| *** arantuna has quit IRC | 00:08 | |
| openstackgerrit | Yash Gupta proposed openstack/kuryr-kubernetes master: Refactor DriverBase.get_instance() https://review.openstack.org/596672 | 01:59 |
|---|---|---|
| openstackgerrit | Yash Gupta proposed openstack/kuryr-kubernetes master: Refactor DriverBase.get_instance() https://review.openstack.org/596672 | 02:14 |
| openstackgerrit | Yash Gupta proposed openstack/kuryr-kubernetes master: Use same pool_driver for different pod_vif_drivers https://review.openstack.org/596673 | 02:14 |
| danil | hello, folks. Please review patches with SRIOV functionality: https://review.openstack.org/#/c/512280/39 , https://review.openstack.org/#/c/512281/40 | 04:31 |
| *** aperevalov has joined #openstack-kuryr | 05:10 | |
| *** threestrands has joined #openstack-kuryr | 05:45 | |
| *** threestrands has quit IRC | 05:45 | |
| *** threestrands has joined #openstack-kuryr | 05:45 | |
| *** ccamposr has joined #openstack-kuryr | 07:02 | |
| *** threestrands has quit IRC | 07:25 | |
| *** pcaruana has joined #openstack-kuryr | 07:31 | |
| *** dims has quit IRC | 07:57 | |
| *** dims has joined #openstack-kuryr | 07:57 | |
| *** garyloug has joined #openstack-kuryr | 07:59 | |
| *** celebdor has joined #openstack-kuryr | 08:01 | |
| openstackgerrit | Merged openstack/kuryr master: Use templates for cover and lower-constraints https://review.openstack.org/600925 | 08:40 |
| celebdor | aperevalov: https://review.openstack.org/#/c/600693/1 | 08:45 |
| celebdor | I made some small suggestion to your patch | 08:45 |
| celebdor | otherwise, it looks good to me | 08:45 |
| aperevalov | celebdor: thanks! | 08:46 |
| celebdor | thanks to you for the patch | 08:47 |
| openstackgerrit | Merged openstack/kuryr-libnetwork master: Use templates for cover and lower-constraints https://review.openstack.org/600924 | 09:01 |
| openstackgerrit | Alexey Perevalov proposed openstack/kuryr-kubernetes master: Produce mount volume only if OVS_HOST_PATH is defined https://review.openstack.org/600693 | 10:24 |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-tempest-plugin master: Add Network Policy tests https://review.openstack.org/589521 | 10:38 |
| aperevalov | we tested containerized cni/controller and didn't find required docker version in docs. Is it 1.13? | 12:00 |
| celebdor | aperevalov: just a sec | 12:01 |
| celebdor | aperevalov: I think it is 1.13.1 nowadays | 12:02 |
| *** rh-jelabarre has joined #openstack-kuryr | 12:05 | |
| aperevalov | celebdor: thanks! we have 2 installation in one we're using 1.12.6 and in another one 1.13.1. There is no --env option in 1.12.6 ;) It's better to mention about it in documentation ) | 12:08 |
| celebdor | aperevalov: good idea | 12:20 |
| celebdor | thanks | 12:20 |
| celebdor | how is it in 1.12.6? Just - | 12:21 |
| celebdor | -e ? | 12:21 |
| * celebdor -> lunch | 12:21 | |
| aperevalov | just error message: flag provided but not defined: --env See 'docker exec --help'. | 12:26 |
| aperevalov | and there is no --env as well as -e | 12:26 |
| danil | hello, folks. Please review patches with SRIOV functionality: https://review.openstack.org/#/c/512280/39 , https://review.openstack.org/#/c/512281/40 | 12:34 |
| *** maysams has joined #openstack-kuryr | 12:36 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Clean up watch resources after watcher.stop() https://review.openstack.org/600142 | 12:51 |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: WIP: Add HA gate https://review.openstack.org/588223 | 12:51 |
| aperevalov | just one more question: I got empty CNI_NETNS in env for my containerized kuryr-cni, looks like it's not normal, and seems it's somewhere in my configiration. | 12:57 |
| celebdor | exec? | 13:05 |
| celebdor | oh, right | 13:05 |
| celebdor | in 13 it was added to exec before only on run | 13:05 |
| aperevalov | celebdor: CNI_NETNS was empty during docker exec | 13:07 |
| celebdor | hmmm | 13:07 |
| celebdor | this does sound familiar | 13:07 |
| celebdor | ltomasbo: do you remember ^^ happening | 13:07 |
| ltomasbo | umm, nop, I don't remember that happening to me... | 13:12 |
| celebdor | ltomasbo: do you recall if adding the loopback net conf to /etc/ cni was enough to fix the pods not starting? | 13:17 |
| celebdor | the host networking pods not starting, I mean | 13:17 |
| ltomasbo | celebdor, it was enough to make it ready | 13:20 |
| celebdor | ok | 13:20 |
| ltomasbo | but the pods were not schedule either, due to the missing csr approval (after the timeout) | 13:20 |
| ltomasbo | celebdor, ^^ | 13:20 |
| celebdor | ltomasbo: but that's because you put the conf file after the timeout | 13:21 |
| celebdor | :P | 13:21 |
| ltomasbo | yep | 13:21 |
| ltomasbo | xD | 13:21 |
| openstackgerrit | Merged openstack/kuryr-tempest-plugin master: Add Network Policy tests https://review.openstack.org/589521 | 13:22 |
| *** garyloug_ has joined #openstack-kuryr | 13:31 | |
| celebdor | ltomasbo: https://github.com/celebdor/origin/tree/d_csr | 13:32 |
| celebdor | I believe this should do it | 13:33 |
| ltomasbo | umm, it actually could be... | 13:33 |
| ltomasbo | I'm stacking at the moment to test another thing, as soon as it finished, I'll give it a try to this one | 13:34 |
| celebdor | very well | 13:34 |
| *** garyloug has quit IRC | 13:34 | |
