| gvincent | apollo13, for 605421 does drf-extensions exist as rpm? IIRC this is why I didn't use it. I think there is no drf-extensions rpm | 04:04 |
|---|---|---|
| *** gvincent has quit IRC | 05:46 | |
| dmsimard | apollo13: I'll review your patches this morning, meanwhile, a bit of docs I came up with: http://logs.openstack.org/29/605529/2/check/build-openstack-sphinx-docs/94a1479/html/arch.html#recording-data-from-ansible | 11:49 |
| dmsimard | gvincent isn't here right now but don't worry about rpm packages, I'm a packager in Fedora and CentOS :p | 11:50 |
| *** pvc has joined #ara | 13:10 | |
| pvc | hi | 13:10 |
| pvc | is there a way that i can use a single server ara than can put all my 5 ansible nodes if ever? | 13:11 |
| pvc | or ara is just a stand alone server? | 13:11 |
| *** tbielawa has joined #ara | 13:11 | |
| pvc | is there a way that i can use a single server ara than can put all my 5 ansible nodes if ever? | 13:12 |
| pvc | or ara is just a stand alone server? | 13:12 |
| pvc | anyone? | 13:17 |
| dmsimard | pvc: hey, sorry I was grabbing breakfast :p | 13:23 |
| dmsimard | pvc: how are you using ansible right now ? are you using ansible on 5 machines ? like, each of these 5 machines run ansible ? | 13:24 |
| pvc | yes thats my scenario im running ansible on 5 machines | 13:25 |
| pvc | is it possible that i can use one ARA web server that contains all of the host that have my ansible | 13:25 |
| pvc | then i can select which host i can view? | 13:25 |
| pvc | or ARA is just a stand alone? | 13:25 |
| dmsimard | the default database backend is sqlite | 13:26 |
| dmsimard | but you can use mysql/postgresql to aggregate data from different locations | 13:26 |
| pvc | so if i want another node to have in my ara, i will register it | 13:26 |
| pvc | then i can view it on ara dashboard | 13:26 |
| dmsimard | well it's not really how it works | 13:26 |
| pvc | i see | 13:26 |
| dmsimard | ara provides an ansible callback plugin which records things into a database | 13:26 |
| dmsimard | ara needs to be installed wherever ansible is running | 13:27 |
| pvc | so one ara one server? | 13:27 |
| dmsimard | and then the callback needs to be configured to send data to a single place (mysql) | 13:27 |
| dmsimard | you can host the web interface wherever you want, so long as it has access to the same database | 13:27 |
| pvc | i see. i want to integrate ara in my company so all their ansible playbook will be saved | 13:27 |
| dmsimard | great idea :) | 13:27 |
| pvc | one ara web server for all ansible playbooks | 13:28 |
| pvc | i mean all ansible nodes | 13:28 |
| dmsimard | I'm in meeting most of the day today but here's a few links that might help | 13:28 |
| dmsimard | https://ara.readthedocs.io/en/stable/faq.html#can-ansible-with-ara-run-on-a-different-server-than-the-web-application | 13:28 |
| pvc | thank you so much dmsimard | 13:28 |
| pvc | ara so great btw :) | 13:28 |
| dmsimard | https://ara.readthedocs.io/en/stable/configuration.html and more importantly https://ara.readthedocs.io/en/stable/configuration.html#ara-database | 13:28 |
| dmsimard | happy to hear about it, thanks :D | 13:28 |
| pvc | thank you so much | 13:29 |
| *** themurph has joined #ara | 13:55 | |
| apollo13 | dmsimard: ok, gotta head home now | 14:15 |
| apollo13 | either way, I tend to isolate my python stuff from the system as much as possible | 14:16 |
| apollo13 | it's just not possible to install two apps which require different django versions etc… | 14:16 |
| apollo13 | hence venvs and/or tools like pex and fpm | 14:16 |
| dmsimard | apollo13: oh, sure, I just meant that as a way to address gvincent's concern | 14:19 |
| *** evrardjp has quit IRC | 14:45 | |
| *** evrardjp has joined #ara | 14:47 | |
| *** evrardjp has quit IRC | 14:58 | |
| *** themurph has quit IRC | 15:09 | |
| *** themurph has joined #ara | 15:15 | |
| *** spiette_ is now known as spiette | 15:21 | |
| apollo13 | dmsimard: that said, rpm allows for rootless installation if they are reloctable right? | 15:24 |
