*** yamamoto has joined #openstack-kuryr | 00:07 | |
*** c00281451 has joined #openstack-kuryr | 01:41 | |
*** c00281451 is now known as zengchen | 01:41 | |
*** hongbin has joined #openstack-kuryr | 01:46 | |
*** salv-orl_ has joined #openstack-kuryr | 02:04 | |
*** salv-orlando has quit IRC | 02:07 | |
openstackgerrit | Eunsoo Park proposed openstack/kuryr-kubernetes master: Watcher restarts watching resources in failure https://review.openstack.org/529832 | 02:22 |
---|---|---|
openstackgerrit | Eunsoo Park proposed openstack/kuryr-kubernetes master: Make Neutron LBaaS Activation Timeout configurable https://review.openstack.org/549945 | 02:24 |
openstackgerrit | Eunsoo Park proposed openstack/kuryr-kubernetes master: Make Neutron LBaaS Activation Timeout configurable https://review.openstack.org/549945 | 02:26 |
*** wangbo has joined #openstack-kuryr | 02:45 | |
*** janki has joined #openstack-kuryr | 03:34 | |
*** maysamacedos has quit IRC | 03:43 | |
*** hongbin has quit IRC | 03:52 | |
*** janonymous has joined #openstack-kuryr | 04:08 | |
*** wangbo has quit IRC | 04:38 | |
*** wangbo has joined #openstack-kuryr | 05:15 | |
*** wangbo has quit IRC | 06:02 | |
*** wangbo has joined #openstack-kuryr | 06:03 | |
*** yboaron has joined #openstack-kuryr | 06:04 | |
irenab | kaliya, regarding https://bugs.launchpad.net/kuryr-kubernetes/+bug/1708173, can be useful | 06:13 |
openstack | Launchpad bug 1708173 in kuryr-kubernetes "Update Vagrant to fit Octavia used for LBaaS requirements" [Low,Triaged] | 06:13 |
*** wangbo has left #openstack-kuryr | 06:39 | |
*** premsankar has quit IRC | 06:51 | |
*** gcheresh_ has joined #openstack-kuryr | 07:05 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr master: Updated from global requirements https://review.openstack.org/549537 | 07:05 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-kubernetes master: Updated from global requirements https://review.openstack.org/549253 | 07:06 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-libnetwork master: Updated from global requirements https://review.openstack.org/549538 | 07:06 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-tempest-plugin master: Updated from global requirements https://review.openstack.org/551497 | 07:06 |
*** pcaruana has joined #openstack-kuryr | 07:48 | |
*** pcaruana has quit IRC | 07:50 | |
*** pcaruana has joined #openstack-kuryr | 07:50 | |
*** pcaruana is now known as pcaruana|wfh| | 07:53 | |
*** yboaron has quit IRC | 08:17 | |
openstackgerrit | Genadi Chereshnya proposed openstack/kuryr-tempest-plugin master: Adding pools and containerized jobs to .zuul https://review.openstack.org/552512 | 08:53 |
*** ispp is now known as burned | 09:04 | |
*** celebdor has joined #openstack-kuryr | 09:22 | |
*** janki has quit IRC | 09:28 | |
dulek | Anyone have a clue why gates fail with POST_FAILURE? Logs look pretty much okay. | 09:33 |
dulek | Oh wait, I see, post fails when copying the logs - I guess. | 09:33 |
*** yboaron has joined #openstack-kuryr | 09:35 | |
dulek | Okay, looks like infra folks are still working on fixing this, meanwhile rechecking the patches doesn't make a lot of sense. | 09:35 |
celebdor | thanks for the info dulek | 09:44 |
*** garyloug has joined #openstack-kuryr | 09:44 | |
dulek | "yes, see my discussion with tobiash earlier. reason not clear yet, we will probably need to wait for some US people to look into it" :( | 10:01 |
dmellado | dulek: post-failure seems to be due to infra side issues | 10:22 |
dmellado | seems that we can't get to keep the infra working for more than 1 day | 10:22 |
dmellado | damn | 10:22 |
dulek | dmellado: Yup, it's affecting all DevStack jobs. | 10:22 |
openstackgerrit | Yossi Boaron proposed openstack/kuryr-kubernetes master: Services: Add support for K8S service's port edit use case https://review.openstack.org/551773 | 10:35 |
