| *** bobh has joined #openstack-mistral | 00:38 | |
| *** zhurong has joined #openstack-mistral | 00:49 | |
| *** zhurong has quit IRC | 00:50 | |
| *** bobh has quit IRC | 01:15 | |
| *** bobh has joined #openstack-mistral | 01:16 | |
| *** bobh has quit IRC | 01:34 | |
| *** rakhmerov has quit IRC | 01:38 | |
| *** bobh has joined #openstack-mistral | 02:22 | |
| *** bobh has quit IRC | 02:30 | |
| *** rakhmerov has joined #openstack-mistral | 04:01 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Remove remaining references to the rpc_backend https://review.openstack.org/609426 | 04:15 |
|---|---|---|
| *** akovi has joined #openstack-mistral | 04:43 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Remove remaining references to the rpc_backend https://review.openstack.org/609426 | 05:02 |
| *** hardikjasani has joined #openstack-mistral | 05:02 | |
| *** pgaxatte has quit IRC | 05:06 | |
| *** pgaxatte has joined #openstack-mistral | 05:10 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Remove remaining references to the rpc_backend https://review.openstack.org/609426 | 05:17 |
| akovi | rakhmerov: Hi Renat! If it's so broken and seemingly nobody has time to iron it out, does it make sense to keep the kombu driver? | 05:27 |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral master: Remove remaining references to the rpc_backend https://review.openstack.org/609426 | 06:08 |
| rakhmerov | akovi: hi Andras, I think I fixed it now | 06:08 |
| rakhmerov | there was just one failing test and I removed it a couple of minutes ago because it's also not relevant | 06:09 |
| rakhmerov | so it should be fine now | 06:09 |
| *** apetrich has quit IRC | 06:20 | |
| akovi | thanks. My question is more towards wouldn't it be beneficial to remove the implementation altogether? Just asking. | 06:21 |
| *** jrist has joined #openstack-mistral | 06:33 | |
| *** apetrich has joined #openstack-mistral | 06:43 | |
| rakhmerov | akovi: hm.. maybe but I'd leave it for now. Still hoping that I'll get back to improving it later | 06:46 |
| openstackgerrit | Hardik Jasani proposed openstack/mistral master: Fix next link in get resource list rest API https://review.openstack.org/608677 | 07:05 |
| *** shardy has joined #openstack-mistral | 07:06 | |
| *** jrist has quit IRC | 07:09 | |
| rakhmerov | d0ugal: hi, https://review.openstack.org/609426 is not done but I'm afraid we got kind of a CI deadlock | 07:12 |
| rakhmerov | it will be failing because of the Rally gate and the patch that fixes the Rally gate will be failing because of these deprecation removals | 07:12 |
| rakhmerov | but let's see.. | 07:12 |
| rakhmerov | we somehow need to push one of them whatever it takes | 07:13 |
| rakhmerov | ... is now done ... | 07:13 |
| rakhmerov | typo | 07:13 |
| pgaxatte | hello | 07:31 |
| pgaxatte | is there a nice way to validate a workflow's DSL without talking to the API? | 07:38 |
| pgaxatte | I'm guessing I'll have to import mistral and use the server's code? | 07:38 |
| d0ugal | rakhmerov: Here now. Will that patch pass CI? | 08:08 |
| rakhmerov | d0ugal: hah.. it seems like Rally gate will fail | 08:11 |
| rakhmerov | again.. | 08:11 |
| rakhmerov | pgaxatte: hi, in order to answer your question we probably need to understand your use case | 08:12 |
| rakhmerov | pgaxatte: Why is API not a good way for you? | 08:12 |
| d0ugal | rakhmerov: so maybe disable the rally voting in that patch? Then I'll enable it again and switch to the lighter workflow in my other patch? | 08:12 |
| rakhmerov | d0ugal: well, it just actually passed | 08:13 |