| *** garyloug_ has quit IRC | 13:35 | |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-tempest-plugin master: Remove direct calls to os_client for SGs https://review.openstack.org/599948 | 14:03 |
| aperevalov | I subscribed to [Openstack] ML, but I didn't see any cancelation of todays meeting | 14:06 |
| celebdor | aperevalov: openstack-dev? | 14:22 |
| *** garyloug_ has joined #openstack-kuryr | 14:27 | |
| *** maysams has quit IRC | 14:46 | |
| *** hongbin has joined #openstack-kuryr | 14:47 | |
| celebdor | is there any case of https://review.openstack.org/#/c/600142/3/kuryr_kubernetes/watcher.py | 15:04 |
| celebdor | where we get to line 148 without the path being in self._watching? | 15:05 |
| celebdor | (original code) | 15:05 |
| dulek | celebdor: I believe there's a race condition. | 15:08 |
| dulek | celebdor: I saw that in threaded case sometimes _graceful_watch_retry() gets executed on _stop_watch() and sometimes not. | 15:09 |
| celebdor | I believe there is for cleaning out code, but shouldn't the if be nested? | 15:09 |
| dulek | celebdor: So I'm clearing it in two places to make sure we're protected in both cases. | 15:09 |
| dulek | celebdor: And I've deleted the self.remove(path) line - because otherwise we're unable to restart the watcher using .start() | 15:10 |
| celebdor | ok | 15:13 |
| celebdor | let me see about that | 15:13 |
| celebdor | dulek: so remove path is never called anymore? | 15:14 |
| dulek | celebdor: Yup. I doubt it's needed really. See that self._resources is only used in start() and in conditionals in watch() which we always combine with self._running check. | 15:16 |
| dulek | celebdor: Hm, maybe this thing: https://github.com/openstack/kuryr-kubernetes/blob/6c3a73cc677ca499592de2b4fb5ce4544503c2fd/kuryr_kubernetes/watcher.py#L192-L193 … | 15:16 |
| * dulek is super confused about that code. | 15:17 | |
| celebdor | dulek: IIRC I put that there so that it would quite when there are no more paths | 15:17 |
| celebdor | being watched | 15:17 |
| celebdor | dulek: don't get me started | 15:17 |
| dulek | celebdor: :D | 15:17 |
| celebdor | I wish I could move it back to coroutines now that we do py3 | 15:17 |
| dulek | celebdor: So basically my issue here is that in threaded case we have multiple _watch() runs in different thread. | 15:18 |
| dulek | threads. | 15:18 |
| dulek | celebdor: And my understanding was that the code isn't really caring about that. | 15:18 |
| dulek | celebdor: I'm at SIG-K8s session and I have hard time doing context switching. Can we try to discuss it on the review? | 15:19 |
| dulek | celebdor: I'm totally open for any suggestions you may have. | 15:19 |
| celebdor | dulek: sure, sorry | 15:20 |
| celebdor | I'll try to think about it some more | 15:20 |
| dulek | celebdor: Just to recap - my main issue I wanted to fix is to make sure doing Watcher.stop; Watcher.start() doesn't end up with stuff removed from self._resources. | 15:20 |
| dulek | celebdor: Because that blocked HA leader switching. | 15:21 |
| celebdor | dulek: pay attention to the SIG | 15:21 |
| celebdor | we can sync later | 15:21 |
| celebdor | ;-) | 15:21 |
| dulek | celebdor: And also I feel like GreenThread.stop() doesn't always kill the GT immediately. | 15:21 |
| dulek | celebdor: Which causes race conditions. | 15:21 |
| celebdor | makes my head hurt | 15:22 |
| aperevalov | celebdor: no List-Id: The OpenStack General mailing list <openstack.lists.openstack.org> | 15:46 |
| celebdor | aperevalov: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev | 15:48 |
| aperevalov | celebdor: thanks!!! | 15:49 |
| celebdor | ;-) | 15:49 |
| *** hongbin_ has joined #openstack-kuryr | 15:55 | |
| *** hongbin has quit IRC | 15:56 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Refactor options definitons https://review.openstack.org/590415 | 15:59 |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-tempest-plugin master: A/P HA tests https://review.openstack.org/590416 | 15:59 |
| dulek | Hey folks, I'd like to highlight https://review.openstack.org/#/c/590415/ - it's a simple refactoring that gets into merge conflicts all the time. | 16:00 |
| dulek | Would be nice to get some attention there. | 16:01 |
| celebdor | dulek: I just +2ed it | 16:11 |
| * celebdor -> family time | 16:11 | |
| *** ccamposr has quit IRC | 16:14 | |
| *** garyloug_ has quit IRC | 17:10 | |
| dulek | celebdor: Quick question (sorry for the hour) - who's actually running Kuryr-Kubernetes (in production?)? I've joined Ops containers session because I see a question about it in the etherpad. | 17:21 |
| *** maysams has joined #openstack-kuryr | 17:31 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Refactor options definitons https://review.openstack.org/590415 | 17:39 |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-tempest-plugin master: A/P HA tests https://review.openstack.org/590416 | 17:39 |
| *** hongbin_ has quit IRC | 18:10 | |
| *** pcaruana has quit IRC | 20:08 | |
| *** celebdor has quit IRC | 20:53 | |
| *** rh-jelabarre has quit IRC | 21:12 | |
| *** hongbin has joined #openstack-kuryr | 21:53 | |
| *** takamatsu has quit IRC | 22:23 | |
| *** hongbin has quit IRC | 22:25 | |
| *** pmannidi has joined #openstack-kuryr | 23:33 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!