| *** bobh has joined #openstack-vitrage | 02:32 | |
| *** bobh has quit IRC | 03:41 | |
| *** eyalb has joined #openstack-vitrage | 06:04 | |
| *** eyalb1 has joined #openstack-vitrage | 06:06 | |
| *** eyalb has quit IRC | 06:08 | |
| *** ifat_afek has joined #openstack-vitrage | 07:02 | |
| openstackgerrit | Idan Hefetz proposed openstack/vitrage master: Api-worker sync threads - lock graph from changing during API calls. https://review.openstack.org/626566 | 07:35 |
|---|---|---|
| openstackgerrit | Merged openstack/python-vitrageclient master: Use template for lower-constraints https://review.openstack.org/626744 | 07:51 |
| openstackgerrit | Merged openstack/vitrage master: Add Trove cluster capacity autoscaling example https://review.openstack.org/627017 | 09:07 |
| openstackgerrit | Eyal proposed openstack/vitrage master: fix some pycharm warning https://review.openstack.org/627063 | 10:03 |
| openstackgerrit | Merged openstack/vitrage master: mock driver - generate consistent uuids https://review.openstack.org/626558 | 10:29 |
| openstackgerrit | Eyal proposed openstack/vitrage master: fix some pycharm warning https://review.openstack.org/627063 | 11:02 |
| *** e0ne has joined #openstack-vitrage | 11:09 | |
| e0ne | hi team. can anybody please help me with tempest tests? | 11:19 |
| ifat_afek | e0ne: hi | 11:20 |
| ifat_afek | what is the problem? | 11:20 |
| e0ne | not really a problem.. I just want to get some information | 11:21 |
| ifat_afek | sure | 11:21 |
| e0ne | why do we have separate vitrage-dsvm-api-py27 vitrage-dsvm-datasources-py27 vitrage-dsvm-mock-py27 jobs instead of one? | 11:22 |
| ifat_afek | the purpose of the mock is to test the graph on large data. it is used as the only datasource (in addition to doctor, for generating alarms) instead of all the regular openstack datasources | 11:23 |
| ifat_afek | because it is a different configuraiton we run it on a different job | 11:24 |
| ifat_afek | different ‘types’ value in vitrage.conf + specific mock configuration | 11:24 |
| e0ne | it makes sense, thanks | 11:25 |
| e0ne | we can't do it in one job if it requires some config changes | 11:25 |
| ifat_afek | regarding the datasources vs. api - I’m not sure. I think one developer wrote it once and we just keep adding tests to the jobs. I’m not sure there is a real reason for the separation | 11:25 |
| * e0ne like to dig into the tests because it helps to understand project better | 11:26 | |
| ifat_afek | for the mock - maybe we could if we tried harder, but I’m not sure it makes sense. for the datasources and api, might be worth checking | 11:26 |
| ifat_afek | go for it :-) | 11:26 |
| e0ne | I do:) | 11:26 |
| e0ne | and one maybe a stupid question | 11:27 |
| ifat_afek | when we started writing the temepst tests, three years ago, I think we have made several mistakes. then we didn’t have time to fix them… but I guess we should | 11:27 |
| e0ne | I didn't try such configuration before (I'm deploying it now), but looking throw the code, it looks like we can't run vitrage and tempest plugin on different nodes. am I right? | 11:28 |
| e0ne | ifat_afek: it's great that you and team understand that we need to fix it :) | 11:28 |
| ifat_afek | right… because the tempest depends on vitrage code and uses it. but once this is fixed, do you see any other problem? | 11:29 |
| e0ne | it's not the issue, because it only requires to install vitrage ad get config | 11:30 |
| e0ne | but things like https://github.com/openstack/vitrage-tempest-plugin/blob/9da198fcfe9200eae97b93701c98a7096514ffcb/vitrage_tempest_plugin/tests/utils.py#L41https://github.com/openstack/vitrage-tempest-plugin/blob/ff69d62f3ff6e4f5b1e5db20464d379d7018c770/vitrage_tempest_plugin/tests/common/vitrage_utils.py#L121 and https://github.com/openstack/vitrage-tempest-plugin/blob/9da198fcfe9200eae97b93701c98a7096514ffcb/vitrage_tempest_plugin/tests/base.py#L6 | 11:32 |
| e0ne | 7 seems to be more critical, IMO | 11:32 |
