| *** apetrich has quit IRC | 02:10 | |
| *** eyalb has joined #openstack-mistral | 05:44 | |
| *** eyalb has quit IRC | 05:47 | |
| *** eyalb has joined #openstack-mistral | 05:49 | |
| *** pgaxatte has joined #openstack-mistral | 06:30 | |
| *** rakhmerov has joined #openstack-mistral | 07:35 | |
| rakhmerov | hi, please review https://review.opendev.org/#/c/685899/ | 07:56 |
|---|---|---|
| *** shardy has joined #openstack-mistral | 08:14 | |
| *** shardy has quit IRC | 08:15 | |
| d0ugal | rakhmerov: LGTM! | 08:15 |
| rakhmerov | thanks! | 08:16 |
| d0ugal | rakhmerov: Any idea what would cause this error? https://zuul.opendev.org/t/openstack/build/6022fb0642614d6a8d2fe586e922e98f/log/logs/undercloud/var/log/containers/mistral/engine.log.txt.gz#95 | 08:16 |
| rakhmerov | looking.. | 08:16 |
| d0ugal | I guess the api can't reach the engine? | 08:16 |
| d0ugal | or the engine can't reach the executor maybe :) | 08:16 |
| rakhmerov | hah, interesting | 08:18 |
| rakhmerov | yeah.. | 08:18 |
| *** shardy has joined #openstack-mistral | 08:19 | |
| d0ugal | ohhh | 08:19 |
| d0ugal | I think I know what it is | 08:19 |
| rakhmerov | well, it's really weird because it fails on return a.is_sync() | 08:19 |
| rakhmerov | seems like | 08:19 |
| d0ugal | I just looked to see what the patch changes | 08:19 |
| d0ugal | https://review.opendev.org/#/c/686181/1/tripleo_common/image/image_uploader.py | 08:19 |
| *** shardy has quit IRC | 08:19 | |
| d0ugal | It adds multiprocessing. iirc multiprocessing doesn't work well with eventlet | 08:19 |
| rakhmerov | ooh the class is not available | 08:19 |
| rakhmerov | yeah, I remember having some issues with that too | 08:20 |
| rakhmerov | I guess you're right | 08:20 |
| rakhmerov | d0ugal: but I would still add logging into mistral/engine/actions.py (line 335) to see what exactly failed | 08:21 |
| d0ugal | rakhmerov: Good idea. I'll do that | 08:22 |
| rakhmerov | it may also be "prepared_input_dict = self._prepare_input(input_dict)" or a = a_m.get_action_class(self.action_def.name)(**prepared_input_dict) | 08:22 |
| rakhmerov | if it's the second thing then the action class simply may not be on the pythonpath | 08:22 |
| rakhmerov | if it's the first thing then I don't know :) | 08:23 |
| rakhmerov | most likely it's this: a_m.get_action_class(self.action_def.name) | 08:23 |
| d0ugal | some of the multiprocessing code would run when the module is imported - so that would make sense why it can't load the class | 08:23 |
| rakhmerov | yeah | 08:23 |
| d0ugal | The action has previously worked, so I think the paths should be good | 08:23 |
| d0ugal | Thanks. I'll add some logging and see what else we can learn. | 08:24 |
| rakhmerov | ok | 08:25 |
| rakhmerov | let me know when you dig something new | 08:25 |
| rakhmerov | I'm also curious what it is | 08:25 |
| *** eyalb1 has joined #openstack-mistral | 08:26 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral stable/train: Fix "root_execution" lazy loading issue and refactor execution.py https://review.opendev.org/686349 | 08:26 |
| rakhmerov | d0ugal: btw, do you know if everything's set up to run TripleO CI jobs against stable/train? | 08:27 |
| rakhmerov | we have some patches to stable/train and they fail on TripleO | 08:27 |
| rakhmerov | as if not everything was just ready | 08:27 |
| d0ugal | I guess we just have our normal packaging etc. issues around release time | 08:28 |
| d0ugal | but I don't know. I can ask around | 08:28 |
| rakhmerov | ok | 08:28 |
