| *** bhagyashris is now known as bhagyashris|out | 03:30 | |
| opendevreview | Mridula Joshi proposed openstack/glance-specs master: Append new tags to existing tags https://review.opendev.org/c/openstack/glance-specs/+/813429 | 06:34 |
|---|---|---|
| opendevreview | Mridula Joshi proposed openstack/glance master: Fix overwriting of existing tags while creating new tags https://review.opendev.org/c/openstack/glance/+/804966 | 06:46 |
| opendevreview | OpenStack Proposal Bot proposed openstack/glance master: Imported Translations from Zanata https://review.opendev.org/c/openstack/glance/+/813565 | 07:04 |
| -opendevstatus- NOTICE: zuul was stuck processing jobs and has been restarted. pending jobs will be re-enqueued | 10:02 | |
| opendevreview | Andre Aranha proposed openstack/glance master: Add qemu-img to test dependencies https://review.opendev.org/c/openstack/glance/+/813787 | 11:18 |
| opendevreview | Andre Aranha proposed openstack/glance master: DNM: Add fips check jobs https://review.opendev.org/c/openstack/glance/+/790536 | 11:20 |
| opendevreview | Andre Aranha proposed openstack/glance master: DNM: Add fips check jobs https://review.opendev.org/c/openstack/glance/+/790536 | 11:20 |
| opendevreview | Pranali Deore proposed openstack/glance-tempest-plugin master: Implement API protection testing for metadef resource types https://review.opendev.org/c/openstack/glance-tempest-plugin/+/802792 | 12:14 |
| opendevreview | Pranali Deore proposed openstack/glance-tempest-plugin master: Implement API protection testing for metadef properties https://review.opendev.org/c/openstack/glance-tempest-plugin/+/802794 | 12:14 |
| opendevreview | Pranali Deore proposed openstack/glance-tempest-plugin master: Implement API protection testing for metadef tags https://review.opendev.org/c/openstack/glance-tempest-plugin/+/802795 | 12:14 |
| opendevreview | Pranali Deore proposed openstack/glance-tempest-plugin master: Move the inlined utilities at common place https://review.opendev.org/c/openstack/glance-tempest-plugin/+/813625 | 12:14 |
| *** pdeore__ is now known as pdeore|afk | 12:16 | |
| *** bhagyashris|out is now known as bhagyashris|mtg | 13:00 | |
| abhishekk | jokke_, rosmaita, dansmith, croelandt glance weekly meeting at #openstack-meeting in 15 minutes | 13:47 |
| *** pdeore|afk is now known as pdeore | 14:01 | |
| *** pdeore is now known as pdeore|afk | 14:27 | |
| * abhishekk will be back in couple of hours | 14:29 | |
| *** bhagyashris|mtg is now known as bhagyashris|away | 15:03 | |
| opendevreview | Andre Aranha proposed openstack/glance master: Add missing packages to test dependencies https://review.opendev.org/c/openstack/glance/+/813787 | 15:45 |
| dansmith | lbragstad: I didn't notice the lock here in the previous patch: https://review.opendev.org/c/openstack/glance-tempest-plugin/+/802792/17/glance_tempest_plugin/tests/rbac/v2/metadefs/test_resource_types.py | 16:34 |
| dansmith | but if we keep stacking those up it will effectively serialize all these tests | 16:34 |
| dansmith | maybe I missed some conversation about this? | 16:34 |
| lbragstad | true - i think i asked about that lock in a review | 17:10 |
| lbragstad | but i can't remember which one | 17:10 |
| lbragstad | https://review.opendev.org/c/openstack/glance-tempest-plugin/+/802793/8 | 17:12 |
| dansmith | ah, it was abhishekk :) | 17:14 |
| lbragstad | he just showed up, too... must have known we were talking about him | 17:14 |
| abhishekk | no idea at all, need to go through logs :D | 17:14 |
| abhishekk | hmmm, It was me who introduced it after discussion with QA team | 17:15 |
| dansmith | abhishekk: just this: https://review.opendev.org/c/openstack/glance-tempest-plugin/+/802792/17/glance_tempest_plugin/tests/rbac/v2/metadefs/test_resource_types.py | 17:16 |
| abhishekk | yep, saw your comment on mail | 17:16 |
| abhishekk | The problem here is we need to run all the tests serially because all tests are creating namespaces first | 17:17 |
| dansmith | but we wouldn't need to if their names didn't clash right? | 17:17 |