| ifat_afek | yes… I guess you are right | 11:35 |
| e0ne | ifat_afek: according to https://storyboard.openstack.org/#!/story/2004053 | 11:35 |
| ifat_afek | what do you mean by ‘7 seems to be more critical’? what 7? | 11:35 |
| e0ne | ifat_afek: it was a message size limit in IRC :( | 11:36 |
| e0ne | it's the end of the previous message | 11:36 |
| ifat_afek | indeed, line 67 makes more sense than line 6 :-) | 11:37 |
| e0ne | :) | 11:37 |
| e0ne | ifat_afek: I'm not sure that I'll be able to finish #2004053 this release, so to speedup this a bit I'm going to use python-clients still with a TODO node but decouple tempest plugin from vitrage itself | 11:38 |
| ifat_afek | and you are right, the tempest tests should only use python-vitrageclient. on the other hand, we have important use cases that require graph restart (to make sure persistency works, for example)… so I’m not sure about it | 11:38 |
| e0ne | does it make sed? | 11:38 |
| e0ne | s/sed/sense | 11:38 |
| ifat_afek | sounds fine to me. any step towards the separation is welcome | 11:38 |
| e0ne | :) | 11:38 |
| ifat_afek | I think it’s a lot of work to finish it... | 11:38 |
| e0ne | I agree with you | 11:39 |
| e0ne | honestly, I started this for two reasons: | 11:39 |
| e0ne | 1) to become more familiar with how vitrage works | 11:39 |
| e0ne | and 2) to finish upgrades stuff | 11:39 |
| ifat_afek | and are the changes that you are doing now enough for the upgrade? | 11:40 |
| e0ne | not yet, but I'm about 2-3 patches before it | 11:41 |
| ifat_afek | good :-) | 11:42 |
| e0ne | I need toremove vitrage conf usages | 11:42 |
| ifat_afek | do you have a change in vitrage that depends on all of your tempest changes? or do you check them one by one? | 11:42 |
| e0ne | here is it: https://review.openstack.org/#/c/612652/ | 11:43 |
| e0ne | it would be good to add the same jobs for tempest plugin too | 11:43 |
| ifat_afek | but you have 3 changes in the tempest, right? | 11:44 |
| ifat_afek | and only one in vitrage? | 11:44 |
| ifat_afek | what do you mean by adding the same jobs? so the tempest changes will run the tempest as well? | 11:44 |
| e0ne | there is a chain of tempest changes | 11:45 |
| e0ne | exactly | 11:45 |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/vitrage master: Do Not Review. Test tempest changes https://review.openstack.org/612652 | 11:46 |
| e0ne | ifat_afek: you ca see at zuul http://zuul.openstack.org/status now that my vitrage patch will be testes with all 3 patches to tempest | 11:47 |
| ifat_afek | I think our problem is that there is a cyclic dependency between vitrage and the tempest. if you run tempest in both places, it might cause problems - e.g. if you change an api and you should change both vitrage and the tempest. | 11:47 |
| e0ne | oh.. you're right | 11:48 |
| ifat_afek | personally I don’t understand the decision to split the tempest tests from the main code. it seems to only mess things up… (although it also forces us to write the code in a correct way) | 11:48 |
| ifat_afek | or at least I don’t understand why the tempest tests are branchless | 11:49 |
| ifat_afek | but that’s a different discussion | 11:49 |
| e0ne | they are branchless because it's supposed that you'll be able to run the same tests against different versions of the openstack | 11:50 |
| e0ne | e.g. it's useful to run the same tests after cloud upgrade | 11:50 |
| e0ne | that's why it's recommended to not use any python-*client in tempest | 11:51 |
| ifat_afek | for upgrade it makes sense. but if we add a new api then it won’t run against an older version | 11:51 |
| ifat_afek | *not* use python client? | 11:52 |
| e0ne | you can use config with feature flags | 11:52 |
| e0ne | ifat_afek: yep | 11:52 |
| ifat_afek | you got me confused. shouldn’t the tempest call python-vitrageclient, python-novaclient, etc.? if not, then what? | 11:53 |
| e0ne | no, it should not | 11:53 |
| e0ne | but as I know nobody checks if for plugins | 11:54 |
