| *** gvincent has joined #ara | 07:03 | |
| *** sshnaidm|off is now known as sshnaidm|rover | 07:07 | |
| *** themroc has joined #ara | 08:30 | |
| *** gvincent has quit IRC | 12:40 | |
| *** themurph has joined #ara | 13:05 | |
| *** tbielawa has joined #ara | 13:07 | |
| *** tbielawa is now known as tbielawa|rdu | 13:07 | |
| *** gvincent has joined #ara | 14:04 | |
| *** themurph has quit IRC | 14:05 | |
| *** themurph has joined #ara | 14:39 | |
| *** tbielawa|rdu is now known as tbielawa | 14:53 | |
| *** tbielawa is now known as tbielawa|rdu | 14:53 | |
| *** spiette has quit IRC | 14:54 | |
| *** spiette has joined #ara | 15:06 | |
| *** cloudnull has quit IRC | 15:22 | |
| dmsimard | apollo13: around to chat about the offline client patch ? | 15:44 |
|---|---|---|
| apollo13 | dmsimard: sure, at least for 10 minutes, maybe a little bit more, then I'll head home | 15:48 |
| apollo13 | so shoot :) | 15:48 |
| dmsimard | so, first of all | 15:48 |
| dmsimard | I really like the approach | 15:48 |
| dmsimard | of re-using the http client | 15:48 |
| dmsimard | and it respects the requirement that we should not require users to stand up an api server for things to work | 15:49 |
| apollo13 | I fear there is a but coming :) | 15:49 |
| dmsimard | it's not a but, it's a question | 15:49 |
| dmsimard | how would this work if a user has concurrent ansible-playbook commands running ?% | 15:50 |
| apollo13 | each of them would spawn their own server | 15:50 |
| dmsimard | on the same host/port ? | 15:50 |
| apollo13 | same host yes, but I bind to port zero | 15:51 |
| apollo13 | which means the kernel assigns a free port above 32k | 15:51 |
| apollo13 | (or whatever the dynamic range is on that system, do not know the sysctl out of my head for that) | 15:51 |
| dmsimard | oh, interesting | 15:52 |
| apollo13 | cat /proc/sys/net/ipv4/ip_local_port_range | 15:53 |
| apollo13 | if you exhaust that you are out of luck (it shows min - max) | 15:53 |
| apollo13 | that also includes ports that are needed to create a tcp connection (ie src port) | 15:53 |
| dmsimard | ok, that's really good | 15:54 |
| apollo13 | gotta run home, will read later | 15:56 |
| dmsimard | that's all I had for now | 15:56 |
| dmsimard | I'll try it out and review the code | 15:56 |
| apollo13 | dmsimard: to try the integration tests on ara-server you need to set allowed_hosts in server/configs/integration.cfg to localhost, it is no longer testserver | 15:59 |
| *** tbielawa|rdu is now known as tbielawa|lunch | 16:28 | |
| dmsimard | apollo13: works for me at a first glance -- how did you send up setting allowed_hosts ? Setting "testserver,localhost" doesn't work (it's probably not expecting it to be a list) and setting the same thing as ALLOWED_HOSTS env var didn't seem to work either.. I ended up setting just "localhost" for now but it seems like we might want more than one host in there are some point so we probably need to figure that | 16:36 |
| dmsimard | out | 16:36 |
| apollo13 | oh yes I also just did set it to localhost; didn't check the other variant | 16:37 |
| apollo13 | have to see why that fails | 16:37 |
| apollo13 | gimme a sec | 16:38 |
| apollo13 | (sorry for not testing that, but multiple allowed hosts is rather an edge case for most use cases) | 16:38 |
| apollo13 | dmsimard: does ALLOWED_HOSTS = env.list("ALLOWED_HOSTS", default=[]) in the settings.py fix it? | 16:39 |
| apollo13 | instead of what master currently has | 16:39 |
| apollo13 | let me read the docs though | 16:40 |
| dmsimard | dunno, let me see | 16:43 |
| apollo13 | na, it does not; I fucked up somewhere along the way | 16:43 |
| apollo13 | I am currently as far as ALLOWED_HOSTS = env.list("ALLOWED_HOSTS", default=[]) but it double encodes somewhere | 16:46 |
| *** openstackgerrit has quit IRC | 16:48 | |
| *** themurph has quit IRC | 16:52 | |
| apollo13 | dmsimard: ah lol, it does not simply pass an ini file but uses config obj :/ | 16:52 |
| *** themroc has quit IRC | 16:52 | |
| dmsimard | apollo13: the offline client stuff just works, I'm not finding any issues with it | 16:53 |
| *** sshnaidm|rover is now known as sshnaidm|afk | 16:53 | |
| dmsimard | Code looks good to go as far as I can tell, are you good to merge it? I guess we need a patch to at least change testserver for localhost | 16:55 |
| apollo13 | dmsimard: yes, fwiw your env variable would have been ARA_ALLOWED_HOSTS | 16:56 |
| apollo13 | all vars are prefixed with ARA | 16:56 |
| apollo13 | like in the ini file where you have [ara] as section | 16:56 |
| dmsimard | apollo13: oh | 16:56 |
| dmsimard | so what threw me off was the django error message | 16:57 |