| d0ugal | oh, good | 08:13 |
| d0ugal | so it sis | 08:13 |
| d0ugal | oops | 08:13 |
| rakhmerov | d0ugal: yeah, we can do that, right, if needed | 08:13 |
| d0ugal | so it did | 08:13 |
| rakhmerov | do we have "rally" in the "gate" jobs too? | 08:14 |
| d0ugal | rakhmerov: Thanks for updating it btw, I was getting a bit tired and confused yesterday afternoon :) | 08:14 |
| d0ugal | rakhmerov: Yeah, we do | 08:14 |
| rakhmerov | if not then we're good | 08:14 |
| rakhmerov | aah.. damn | 08:14 |
| rakhmerov | d0ugal: np, sure | 08:14 |
| pgaxatte | rakhmerov: I'd like to validate the workbooks/workflows before packaging them for deployment and I'm going to do that in a CI/CD environment | 08:17 |
| rakhmerov | aah.. I see | 08:18 |
| pgaxatte | rakhmerov: I don't want to depend on a mistral server running somewhere or running it myself before checking | 08:18 |
| d0ugal | You can do it by importing mistral, but it isn't really a supported use | 08:18 |
| rakhmerov | maybe worth installing the simplest possible version of Mistral in your CI? | 08:18 |
| rakhmerov | pgaxatte: yeah, using server side code is a dangerous idea, you should understand that | 08:19 |
| rakhmerov | because it may change any time w/o warning users | 08:19 |
| d0ugal | but in CI/CD it could be okay. Worst case you can disable that check | 08:19 |
| pgaxatte | rakhmerov: of course but I'd be running a fixed version of the server code | 08:19 |
| d0ugal | and if you pin mistral version in CI it wouldn't be that risky either - but upgrading could be hard if things change. | 08:20 |
| rakhmerov | pgaxatte: ok then | 08:20 |
| pgaxatte | the same as we'll use on production side where i'll deploy the workflows | 08:20 |
| d0ugal | I started writing a tool to check workflows, but I didn't get very far with it :) https://github.com/d0ugal/mistral-lint | 08:21 |
| rakhmerov | d0ugal: btw, I was thinking about it some time ago. We could probably maintain workflow language things in a separate module where we could open an API | 08:21 |
| rakhmerov | as a library | 08:21 |
| d0ugal | rakhmerov: +1 | 08:21 |
| pgaxatte | rakhmerov: yep I was thinking the same while working on this project | 08:21 |
| rakhmerov | I was considering something else but it'd be aligned with that too | 08:21 |
| pgaxatte | besides the DSL version has nothing to do with mistral's api version | 08:21 |
| pgaxatte | so these could be separate | 08:22 |
| d0ugal | pgaxatte: well, in theroy. We are very bad at managing the DSL version | 08:22 |
| d0ugal | it has been 2.0 for a long time and we have added many features :/ | 08:22 |
| rakhmerov | yes, kind of | 08:22 |
| pgaxatte | d0ugal: yeah I remembered some discussions about that :D | 08:22 |
| rakhmerov | it's still 2.0 although we made lots of changes | 08:22 |
| rakhmerov | what we 100% obey though is backward compatibility | 08:22 |
| rakhmerov | all of the changes have been compatible | 08:23 |
| pgaxatte | so keeping 2.0 is not such a bad idea since there are no breaking changes | 08:23 |
| d0ugal | Right | 08:23 |
| pgaxatte | but you can't anticipate if some feature will be present | 08:23 |
| d0ugal | It means the first 2.0 workflow should still work, but a rocky workflow will never work with a newton server (I forget when 2.0 was introduced?) | 08:24 |
| rakhmerov | to maintain DSL versions I think we exactly need a separate lib plus some additional code that manages versions | 08:24 |
| rakhmerov | d0ugal: ~ 4 years ago | 08:24 |
| rakhmerov | :) | 08:24 |
| pgaxatte | yeah the separation would make the changes in version easier to manage | 08:24 |