*** garyloug_ has joined #openstack-kuryr | 10:42 | |
*** garyloug_ has quit IRC | 10:42 | |
openstackgerrit | Yossi Boaron proposed openstack/kuryr-kubernetes master: Services: Add support for K8S service's port edit use case https://review.openstack.org/551773 | 10:42 |
openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: Split .zuul.yaml into smaller files https://review.openstack.org/548309 | 10:48 |
*** yamamoto has quit IRC | 11:04 | |
*** snapiri- has joined #openstack-kuryr | 11:18 | |
*** caowei_ has joined #openstack-kuryr | 11:19 | |
dmellado | dulek: https://review.openstack.org/#/c/553329/ | 11:21 |
*** pcaruana|wfh| has quit IRC | 11:27 | |
*** caowei has quit IRC | 11:27 | |
*** snapiri has quit IRC | 11:27 | |
*** kzaitsev_pi has quit IRC | 11:27 | |
*** caowei_ is now known as caowei | 11:27 | |
*** pcaruana|wfh| has joined #openstack-kuryr | 11:28 | |
*** kzaitsev1pi has joined #openstack-kuryr | 11:28 | |
*** alraddarla has joined #openstack-kuryr | 11:32 | |
*** dmellado has quit IRC | 11:50 | |
*** dmellado has joined #openstack-kuryr | 11:58 | |
*** yamamoto has joined #openstack-kuryr | 12:04 | |
*** dmellado has quit IRC | 12:05 | |
*** alraddarla has quit IRC | 12:08 | |
*** yamamoto has quit IRC | 12:09 | |
*** pcaruana|wfh| is now known as pcaruana | 12:14 | |
*** maysamacedos has joined #openstack-kuryr | 12:15 | |
*** yamamoto has joined #openstack-kuryr | 12:16 | |
*** maysamacedos has quit IRC | 12:17 | |
*** dmellado has joined #openstack-kuryr | 12:23 | |
*** alraddarla_ has joined #openstack-kuryr | 12:24 | |
*** snapiri- is now known as snapiri | 12:27 | |
*** snapiri has quit IRC | 12:27 | |
*** snapiri has joined #openstack-kuryr | 12:28 | |
snapiri | dmellado: Isn't there a way around it? Some definition one can add to the job? | 12:30 |
*** dmellado has quit IRC | 12:30 | |
*** snapiri has quit IRC | 12:41 | |
*** snapiri has joined #openstack-kuryr | 12:41 | |
*** garyloug has quit IRC | 12:46 | |
dulek | snapiri: Not really, it's a change in ansible: https://review.openstack.org/#/c/553329/1/roles/devstack-project-conf/tasks/main.yaml | 12:49 |
snapiri | dulek: thanks. So we will have to wait for the infra guys to merge their rollback :) | 12:51 |
*** janonymous has quit IRC | 13:02 | |
*** atoth has joined #openstack-kuryr | 13:06 | |
*** maysamacedos has joined #openstack-kuryr | 13:07 | |
*** yamamoto has quit IRC | 13:18 | |
*** yamamoto has joined #openstack-kuryr | 13:19 | |
*** dmellado has joined #openstack-kuryr | 13:19 | |
celebdor | yboaron: regarding https://review.openstack.org/#/c/547822/ | 13:20 |
*** dmellado has quit IRC | 13:20 | |
celebdor | I'm not a fan of factories and this approach seems a bit restrictive | 13:20 |
*** garyloug has joined #openstack-kuryr | 13:20 | |
celebdor | yboaron: dulek: I'm thinking about merging it to get the functionality in and then changing it so that it works similarly to neutron service plugins | 13:21 |
yboaron | celebdor, so what do you suggest? alternative ? | 13:21 |
celebdor | based on entrypoints | 13:21 |
irenab | celebdor, merging what? | 13:21 |
celebdor | merging yboaron's patch | 13:21 |
celebdor | that introduces the factory | 13:21 |
yboaron | irenab, https://review.openstack.org/#/c/547822/ | 13:21 |
celebdor | and then swapping that part out | 13:21 |
dulek | celebdor: You mean using stevedore. I'm fine with that. | 13:22 |
celebdor | so that you can just put in config comma separated entrypoints | 13:22 |
yboaron | irenab, add support for Openshift | 13:22 |
dulek | celebdor: That would make it extendable. | 13:22 |
celebdor | (which can be declared by other python libraries) | 13:22 |
celebdor | dulek: that's exactly what I'm saying | 13:22 |
irenab | yboaron, taking a look | 13:22 |
celebdor | the only missing piece in my plan is to have pre-defined collections | 13:22 |
