| *** zbr has quit IRC | 00:09 | |
| *** cweeks has quit IRC | 08:08 | |
| *** zbr has joined #ara | 08:31 | |
| *** zbr has quit IRC | 08:35 | |
| *** zbr has joined #ara | 08:37 | |
| *** gvincent has joined #ara | 09:05 | |
| *** sshnaidm|off is now known as sshnaidm|rover | 09:30 | |
| *** irclogbot_3 has quit IRC | 09:39 | |
| *** irclogbot_2 has joined #ara | 09:39 | |
| *** openstackgerrit has quit IRC | 12:41 | |
| *** sshnaidm|rover is now known as sshnaidm|ruck | 12:55 | |
| *** dbpiv has joined #ara | 13:12 | |
| *** dbpiv has quit IRC | 14:50 | |
| *** openstackgerrit has joined #ara | 14:55 | |
| openstackgerrit | David Moreau Simard proposed recordsansible/ara master: WIP: Reverse default sort order for playbooks, plays, tasks and results https://review.opendev.org/693855 | 14:55 |
|---|---|---|
| *** dbpiv has joined #ara | 14:58 | |
| *** etienne has joined #ara | 18:03 | |
| *** cweeks has joined #ara | 19:24 | |
| *** sshnaidm|ruck is now known as sshnaidm|afk | 19:42 | |
| *** crag has joined #ara | 19:59 | |
| crag | is generating static HTML from ara not a thing anymore, or am i just missing/misinterpreting docs? | 19:59 |
| dmsimard | crag: it was re-introduced in 1.2.0 which was released last week | 20:00 |
| dmsimard | mentioned in the blog post for the release: https://ara.recordsansible.org/blog/2019/11/06/announcing-the-release-of-ara-records-ansible-1.2/ | 20:01 |
| crag | dmsimard: aha nice. so for some clarification - to utilize this, do I actually need an instance of ara running, or can i make my records of runs locally and store these static results? | 20:03 |
| dmsimard | crag: you don't need to run a persistent instance to generate the static files | 20:23 |
| crag | okay, so for instance - my targets will be RHEL 7 + 8 systems, and i should be able to pip ara there, generate the static HTML and offload it | 20:24 |
| dmsimard | crag: ara needs to be installed once only where ansible is running from -- no need to install it everywhere | 20:25 |
| crag | well, my tests are likely going to be running on each test system individually, not from a remote host. | 20:35 |
| crag | as the versions of RHEL and the versions of ansible are the variables and key to the tests., | 20:35 |
| crag | but i get what you are saying | 20:35 |
| dmsimard | crag: if you run ansible against localhost on every node, you can generate the html report interface for each node and send them somewhere to host them -- this is a common use case for CI (i.e, one report per job) | 22:35 |
| dmsimard | the new thing in 1.x is that it's easier to send data to a remote server so it's aggregated in one place | 22:36 |
| dmsimard | so, for example, your RHEL nodes could be configured to use the HTTP API client and a specific persistent server | 22:38 |
| dmsimard | in practice, I got results from openstack-ansible with just this: https://review.opendev.org/#/c/682962/3/scripts/openstack-ansible.sh | 22:38 |
| dmsimard | there is a performance impact when sending data remotely because of the latency to the API server but depending on the use case, it can be worth it | 22:39 |
| dmsimard | the interesting part about having data in a live API server is that you can then query it for data | 22:40 |
| dmsimard | if you want to keep data isolated per-node or per-job, another approach would be the distributed sqlite backend: https://ara.readthedocs.io/en/latest/distributed-sqlite-backend.html | 22:41 |
| dmsimard | the idea is that instead of generating the html files (and uploading them, and storing them), you just host the sqlite databases | 22:41 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!