| apollo13 | something I might look into one day :) | 15:24 |
| dmsimard | apollo13: rootless ? | 15:24 |
| apollo13 | well you can install it into your home directory and don't need sudo rights for installing iirc | 15:24 |
| apollo13 | ah maybe not without root if it still should end up in the rpmdb | 15:27 |
| dmsimard | apollo13: oh, well -- regardless of the use case, ara (0.x) is already packaged in fedora | 15:35 |
| dmsimard | I looked at drf quicky and it doesn't seem like a big deal | 15:35 |
| dmsimard | I mean, drf-extensions | 15:35 |
| apollo13 | oh yeah, it is just a standard python thingy | 15:36 |
| apollo13 | whatever macros rpm has should suffice | 15:36 |
| dmsimard | we do a lot of python packaging in openstack :P | 15:36 |
| apollo13 | hehe | 15:37 |
| dmsimard | apollo13: left a few comments in https://review.openstack.org/#/c/605421/ | 16:13 |
| apollo13 | dmsimard: is that a run of ara-server or ara-client? http://logs.openstack.org/03/605503/1/check/ara-server-ansible-integration/0a85a04/ara-report/result/0e7334be-0f66-46bf-83a3-2d6c6fed794a/ | 16:27 |
| apollo13 | mhm, probably server, let me run that locally | 16:28 |
| dmsimard | apollo13: ansible-integration installs ara-server, ara-client and ara-plugins and then it runs an ansible playbook command with the callback enabled | 16:28 |
| apollo13 | Out[2]: <module 'ara' from '/home/florian/sources/ara-server/.tox/ansible-integration/lib/python3.6/site-packages/ara/__init__.py'> | 16:34 |
| apollo13 | aha | 16:34 |
| apollo13 | something installed an __init__.py which breaks namespacing | 16:34 |
| apollo13 | dmsimard: oh, the issue is that ansible-integration seems to manage to also install ara 0.x | 16:36 |
| apollo13 | ansible-integration installed: alembic==1.0.0,ansible==2.6.4,ara==0.16.1 <-- there | 16:36 |
| apollo13 | this breaks the namespacing | 16:36 |
| apollo13 | because ara 0.16.1 has an __init__.py in ara/ and therefor "destroys" the package I think | 16:36 |
| apollo13 | have to figure out what excatly requires ara currently | 16:37 |
| apollo13 | as soon as I remove that __init__.py it works as expected: https://dpaste.de/xNjM | 16:38 |
| apollo13 | so namespacing in theory does work, but ara 0.16.1 shouldn't be installed | 16:38 |
| apollo13 | why is it installed? | 16:39 |
| dmsimard | oh, because they're a namespaced package they install ara ? | 16:39 |
| apollo13 | no | 16:39 |
| apollo13 | should not be the case | 16:39 |
| dmsimard | if I had to take a guess, it's probably coming from https://github.com/openstack/ara-server/blob/master/setup.cfg#L28-L33 | 16:39 |
| apollo13 | nope | 16:39 |
| apollo13 | wrong guess | 16:40 |
| apollo13 | Collecting ara>=clients (from ara-plugins) | 16:40 |
| apollo13 | now wtf is pip smoking there | 16:41 |
| apollo13 | I want it too fwiw | 16:41 |
| dmsimard | hang on | 16:43 |
| dmsimard | well, ara-clients also has the same setup.cfg tidbits fwiw | 16:43 |
| apollo13 | yes, it is not about that, it is about requirements | 16:43 |
| apollo13 | does pbr use requirements.txt as setup requirements? | 16:43 |
| dmsimard | it does | 16:44 |
| apollo13 | I wonder if it missinterprets this from requirements.txt: git+https://git.openstack.org/openstack/ara-clients@master#egg=ara-clients | 16:44 |
| dmsimard | let me check something | 16:45 |
| apollo13 | jupp that's it, just no idea how I can tell it to behave | 16:46 |
| apollo13 | jupp if I remove that line from requirements.txt pip does not install ara | 16:51 |
| apollo13 | otherwise it installs ara>=clients | 16:51 |
| dmsimard | apollo13: oh, just found the issue for black | 16:54 |
| dmsimard | it requires py36+ and the node it ran on was a ubuntu 16.04 which ships 3.5 | 16:54 |
| apollo13 | ah | 16:54 |
| dmsimard | let me add a comment on the patch | 16:54 |
| dmsimard | there, see my last comment on https://review.openstack.org/#/c/605439/ | 16:57 |
| apollo13 | aha pbr parses the url wrongly | 16:59 |
| dmsimard | That's odd | 17:00 |
| apollo13 | https://github.com/openstack-dev/pbr/blob/master/pbr/packaging.py#L135-L165 | 17:01 |
