| *** yamamoto has quit IRC | 00:01 | |
| *** yamamoto has joined #openstack-kuryr | 00:08 | |
| *** yamamoto has quit IRC | 00:22 | |
| *** yamamoto has joined #openstack-kuryr | 00:22 | |
| *** yamamoto has quit IRC | 00:28 | |
| *** yamamoto has joined #openstack-kuryr | 00:29 | |
| *** yamamoto has quit IRC | 00:32 | |
| *** yamamoto has joined #openstack-kuryr | 00:35 | |
| *** yamamoto has quit IRC | 00:40 | |
| *** yamamoto has joined #openstack-kuryr | 00:41 | |
| *** yamamoto has quit IRC | 00:42 | |
| *** yamamoto has joined #openstack-kuryr | 00:43 | |
| *** yamamoto has quit IRC | 00:46 | |
| *** yamamoto has joined #openstack-kuryr | 00:51 | |
| *** pc_m has quit IRC | 00:53 | |
| *** pc_m has joined #openstack-kuryr | 00:54 | |
| *** yamamoto has quit IRC | 00:55 | |
| *** rh-jelabarre has quit IRC | 01:16 | |
| *** pmannidi has joined #openstack-kuryr | 01:20 | |
| *** maysamacedos has joined #openstack-kuryr | 02:36 | |
| *** maysamacedos has quit IRC | 02:45 | |
| *** threestrands has quit IRC | 04:02 | |
| *** yamamoto has joined #openstack-kuryr | 04:30 | |
| *** pcaruana has joined #openstack-kuryr | 04:44 | |
| *** pcaruana has quit IRC | 05:03 | |
| *** janki has joined #openstack-kuryr | 05:19 | |
| *** gcheresh_ has joined #openstack-kuryr | 06:29 | |
| *** gcheresh_ has quit IRC | 06:35 | |
| *** pcaruana has joined #openstack-kuryr | 06:46 | |
| *** natanbro has joined #openstack-kuryr | 06:51 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Add namespace subnet driver for namespace creation https://review.openstack.org/562247 | 07:51 |
|---|---|---|
| *** threestrands has joined #openstack-kuryr | 08:11 | |
| *** garyloug has joined #openstack-kuryr | 08:15 | |
| *** celebdor1 has joined #openstack-kuryr | 08:41 | |
| *** threestrands has quit IRC | 08:46 | |
| *** gcheresh_ has joined #openstack-kuryr | 08:49 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Make `body` arg conditional when deleting service https://review.openstack.org/569945 | 09:16 |
| dulek | ltomasbo: This should help with your stable/queens patch. ^ | 09:17 |
| dulek | celebdor1: BTW - I've found the cause of the disconnections when watching through loadbalancer. LBaaS v2 has 50 seconds timeout for connections hardcoded. | 09:36 |
| dulek | celebdor1: It's configurable in Octavia. In Neutron I can only provide different template for HA proxy configuration. | 09:36 |
| ltomasbo | thanks dulek, I'll add the dependency and check | 09:39 |
| celebdor1 | dulek: great | 09:40 |
| celebdor1 | dulek: is it configurable per LB or only for all LBs in Octavia? | 09:41 |
| dulek | celebdor1: Per listener if I understood correctly. | 09:41 |
| celebdor1 | perfect | 09:44 |
| celebdor1 | dulek: for lbaasv2 it will be the template used for everything, right? | 09:44 |
| dulek | celebdor1: Yes. I don't really like that, but there's no other way besides relying on connection being retried every 50 seconds. | 09:45 |
| celebdor1 | dulek: that's fine | 09:46 |
| celebdor1 | let's fix it in the gate template then | 09:46 |
| celebdor1 | and document it | 09:46 |
| dulek | celebdor1: Meanwhile - anything out of ordinary here: http://logs.openstack.org/95/569095/6/check/kuryr-kubernetes-tempest-lbaasv2/9b37cb9/controller/logs/iptables.txt.gz ? | 09:47 |
| dulek | celebdor1: Maybe some rule prevents traffic from host to LB. Seems unlikely to me, but still. | 09:48 |
| dulek | celebdor1: Oh, wait, maybe I have an idea about the cause. | 09:48 |
| dulek | celebdor1: Crap, RetryHandler only tries to reconnect for 180 seconds. And this is too little for LB to become active. | 09:54 |
| dulek | It works in containers because we spawn containers in test-config phase. | 09:54 |
| dulek | I think extending the timeout for RetryHandler in both controller and daemon watches is what we should do. | 09:55 |
| dulek | Actually I'd be tempted to just make them retry forever. | 09:55 |
| *** gcheresh_ has quit IRC | 09:55 | |
| celebdor1 | dulek: that was my first thought, iptables | 09:57 |
| celebdor1 | oh | 09:57 |