| d0ugal | We maybe don't need to support every version, but if we increase the minor number each release it would at least make it clearer | 08:25 |
| rakhmerov | yep | 08:25 |
| d0ugal | i.e. my workflow is "2.1" then it works on a server that supports "2.1" or higher | 08:25 |
| rakhmerov | I wish we could do that | 08:25 |
| d0ugal | Why can't we? :) | 08:25 |
| rakhmerov | as usually | 08:25 |
| rakhmerov | priorities and limited resources | 08:25 |
| pgaxatte | lack of time/resources? | 08:25 |
| d0ugal | Right | 08:25 |
| pgaxatte | :D | 08:25 |
| rakhmerov | + not very stable team | 08:25 |
| d0ugal | That small issue :) | 08:25 |
| pgaxatte | d0ugal, I met a french colleague of you recently (Sylvain Bauzas, nova core-team) who told me more and more people at RedHat are getting interested in Mistral | 08:28 |
| pgaxatte | didn't have much time to discuss this further, but do you know if Mistral is taking off at RedHat? | 08:29 |
| d0ugal | pgaxatte: I think we only use it in TripleO, so I don't feel like it is taking off :) | 08:29 |
| d0ugal | but it is very important at the moment for connecting together everything in our main OpenStack product | 08:30 |
| pgaxatte | d0ugal: arf so it would be hard to justify bringing more people in? :D | 08:35 |
| pgaxatte | on Mistral I mean | 08:35 |
| pgaxatte | not in RedHat you guys are big enough :P | 08:35 |
| d0ugal | pgaxatte: lol, we are pretty big. | 08:41 |
| d0ugal | ... but we are always hiring too ;) | 08:45 |
| pgaxatte | d0ugal: gotta keep growing :D | 09:00 |
| pgaxatte | any of you guys are going to Berlin for the summit, rakhmerov d0? | 09:07 |
| pgaxatte | d0ugal ^ | 09:08 |
| rakhmerov | pgaxatte: unfortunately, I'm not | 09:08 |
| rakhmerov | will do my best to be in Denver in april | 09:08 |
| pgaxatte | ahh too bad | 09:08 |
| rakhmerov | yeah.. | 09:08 |
| rakhmerov | sorry | 09:08 |
| *** jistr is now known as jistr|call | 09:12 | |
| d0ugal | I can't make it either | 09:18 |
| d0ugal | I believe apetrich will be there! | 09:18 |
| apetrich | rakhmerov, d0ugal I hope so. I can see the venue from my balcony | 09:19 |
| rakhmerov | apetrich: ooh, really? :)) | 09:22 |
| rakhmerov | that's awesome! | 09:23 |
| pgaxatte | apetrich: so no excuse to not go then | 09:23 |
| pgaxatte | ;) | 09:23 |
| rakhmerov | yeaah :)) | 09:23 |
| apetrich | :) | 09:23 |
| apetrich | I hope I don't have to scream my talk all the way from here | 09:24 |
| apetrich | it is like 3 subways stations away but because you have to do an L | 09:24 |
| *** gkadam has quit IRC | 09:25 | |
| rakhmerov | if needed, scream man! :) | 09:26 |
| rakhmerov | then even more people will hear you :) | 09:27 |
| rakhmerov | and get to know about OpenStack/Mistral/TripleO | 09:27 |
| d0ugal | lol | 09:28 |
| *** jtomasek has joined #openstack-mistral | 09:46 | |
| *** jistr|call is now known as jistr | 11:00 | |
| openstackgerrit | Hardik Jasani proposed openstack/mistral-tempest-plugin master: Add test for pagination url https://review.openstack.org/608686 | 11:01 |
| openstackgerrit | Oleg Ovcharuk proposed openstack/mistral master: Refactor size limit check mechanism https://review.openstack.org/608469 | 11:09 |
| openstackgerrit | Hardik Jasani proposed openstack/mistral-tempest-plugin master: Add test for pagination url https://review.openstack.org/608686 | 11:10 |
| openstackgerrit | Oleg Ovcharuk proposed openstack/mistral master: Divide yaml input to save it into definitions separately in case of create/update multiple workflows from one yaml. https://review.openstack.org/603208 | 11:26 |
