |
| *** rstarmer has quit IRC | 22:42 | |
| *** rstarmer has joined #openstack-ansible | 22:56 | |
| *** Guest59655 has joined #openstack-ansible | 22:56 | |
| *** Guest59655 has quit IRC | 23:02 | |
| *** thorst_ has quit IRC | 23:04 | |
| *** thorst has joined #openstack-ansible | 23:05 | |
| *** rstarmer has quit IRC | 23:08 | |
| *** jthorne has joined #openstack-ansible | 23:10 | |
| *** thorst has quit IRC | 23:13 | |
| *** jthorne has quit IRC | 23:14 | |
| *** jthorne has joined #openstack-ansible | 23:14 | |
| *** tlbr_ has joined #openstack-ansible | 23:22 | |
| *** tlbr_ has quit IRC | 23:27 | |
| *** rstarmer has joined #openstack-ansible | 23:28 | |
| *** kulinacs has quit IRC | 23:30 | |
| *** tlbr has joined #openstack-ansible | 23:34 | |
| *** tlbr is now known as Guest16207 | 23:34 | |
| *** Guest16207 has quit IRC | 23:38 | |
| *** rstarmer has quit IRC | 23:44 | |
| *** rstarmer has joined #openstack-ansible | 23:45 | |
| *** rstarmer has quit IRC | 23:51 | |
| *** tlbr_ has joined #openstack-ansible | 23:58 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_cloudkitty: Implement Multi-OS support in os_cloudkitty https://review.openstack.org/346952 | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!