| abhishekk | so if we don't run it one by one then there could be problem | 17:17 |
| abhishekk | get all | 17:17 |
| abhishekk | will be problematic | 17:17 |
| dansmith | we just ignore any that aren't using our namespace right? | 17:18 |
| dansmith | I mean, otherwise you can't run this against a working cloud anyway, which I think is the intent of tempest (although admin-y things might be a gray area) | 17:18 |
| dansmith | self.assertEqual([], [x for x in get_all() if x.startswith(self.our_namespace)]) | 17:19 |
| abhishekk | I am saying if NameSpace and Object tests are running in parallel then get all namespaces might fetch namespaces created by Objects | 17:19 |
| dansmith | sure, | 17:19 |
| dansmith | but if they are named accordingly that won't be a problem | 17:19 |
| dansmith | I mean namespace the Namespaces :) | 17:19 |
| abhishekk | hmm, so you are saying if the namespace is not related to our test we will ignore it ? | 17:20 |
| dansmith | create_namespaces() can use names like 'namespace_%s' % self.__class__.__name__ | 17:20 |
| dansmith | yeah | 17:20 |
| dansmith | otherwise you're just assuming the cloud is empty, else the tests will fail | 17:20 |
| abhishekk | hmm, we can do that | 17:21 |
| dansmith | gmann: isn't tempest supposed to work on a production cloud, when possible? meaning we shouldn't fail if they have some resources on the cloud before the tests start? | 17:22 |
| abhishekk | I think we need to link this chat to review so that pdeore will understand what we were talking | 17:22 |
| dansmith | sure | 17:23 |
| dansmith | lbragstad: maybe you could quickly confirm that your comments from the previous version were addressed on this? https://review.opendev.org/c/openstack/glance-tempest-plugin/+/802792/ | 17:24 |
| abhishekk | @all, just FYI tomorrow is holiday in India so me and pdeore will not be around | 17:28 |
| dansmith | abhishekk: cool, enjoy | 17:29 |
| dansmith | so is she probably already gone until next week? | 17:29 |
| abhishekk | yep | 17:29 |
| dansmith | okay | 17:29 |
| abhishekk | its almost 11 PM now | 17:30 |
| * abhishekk signing out now | 17:45 | |
| dansmith | abhishekk: go! | 17:46 |
| abhishekk | haha | 17:46 |
| lbragstad | yeah - i think https://review.opendev.org/c/openstack/glance-tempest-plugin/+/802792/17 addresses my concerns from ps14 | 18:09 |
| gmann | dansmith: yes, it should work irrespective of what resources we have in testing env. that is why assert are are tempest created resources only, and most of the test create resources with prefix of tempest-* etc so what we do not conflict with resource in production env. If it clash then its test bug | 22:42 |
| dansmith | gmann: ack, okay well, all these metadef tests are violating that principle unfortunately | 22:43 |
| dansmith | thanks for the confirmation | 22:43 |
| gmann | dansmith: for namespace resource and using lock, I remember the discussion that glance cannot create separate resource etc so we need to run them in serial. in that case i suggested to use lock instead of serial run | 22:43 |
| dansmith | gmann: yeah, and I think I was part of that as well, | 22:44 |
| dansmith | however, I think the actual problem is just that the code relies on the cloud being empty, | 22:44 |
| dansmith | which it doesn't need to | 22:44 |
| gmann | humm | 22:44 |
| dansmith | there is no tenant isolation, so everything is global, | 22:45 |
| dansmith | however we do not need to assert empty, create stuff, delete those things, assert empty | 22:45 |
| dansmith | we can create and delete specific resources and assert those are deleted, | 22:45 |
| dansmith | and not self.assertEqual([], client.get_all()) | 22:45 |
| dansmith | and then we don't have to use a lock and serialize every single metadef test :) | 22:45 |
| gmann | k, and test resource name can be used as self.__class__.__name__ + <resource> for easy in debugging | 22:47 |
| dansmith | yeah as I said above and in the review | 22:47 |
| gmann | +1 | 22:47 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!