| openstackgerrit | Oleg Ovcharuk proposed openstack/mistral master: Divide yaml input to save it into definitions separately. https://review.openstack.org/603208 | 11:29 |
| *** toure|gone is now known as toure | 12:01 | |
| *** bobh has joined #openstack-mistral | 13:05 | |
| jlejeune | hi guys | 13:05 |
| jlejeune | I've a small question for you | 13:05 |
| jlejeune | I tried to create my first custom action | 13:05 |
| jlejeune | I need to get the root_execution_id value in my action | 13:06 |
| jlejeune | do you know if there is a way to get it inside actions in sub workflows ? | 13:07 |
| akovi | jlejeune: I don't have a mistral running at hand so I can't test the following | 13:16 |
| akovi | you'll have the context that is created in mistral_lib/actions/context.py ExecutionContext | 13:17 |
| akovi | unfortunately that does not contain the root id | 13:17 |
| akovi | to check what is there, I would start a mistral locally and stop the execution in mistral.executors.default_executor.DefaultExecutor#run_action | 13:18 |
| jlejeune | actually I already dig into ExecutionContext class | 13:19 |
| jlejeune | indeed, this class doesn't have any root_execution_id key... | 13:20 |
| akovi | and I think the context doesn't have it either | 13:22 |
| jlejeune | yep... | 13:22 |
| akovi | this is what can be checked in the run_action | 13:22 |
| akovi | it's constructed here mistral.engine.actions.PythonAction#_prepare_execution_context | 13:25 |
| akovi | so definitely, the root is not there | 13:25 |
| jlejeune | too bad :/ | 13:25 |
| akovi | it can be added and you can use it in the next release :D | 13:25 |
| jlejeune | :) | 13:26 |
| akovi | can you please file a bug | 13:26 |
| *** hardikjasani has quit IRC | 13:32 | |
| *** jrist has joined #openstack-mistral | 13:33 | |
| *** jrist has quit IRC | 14:20 | |
| *** jtomasek has quit IRC | 14:36 | |
| *** openstackgerrit has quit IRC | 14:58 | |
| *** openstackgerrit has joined #openstack-mistral | 14:58 | |
| *** openstackgerrit has quit IRC | 15:22 | |
| d0ugal | Yeah, we should add it. That is probably my fault :) | 15:24 |
| *** akovi has quit IRC | 15:46 | |
| *** jrist has joined #openstack-mistral | 15:50 | |
| *** jrist has quit IRC | 15:59 | |
| bobh | Just ran into a problem with task name - there do not seem to be any restrictions on task name characters (besides length), but the OnClauseSpec now requires the task name to be \w+ - that means I can define a task name with a "-" in it and it will work, but I can't on-anything to that task because the spec disallows it | 17:31 |
| bobh | so do we restrict task names to \w+ or expand the OnClauseSpec definition of a task name to include a "-"? | 17:32 |
| *** shardy has quit IRC | 17:47 | |
| bobh | I think the on-clase spec is too restrictive, YAML will let you use just about any string for a key name | 17:49 |
| *** openstackgerrit has joined #openstack-mistral | 19:34 | |
| openstackgerrit | Bob Haddleton proposed openstack/mistral master: Update OnClauseSPec task name criteria to non-whitespace characters https://review.openstack.org/609794 | 19:34 |
| *** bobh has quit IRC | 21:05 | |
| *** bobh has joined #openstack-mistral | 21:07 | |
| *** bobh has quit IRC | 21:11 | |
| *** bobh has joined #openstack-mistral | 21:40 | |
| *** bobh has quit IRC | 21:44 | |
| *** bobh has joined #openstack-mistral | 21:49 | |
| *** bobh has quit IRC | 22:12 | |
| *** openstackgerrit has quit IRC | 22:19 | |
| *** openstackgerrit has joined #openstack-mistral | 22:19 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!