| dmsimard | that mentions ALLOWED_HOSTS | 16:57 |
| dmsimard | (not ARA_ALLOWED_HOSTS) | 16:57 |
| apollo13 | yeah, because for Django it is ALLOWED_HOSTS, the thing is I do not want to set stuff like DEBUG in the env | 16:57 |
| apollo13 | namespacing seems nicer | 16:57 |
| dmsimard | makes sense to namespace ARA things | 16:58 |
| dmsimard | django things, maybe not | 16:58 |
| dmsimard | like SECRET_KEY and ALLOWED_HOSTS | 16:59 |
| apollo13 | I fear endless confusin | 16:59 |
| apollo13 | how would you know what ends up in [ara] in the ini file and what outside it | 16:59 |
| apollo13 | or maybe even add [django]? :D | 16:59 |
| dmsimard | hmmm | 16:59 |
| dmsimard | so those settings are in the server directory now | 17:00 |
| dmsimard | despite that, would it mean that API settings would end up being there too ? | 17:01 |
| apollo13 | what do you mean? | 17:01 |
| dmsimard | there's ara/server/settings.py | 17:02 |
| dmsimard | but there's no ara/api/settings.py | 17:02 |
| dmsimard | is that ok ? | 17:02 |
| apollo13 | yes because ara/server is the django project which has settings for all apps | 17:02 |
| apollo13 | ara/api is just an INSTALLED_APP like django.contrib.admin | 17:02 |
| dmsimard | ok, makes sense | 17:03 |
| *** openstackgerrit has joined #ara | 17:08 | |
| openstackgerrit | Merged openstack/ara-clients master: Switched the offline client to use an actual threaded http server. https://review.openstack.org/615439 | 17:08 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Fixes parsing of lists in config files. https://review.openstack.org/615609 | 17:12 |
| apollo13 | ^ this should allow you to use lists in config files; I'll see that I can upstream that patch | 17:12 |
| apollo13 | for now it would do | 17:12 |
| apollo13 | ups, need to run black and isort | 17:12 |
| apollo13 | I have to disable recreation of the linter env in tox.ini | 17:13 |
| apollo13 | btw the tox -e linters seems to be broken; it says linters: commands succeeded even though isort and black failed | 17:14 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Fixes parsing of lists in config files. https://review.openstack.org/615609 | 17:14 |
| dmsimard | apollo13: ok I'll check it out, thanks | 17:16 |
| apollo13 | lets see https://github.com/willkg/everett/pull/71 | 17:27 |
| apollo13 | time for sport | 17:27 |
| *** tbielawa|lunch has quit IRC | 18:01 | |
| *** themurph has joined #ara | 18:05 | |
| *** themurph has quit IRC | 19:19 | |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Set exit code to 1 if at least one linter fails. https://review.openstack.org/615634 | 19:34 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Fixes parsing of lists in config files. https://review.openstack.org/615609 | 19:45 |
| apollo13 | dmsimard: lol, I fail at typing | 19:53 |
| dmsimard | ¯\_(ツ)_/¯ | 19:53 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Set exit code to 1 if at least one linter fails. https://review.openstack.org/615634 | 19:54 |
| apollo13 | that's it for me today codewise though | 19:54 |
| apollo13 | btw how do you feel about the IBM acquisition? | 19:54 |
| dmsimard | ¯\_(ツ)_/¯ | 19:55 |
| dmsimard | haha | 19:55 |
| dmsimard | I was in PTO last week so missed out on a lot and haven't fully caught up yet | 19:55 |
| dmsimard | my "public" statement about the acquisition https://twitter.com/dmsimard/status/1057267958253268993 | 19:56 |
| apollo13 | that's a fair statement, at least assuming that IBM doesn't force new policies down wrt open source at red hat. though that would probably be the end of red hat anyways | 19:57 |
| dmsimard | yeah, I mean, $34b is a lot of money | 19:58 |
| dmsimard | open source software is public and licensed so they're not really buying any of that | 19:59 |
| apollo13 | true, but take ansible as an example; would tower still be opensource if red hat were to acquire it now… that said, one can just hope for the best :) | 20:00 |
| dmsimard | I've definitely seen a lot of concern on social media about not just ansible but fedora, centos, ceph, etc. | 20:01 |
| apollo13 | guess we will see what the future brings | 20:02 |
| dmsimard | hopefully it doesn't turn out like oracle and sun :( | 20:02 |
| apollo13 | ha yeah | 20:02 |
| *** themurph has joined #ara | 20:19 | |
| dmsimard | apollo13: the linters job failed but it's because it's running on xenial, not bionic | 20:24 |
| dmsimard | I'll fix it | 20:24 |
| *** themurph has quit IRC | 22:03 | |
| *** Xaroth has quit IRC | 22:19 | |
| *** apollo13 has quit IRC | 22:21 | |
| *** apollo13 has joined #ara | 22:23 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!