celebdor | of entrypoints | 13:22 |
celebdor | but I'm sure it can be done somehow | 13:22 |
irenab | yboaron, celebdor so this patch is just the framework? | 13:23 |
*** yamamoto has quit IRC | 13:23 | |
yboaron | yes | 13:23 |
irenab | This should be properly defined since itg defines a contract | 13:24 |
irenab | celebdor, let me take a look. I prefer not to merge fundamentals when there is an agreement to change it | 13:25 |
celebdor | irenab: so you would advocate making a devref and going for the approach I'm saying directly? | 13:25 |
irenab | give me a sec, missed this patch before | 13:25 |
celebdor | ok | 13:25 |
irenab | but as a general say, I would probably advocate for devref and what you say (you know me pretty well :-)) | 13:26 |
celebdor | alrighty then | 13:27 |
celebdor | yboaron: let's sync on this after the meeting | 13:28 |
yboaron | celebdor, OK | 13:29 |
*** snapiri has quit IRC | 13:29 | |
irenab | celebdor, this looks a bit similar to the direction we discussed about a year ago with Ilya, the profiles | 13:31 |
*** snapiri has joined #openstack-kuryr | 13:32 | |
irenab | then it was more focused on 'south' side (the drivers), but probably should be also for the 'north' (the watchers) | 13:33 |
*** janki has joined #openstack-kuryr | 13:52 | |
*** natanbro has joined #openstack-kuryr | 14:00 | |
*** neiljerram has joined #openstack-kuryr | 14:05 | |
*** kiennt26 has joined #openstack-kuryr | 14:07 | |
*** natanbro has quit IRC | 14:11 | |
*** yamamoto has joined #openstack-kuryr | 14:20 | |
*** yamamoto has quit IRC | 14:25 | |
celebdor | irenab: I think since yossi's patch is not changing the fact that they are hardcoded | 14:26 |
celebdor | we can merge it, and then I'll make a patch with the devref and the stevedore approach | 14:27 |
irenab | celebdor, I am not sure we need platform abstraction | 14:27 |
irenab | in the code | 14:28 |
irenab | this is just template/profile to deploy which mixture of watchers and drivers | 14:28 |
irenab | Do you think platform is needed? | 14:29 |
celebdor | you know me irenab | 14:35 |
celebdor | irenab: personally I'd just have done a if get_platform == OSP: register_watchers(osp_watchers) | 14:36 |
celebdor | I despise factories and inheritance | 14:36 |
irenab | celebdor, I just not sure Platform is something that should exist in kuryr code, it is more of a deployment environment that kuryr is used in | 14:38 |
celebdor | irenab: I agree with you | 14:38 |
celebdor | that's why I propose to have the handlers config variable | 14:39 |
celebdor | that in deployment time you'll just put your handlers comma separated there | 14:39 |
irenab | yes, this is more aligned with the direction, agreed | 14:39 |
celebdor | I would only merge it to allow the osp and k8s handlers in the meantime | 14:39 |
celebdor | but I would scrap it | 14:39 |
celebdor | after we have it deployment config ready | 14:40 |
celebdor | maybe I'd add default k8s and osp entrypoints that encompass the other ones that they need | 14:40 |
irenab | Maybe just removing the Factory and follow more lightweight approach you suggest? | 14:40 |
celebdor | but that's about it | 14:40 |
*** gcheresh has joined #openstack-kuryr | 14:42 | |
irenab | and going forward, we really need to do some steps towards the profiles, since number of options start to grow | 14:42 |
irenab | celebdor, entrypoints for watchers? | 14:43 |
*** gcheresh_ has quit IRC | 14:43 | |
celebdor | yes, entrypoints for watchers | 14:44 |
celebdor | you'll have a config like: | 14:44 |
celebdor | pipeline_watchers = vif,service,ingress | 14:45 |
celebdor | for example | 14:45 |
*** dmellado_ has joined #openstack-kuryr | 14:46 | |
*** phuoc_ has joined #openstack-kuryr | 14:47 | |
irenab | celebdor, I agree. I prefer kuryr code to be less 'smart' with regards to what environment it lives in | 14:49 |