| celebdor1 | it works for containerized case, though, right? dulek ? | 09:58 |
| dulek | celebdor1: Yes, because in containerized DevStack deployments we're starting Kuryr containers in post-config phase - that is after LB is created and active. | 09:58 |
| celebdor1 | dulek: oh, I thought it would also survive because k8s/openshift would just restart the containers | 09:59 |
| dulek | celebdor1: It's possible this happens as well, lemme see. | 10:00 |
| dulek | celebdor1: Nah, this isn't happening: "restartCount: 0" on both pods. | 10:01 |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Use K8s API LB IP instead of HOST_IP in DevStack https://review.openstack.org/569095 | 10:11 |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Use K8s API LB IP instead of HOST_IP in DevStack https://review.openstack.org/569095 | 10:13 |
| celebdor1 | ok | 10:14 |
| celebdor1 | ltomasbo: please, check the review on the namespace subnet handler | 10:41 |
| ltomasbo | celebdor1, ohh, I was about to upload the unittest fixes... | 10:41 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Add namespace subnet driver for namespace creation https://review.openstack.org/562247 | 10:42 |
| ltomasbo | going to check... | 10:42 |
| *** janki has quit IRC | 10:44 | |
| celebdor1 | thanks ltomasbo | 10:44 |
| ltomasbo | celebdor1, umm, I'm not sure about that, it is handler dependant but it is definitely a neutron config | 10:45 |
| ltomasbo | easy to move but not sure if more intuitive or not | 10:45 |
| *** pc_m has quit IRC | 10:46 | |
| *** pc_m has joined #openstack-kuryr | 10:48 | |
| ltomasbo | celebdor1, ok, I'll move it -- after a second thought, it is better to move it, and then if other drivers need the same info, move it back to the neutron defaults | 10:50 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-tempest-plugin master: Add tempest coverage for namespace creation https://review.openstack.org/567179 | 10:52 |
| *** pcaruana has quit IRC | 10:53 | |
| celebdor1 | ltomasbo: agreed, thanks | 10:58 |
| ltomasbo | celebdor1, and I should move pod router and pod subnet_ pool, right? | 10:59 |
| celebdor1 | ltomasbo: that would need a major version change :'( | 10:59 |
| celebdor1 | wouldn't it? | 10:59 |
| *** rh-jelabarre has joined #openstack-kuryr | 11:01 | |
| ltomasbo | don't think so | 11:04 |
| ltomasbo | I'll go for lunch and have the patch ready after that! | 11:05 |
| ltomasbo | celebdor1, ^^ | 11:05 |
| *** pcaruana has joined #openstack-kuryr | 11:09 | |
| *** janki has joined #openstack-kuryr | 11:17 | |
| celebdor1 | ltomasbo: well, I guess the oslo deprecated config facility would help avoiding the major version | 11:30 |
| *** natanbro has quit IRC | 11:32 | |
| *** garyloug_ has joined #openstack-kuryr | 11:49 | |
| *** garyloug has quit IRC | 11:49 | |
| *** snapiri has joined #openstack-kuryr | 12:17 | |
| *** janki has quit IRC | 12:29 | |
| *** janki has joined #openstack-kuryr | 12:30 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Add namespace subnet driver for namespace creation https://review.openstack.org/562247 | 12:43 |
| *** pcaruana has quit IRC | 12:54 | |
| ltomasbo | celebdor1, ^^ | 12:55 |
| celebdor1 | thanks | 12:57 |
| *** gcheresh_ has joined #openstack-kuryr | 13:05 | |
| ltomasbo | celebdor1, I just tested and it seems to work, going to rebase the other 2 and add the unittest | 13:06 |
| ltomasbo | celebdor1, also, this got the other +2 https://review.openstack.org/#/c/569431/ | 13:06 |
| *** gcheresh_ has quit IRC | 13:09 | |
| *** pcaruana has joined #openstack-kuryr | 13:13 | |
| celebdor1 | ltomasbo: done | 13:34 |
| celebdor1 | ltomasbo: you tested it on bm I suppose, right? | 13:34 |
| ltomasbo | celebdor1, yep, but I can test them all together later on nested | 13:41 |
| celebdor1 | ltomasbo: I don't think the behavior will change | 13:44 |
| ltomasbo | celebdor1, nop, it should not | 13:45 |
| ltomasbo | steps are the same for both cases | 13:45 |
| celebdor1 | fine then | 13:50 |