| apollo13 | na, regex | 17:01 |
| apollo13 | now you got another problem | 17:01 |
| apollo13 | also pip install -r properly installs ara and not ara>=clients | 17:01 |
| dmsimard | apollo13: were you trying to fix something in specific with the namespace stuff ? or was it part of the cleanup | 17:02 |
| apollo13 | dmsimard: cleanup | 17:02 |
| apollo13 | but as you can see it already revealed a bug in pbr :D | 17:02 |
| apollo13 | https://github.com/openstack-dev/pbr/blob/master/pbr/packaging.py#L104-L110 | 17:04 |
| apollo13 | there we go | 17:04 |
| apollo13 | this is the bug | 17:04 |
| apollo13 | the egg fragment is #egg=ara-clients, but clients is not a version | 17:04 |
| dmsimard | apollo13: oh, it's the dash in ara-clients that's screwing things up | 17:06 |
| dmsimard | ... nice | 17:06 |
| apollo13 | it is a valid package name though, so imo pbr is screwing up | 17:06 |
| dmsimard | yeah, let me get a pbr expert :D | 17:07 |
| apollo13 | much appreciated! | 17:07 |
| apollo13 | arg | 17:08 |
| apollo13 | The "project name" component of the url suffix "egg=<project name>-<version>" is used by pip in its dependency logic to identify the project prior to pip downloading and analyzing the metadata. The optional "version" component of the egg name is not functionally important. It merely provides a human-readable clue as to what version is in use. | 17:08 |
| apollo13 | in https://pip.pypa.io/en/stable/reference/pip_install/#vcs-support | 17:08 |
| apollo13 | so pbr is not really doing that much wrong | 17:08 |
| apollo13 | maybe we should normalize the package name to ara_client (for the egg link) let me try that | 17:08 |
| dmsimard | apollo13: yeah the underscore probably wouldn't trip the regex issue | 17:09 |
| apollo13 | and package normalization seems to replace - with _ in pip logic, let me find the code | 17:10 |
| apollo13 | that said I think we can safely replace ara-client with ara_clients in the egg name | 17:10 |
| apollo13 | https://github.com/pypa/pip/blob/9.0.1/pip/_vendor/packaging/utils.py | 17:13 |
| apollo13 | jupp, pip replaces -_. with - internally anyways; let me fix that in all three repos | 17:13 |
| dmsimard | nice find | 17:13 |
| *** openstackgerrit has joined #ara | 17:15 | |
| openstackgerrit | Florian Apolloner proposed openstack/ara-plugins master: Fixed namespacing to use python3 style namespace packages. https://review.openstack.org/605817 | 17:15 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-clients master: Fixed namespacing to use python3 style namespace packages. https://review.openstack.org/605819 | 17:16 |
| apollo13 | you probably have to trust me and just merged those two. then we can retest the integration on ansible server | 17:17 |
| apollo13 | ara-server sorry | 17:17 |
| apollo13 | ups, ara-plugins is wrong | 17:18 |
| apollo13 | I made it editable | 17:18 |
| apollo13 | shouldn't be | 17:18 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-plugins master: Fixed namespacing to use python3 style namespace packages. https://review.openstack.org/605817 | 17:18 |
| apollo13 | gotta leave for now, will look at ara-server tomorrow and address your comments at work. would be great if you could merge ara-plugins/clients in the meantime | 17:19 |
| *** tbielawa is now known as tbielawa|lunch | 17:41 | |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Added CI for black/isort. https://review.openstack.org/605439 | 17:45 |
| *** sshnaidm is now known as sshnaidm|off | 18:03 | |
| *** tbielawa|lunch is now known as tbielawa | 18:52 | |
| openstackgerrit | Merged openstack/ara-server master: Update and improve README to be more accurate https://review.openstack.org/605528 | 19:10 |
| openstackgerrit | Merged openstack/ara-server master: Bootstrap documentation and publish it to readthedocs https://review.openstack.org/605529 | 19:12 |
| *** themurph has quit IRC | 21:16 | |
| *** tbielawa is now known as tbielawa|g0n3 | 21:25 | |
| *** tbielawa|g0n3 has quit IRC | 21:26 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!