| rakhmerov | d0ugal: for example https://e0f905401f3bac5f7468-fb5f8b6fa903b511ee0656ffba88adbe.ssl.cf2.rackcdn.com/685477/1/check/tripleo-ci-centos-7-containers-multinode/9f3353e/job-output.txt | 08:29 |
| rakhmerov | (I wonder why it's not possible to point to a line anymore) | 08:29 |
| rakhmerov | but it says "mit_releases_file.py: error: argument --stable-release: invalid choice: 'train' (choose from 'newton', 'ocata', 'pike', 'queens', 'rocky', 'stein', 'master')" | 08:29 |
| *** eyalb has quit IRC | 08:29 | |
| rakhmerov | so yes, please ask | 08:31 |
| openstackgerrit | Dougal Matthews proposed openstack/mistral master: Log the original exception in is_sync https://review.opendev.org/686353 | 08:31 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral stable/train: Add the explicit dependency from kombu https://review.opendev.org/686354 | 08:33 |
| openstackgerrit | ali proposed openstack/mistral master: Task publish is ignored silently:- https://review.opendev.org/684711 | 09:07 |
| *** ricolin has quit IRC | 09:15 | |
| openstackgerrit | ali proposed openstack/mistral master: It's impossible to set "equal to None" filter when calling REST API:- https://review.opendev.org/684718 | 09:36 |
| openstackgerrit | ali proposed openstack/mistral master: It's impossible to set "equal to None" filter when calling REST API:- https://review.opendev.org/684718 | 09:40 |
| openstackgerrit | ali proposed openstack/mistral master: It's impossible to set "equal to None" filter when calling REST API:- https://review.opendev.org/684718 | 09:41 |
| openstackgerrit | ali proposed openstack/mistral master: It's impossible to set "equal to None" filter when calling REST API https://review.opendev.org/684718 | 10:09 |
| openstackgerrit | Merged openstack/mistral master: Fix "root_execution" lazy loading issue and refactor execution.py https://review.opendev.org/685899 | 10:26 |
| rakhmerov | please review https://review.opendev.org/#/c/679782/, should be pretty useful additional tool for many people | 10:42 |
| *** pgaxatte has quit IRC | 10:55 | |
| openstackgerrit | Merged openstack/mistral stable/train: Add the explicit dependency from kombu https://review.opendev.org/686354 | 11:37 |
| *** pgaxatte has joined #openstack-mistral | 11:52 | |
| d0ugal | rakhmerov: https://zuul.opendev.org/t/openstack/build/321e651dbbec4462add2a9e84eaa0b0b/log/logs/undercloud/var/log/containers/mistral/engine.log.txt.gz#68-94 | 12:16 |
| d0ugal | so the original exception is coming from multiprocessing, so I think that confirms my theory | 12:17 |
| rakhmerov | yeah | 12:17 |
| rakhmerov | seems like | 12:17 |
| d0ugal | rakhmerov: I think the cookiecutter might be better in Mistral lib? | 12:18 |
| rakhmerov | d0ugal: hm... | 12:18 |
| rakhmerov | :) | 12:18 |
| d0ugal | but I don't really mind | 12:18 |
| rakhmerov | I actually think you're right ) | 12:18 |
| rakhmerov | leave a comment please | 12:18 |
| d0ugal | Makes sense to keep everything external developers need in one place | 12:18 |
| d0ugal | will do | 12:18 |
| rakhmerov | yes | 12:19 |
| rakhmerov | it's easy to move it over there | 12:19 |
| d0ugal | yup | 12:19 |
| rakhmerov | eyalb1: ^ | 12:19 |
| rakhmerov | eyalb1: d0ugal is making a good point | 12:19 |
| eyalb1 | ok | 12:19 |
| eyalb1 | so I should create a tools directory under mistral-lib ? | 12:20 |
| d0ugal | Yeah, I think so | 12:20 |
| eyalb1 | but mistral-lib is a library which is installed as a dependency for mistral | 12:21 |
| eyalb1 | and the tools will be under /usr/lib/python... | 12:22 |
| eyalb1 | how can someone can find it | 12:22 |
| d0ugal | Documentation? | 12:22 |