| e0ne | e.g. cinder client implementation for tempest https://github.com/openstack/tempest/blob/master/tempest/lib/services/volume/v3/volumes_client.py | 11:54 |
| ifat_afek | regarding the changes - I see the chain in zuul, but I don’t see any depends-on in https://review.openstack.org/#/c/627052/. what am I missing? | 11:54 |
| e0ne | tempest should call API directly | 11:54 |
| ifat_afek | right… eyalb explained it to me now. I knew that there is a common way to use clients in tempest tests, but I assumed these are simply methods to access the regular python clients | 11:55 |
| e0ne | here is a git chain: master branch => Use Tempest config => Add constants for datasources => Delete unused code | 11:56 |
| ifat_afek | so you pushed 3 commits one on top of the other? wow… we are afraid of doing such things ;-) | 11:56 |
| e0ne | https://docs.openstack.org/infra/manual/developers.html#adding-a-dependency | 11:57 |
| e0ne | ifat_afek: yes. why you you afraid to do it? | 11:57 |
| ifat_afek | cool! we could have used it several times | 11:57 |
| e0ne | *why are you | 11:57 |
| ifat_afek | because git sometime messes things up if you don’t do it right… and we weren’t aware of this specific option | 11:58 |
| ifat_afek | and also - what happens if you need to make a change in the middle commit? | 11:58 |
| e0ne | I can do it with git rebase | 11:59 |
| ifat_afek | ok, I learned something new :-) | 11:59 |
| e0ne | I'm glad to help you | 12:00 |
| ifat_afek | :-) | 12:03 |
| *** e0ne has quit IRC | 12:10 | |
| *** e0ne has joined #openstack-vitrage | 12:12 | |
| openstackgerrit | Idan Hefetz proposed openstack/vitrage master: Api-worker lock graph from changing during API calls https://review.openstack.org/626566 | 12:28 |
| openstackgerrit | Idan Hefetz proposed openstack/vitrage master: Fix topology api crashes https://review.openstack.org/627072 | 12:28 |
| openstackgerrit | Idan Hefetz proposed openstack/vitrage master: Api-worker sync threads - lock graph from changing during API calls. https://review.openstack.org/626566 | 12:31 |
| *** e0ne has quit IRC | 12:47 | |
| *** e0ne has joined #openstack-vitrage | 12:51 | |
| *** ifat_afek has quit IRC | 13:02 | |
| *** ifat_afek has joined #openstack-vitrage | 13:03 | |
| *** e0ne has quit IRC | 13:18 | |
| *** e0ne has joined #openstack-vitrage | 13:28 | |
| *** e0ne has quit IRC | 13:30 | |
| *** ifat_afek has quit IRC | 13:41 | |
| *** ifat_afek has joined #openstack-vitrage | 13:58 | |
| openstackgerrit | Bartosz Zurkowski proposed openstack/vitrage master: Add Monasca datasource https://review.openstack.org/622899 | 14:50 |
| *** eyalb1 has quit IRC | 15:05 | |
| openstackgerrit | Merged openstack/vitrage master: Api-worker sync threads - lock graph from changing during API calls. https://review.openstack.org/626566 | 15:06 |
| openstackgerrit | Bartosz Zurkowski proposed openstack/vitrage master: Add Monasca datasource https://review.openstack.org/622899 | 15:06 |
| openstackgerrit | Idan Hefetz proposed openstack/vitrage master: Fix topology api crashes https://review.openstack.org/627072 | 15:06 |
| *** bobh has joined #openstack-vitrage | 15:49 | |
| *** bobh has quit IRC | 15:53 | |
| *** bobh_ has joined #openstack-vitrage | 15:53 | |
| *** bobh_ has quit IRC | 16:53 | |
| *** bobh has joined #openstack-vitrage | 17:12 | |
| *** bobh has quit IRC | 17:16 | |
| *** bobh has joined #openstack-vitrage | 17:28 | |
| openstackgerrit | Bartosz Zurkowski proposed openstack/vitrage master: Add Monasca datasource https://review.openstack.org/622899 | 17:34 |
| *** ifat_afek has quit IRC | 17:45 | |
| *** bobh has quit IRC | 18:13 | |
| *** e0ne has joined #openstack-vitrage | 18:56 | |
| *** e0ne has quit IRC | 18:57 | |
| *** e0ne has joined #openstack-vitrage | 20:07 | |
| *** e0ne has quit IRC | 20:13 | |
| *** e0ne has joined #openstack-vitrage | 20:51 | |
| *** e0ne has quit IRC | 20:51 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!