| celebdor1 | ltomasbo: why does https://review.openstack.org/#/c/567179/ depend on the implementation? | 13:51 |
| celebdor1 | Shouldn't it be the other way around? | 13:51 |
| celebdor1 | and then have a zuul tests that enables the namespace driver tests | 13:52 |
| ltomasbo | celebdor1, I don't think so, the tempest test should depend on the implementation, right? and then I'll add a gate (experimental) on kuryr-kubernetes that will trigger that tempest test, right? | 13:53 |
| openstackgerrit | Merged openstack/kuryr-tempest-plugin master: Make `body` arg conditional when deleting service https://review.openstack.org/569945 | 13:54 |
| celebdor1 | ltomasbo: I'm always confused about these things | 13:58 |
| celebdor1 | for me the logical thing is for code to depends on its tests | 13:58 |
| ltomasbo | me too! | 13:58 |
| celebdor1 | so it only gets merged if the tests pass | 13:58 |
| ltomasbo | ok, but as the tests are in a different repo, then you need to ensure the tempest test are running with the new functionality | 13:59 |
| ltomasbo | otherwise it will not create the namespaces resources | 13:59 |
| ltomasbo | I think that was the case when tempest was in-tree, instead of in a different repo | 14:00 |
| celebdor1 | ltomasbo: it would be good to get dmellado to confirm | 14:03 |
| celebdor1 | before we botch it up | 14:03 |
| ltomasbo | sure! | 14:04 |
| ltomasbo | no hurry | 14:05 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Namespace deletion functionality for namespace_subnet driver https://review.openstack.org/562249 | 14:18 |
| *** lihi has quit IRC | 14:28 | |
| *** irenab has quit IRC | 14:28 | |
| *** oanson has quit IRC | 14:28 | |
| *** snapiri has quit IRC | 14:28 | |
| *** oanson has joined #openstack-kuryr | 14:50 | |
| *** irenab has joined #openstack-kuryr | 14:50 | |
| *** lihi has joined #openstack-kuryr | 14:51 | |
| *** garyloug_ has quit IRC | 14:51 | |
| *** pcaruana has quit IRC | 14:58 | |
| *** yamamoto has quit IRC | 15:39 | |
| *** maysamacedos has joined #openstack-kuryr | 15:49 | |
| *** maysamacedos has quit IRC | 15:51 | |
| *** maysamacedos has joined #openstack-kuryr | 15:53 | |
| *** gcheresh_ has joined #openstack-kuryr | 16:18 | |
| *** gcheresh_ has quit IRC | 16:24 | |
| *** maysamacedos has quit IRC | 16:37 | |
| dulek | celebdor1: Sorry it took me so long, I was writing patch to python-octaviaclient (guess what, timeouts are not yet exposed there :D). | 16:43 |
| dulek | celebdor1: Answer is inline. Do you think I should move run_kuryr_kubernetes to test-config phase? | 16:44 |
| *** maysamacedos has joined #openstack-kuryr | 16:58 | |
| *** janki has quit IRC | 17:40 | |
| *** maysamacedos has quit IRC | 17:48 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Use K8s API LB IP instead of HOST_IP in DevStack https://review.openstack.org/569095 | 17:59 |
| *** yamamoto has joined #openstack-kuryr | 18:20 | |
| *** yamamoto has quit IRC | 18:21 | |
| *** maysamacedos has joined #openstack-kuryr | 18:50 | |
| *** maysamacedos has quit IRC | 18:57 | |
| *** maysamacedos has joined #openstack-kuryr | 18:58 | |
| *** maysamacedos has quit IRC | 18:59 | |
| *** gcheresh_ has joined #openstack-kuryr | 19:32 | |
| *** gcheresh_ has quit IRC | 20:07 | |
| *** kzaitsev1pi has quit IRC | 20:37 | |
| *** kzaitsev_pi has joined #openstack-kuryr | 20:40 | |
| *** maysamacedos has joined #openstack-kuryr | 21:21 | |
| *** maysamacedos has quit IRC | 21:27 | |
| *** threestrands has joined #openstack-kuryr | 22:39 | |
| *** celebdor1 has quit IRC | 22:43 | |
| *** kzaitsev_pi has quit IRC | 22:49 | |
| *** kzaitsev_pi has joined #openstack-kuryr | 22:49 | |
| *** yamamoto has joined #openstack-kuryr | 22:59 | |
| *** yamamoto has quit IRC | 23:10 | |
| *** yamamoto has joined #openstack-kuryr | 23:38 | |
| *** yamamoto has quit IRC | 23:40 | |
| *** isssp has quit IRC | 23:42 | |
| *** isssp has joined #openstack-kuryr | 23:42 | |
| *** yamamoto has joined #openstack-kuryr | 23:42 | |
| *** yamamoto has quit IRC | 23:47 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!