| eyalb1 | today when you install mistral the tools folder are not installed | 12:23 |
| d0ugal | Indeed, I wouldn't expect it to be as it isn't really part of the Python code | 12:23 |
| d0ugal | I forget how you use a cookiecutter template - doesn't it get it directly from git? | 12:23 |
| eyalb1 | it can but doesn't have to | 12:24 |
| eyalb1 | you give it a url | 12:24 |
| eyalb1 | it can be a files path | 12:24 |
| d0ugal | Another option is to create a dedicated repo for it | 12:29 |
| d0ugal | https://github.com/openstack/?utf8=%E2%9C%93&q=cookie&type=&language= | 12:29 |
| d0ugal | ... but that is a less trivial process :) | 12:29 |
| rakhmerov | yeahy | 12:34 |
| rakhmerov | I'd avoid creating a new repo for now ) | 12:34 |
| rakhmerov | well, so | 12:34 |
| rakhmerov | we already have the "tools" folder under mistral | 12:34 |
| rakhmerov | that has a number of scripts etc. | 12:35 |
| rakhmerov | so may be let's leave it there for now? Although I understand that it's not fully right | 12:35 |
| d0ugal | but mostly tooks for working on mistral itself | 12:35 |
| rakhmerov | hm.. yeah | 12:35 |
| d0ugal | but it can stay there, I don't really mind | 12:35 |
| rakhmerov | Eyal's argument about mistral-lib being a library makes sense to me | 12:36 |
| rakhmerov | well, I guess if it's documented well may be it's not a problem to have it under lib | 12:36 |
| rakhmerov | eyalb1: what do you think? | 12:36 |
| rakhmerov | or there's any serious issues with that? | 12:37 |
| rakhmerov | maybe I'm missing something | 12:37 |
| rakhmerov | I guess we can make these all artefacts available on the pythonpath, right? | 12:38 |
| eyalb1 | like you saud we allready have a tools folder with scripts | 12:38 |
| rakhmerov | if so, we can just document it well | 12:39 |
| eyalb1 | people ususally clone the mistral project not the mistral-lib | 12:39 |
| eyalb1 | so they won't see the tools if it is in the mistral-lib | 12:39 |
| eyalb1 | they have to clone it | 12:40 |
| eyalb1 | best solution is to create a repo for it but like you said its a "process" | 12:40 |
| rakhmerov | so | 12:41 |
| rakhmerov | I'd suggest we leave it in mistral for now | 12:41 |
| rakhmerov | we'll be doing a lot of refactoring related to actions, repos etc. etc. | 12:42 |
| rakhmerov | and we'll find a better place for it | 12:42 |
| rakhmerov | d0ugal: sounds ok for you? | 12:42 |
| d0ugal | sure | 12:42 |
| rakhmerov | because we're going to change the ways how we create actions anyway | 12:42 |
| rakhmerov | what we have now is not ok (IMO) | 12:43 |
| openstackgerrit | ali proposed openstack/mistral master: Task publish is ignored silently:- https://review.opendev.org/684711 | 13:19 |
| openstackgerrit | ali proposed openstack/mistral master: It's impossible to set "equal to None" filter when calling REST API https://review.opendev.org/684718 | 13:54 |
| openstackgerrit | ali proposed openstack/mistral master: It's impossible to set "equal to None" filter when calling REST API https://review.opendev.org/684718 | 14:02 |
| *** ricolin has joined #openstack-mistral | 14:17 | |
| *** pgaxatte has quit IRC | 14:30 | |
| *** eyalb1 has quit IRC | 14:48 | |
| *** openstackgerrit has quit IRC | 17:21 | |
| *** ricolin has quit IRC | 17:35 | |
| *** jtomasek has quit IRC | 23:07 | |
| *** jtomasek has joined #openstack-mistral | 23:42 | |
| *** jtomasek_ has joined #openstack-mistral | 23:51 | |
| *** jtomasek_ has quit IRC | 23:54 | |
| *** jtomasek has quit IRC | 23:54 | |
| *** jtomasek has joined #openstack-mistral | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!