irenab | so the more can be provided during the deployment time, the better | 14:49 |
*** phuoc has quit IRC | 14:51 | |
*** phuoc has joined #openstack-kuryr | 14:52 | |
*** phuoc_ has quit IRC | 14:52 | |
*** dmellado_ is now known as dmellado | 14:54 | |
dmellado | +1 | 14:56 |
dmellado | maybe we can just add that as an option in kuryr.conf and have the user handle it | 14:56 |
*** gcheresh has quit IRC | 14:58 | |
*** gcheresh has joined #openstack-kuryr | 14:58 | |
irenab | dmellado, yes, this is the idea. The cluster manager will provide the config based on deployment environment | 15:01 |
irenab | celebdor, dmellado need to go now, ping me if some follow-up is required | 15:02 |
irenab | dmellado, enjoy your time in BCN | 15:03 |
*** gcheresh has quit IRC | 15:09 | |
celebdor | irenab: he already left | 15:09 |
celebdor | to the center to shop and then to the train | 15:09 |
*** yamamoto has joined #openstack-kuryr | 15:21 | |
*** maysamacedos has quit IRC | 15:27 | |
*** yamamoto has quit IRC | 15:27 | |
yboaron | celebdor, another point for having configurable watchers is integration with Istio, where only the pod and NetPolicy will be required in that case | 15:32 |
*** garyloug_ has joined #openstack-kuryr | 15:36 | |
*** maysamacedos has joined #openstack-kuryr | 15:38 | |
*** garyloug has quit IRC | 15:38 | |
*** kiennt26 has quit IRC | 15:44 | |
celebdor | alright. I'll draft something up then | 15:45 |
*** maysamacedos has quit IRC | 15:56 | |
openstackgerrit | MichaĆ Dulko proposed openstack/kuryr-kubernetes master: WiP: Preview of CNI daemon side VIF choice https://review.openstack.org/527243 | 15:58 |
dulek | celebdor: ^ This should be rebased and was tested. I'd love to get some feedback on how this works on the lowest level - CNI binding and stuff. Please note that only bridge drivers are currently done. | 15:59 |
dulek | celebdor: And I haven't incorporated the switch from marking VIFs reservation in pods instead of CRDs yet. | 16:00 |
* dulek needs to go now. | 16:02 | |
celebdor | dulek: will check tomorrow | 16:05 |
celebdor | thanks | 16:05 |
dulek | celebdor: I'd be grateful. Meanwhile I'll take a Friday break from this and will work on other stuff from the PTG. ;) | 16:06 |
celebdor | dulek: like? | 16:07 |
yboaron | celebdor, we plan to allow configuring not just the resources to watch, but also the handler functions right ? | 16:17 |
celebdor | no, jus tthe handlers | 16:18 |
celebdor | just | 16:18 |
celebdor | the handlers decide what they watch | 16:19 |
yboaron | OK, right! | 16:20 |
*** yamamoto has joined #openstack-kuryr | 16:23 | |
*** hongbin has joined #openstack-kuryr | 16:25 | |
*** yamamoto has quit IRC | 16:29 | |
*** garyloug_ has quit IRC | 16:35 | |
*** yboaron has quit IRC | 16:38 | |
*** janki has quit IRC | 16:46 | |
*** garyloug_ has joined #openstack-kuryr | 16:52 | |
*** celebdor has quit IRC | 16:56 | |
*** yamamoto has joined #openstack-kuryr | 17:25 | |
*** yamamoto has quit IRC | 17:30 | |
*** maysamacedos has joined #openstack-kuryr | 17:32 | |
*** garyloug_ has quit IRC | 17:42 | |
*** yamamoto has joined #openstack-kuryr | 18:27 | |
*** yamamoto has quit IRC | 18:33 | |
*** hongbin has quit IRC | 19:25 | |
*** hongbin has joined #openstack-kuryr | 19:26 | |
*** yamamoto has joined #openstack-kuryr | 19:29 | |
*** yamamoto has quit IRC | 19:33 | |
*** yamamoto has joined #openstack-kuryr | 20:29 | |
*** yamamoto has quit IRC | 20:35 | |
*** gcheresh has joined #openstack-kuryr | 20:53 | |
*** s1061123 has quit IRC | 20:58 | |
*** s1061123 has joined #openstack-kuryr | 21:00 | |
*** mburrows has quit IRC | 21:21 | |
*** gcheresh has quit IRC | 21:26 | |
*** yamamoto has joined #openstack-kuryr | 21:31 | |
*** yamamoto has quit IRC | 21:36 | |
*** celebdor has joined #openstack-kuryr | 21:59 | |
*** celebdor has quit IRC | 22:19 | |
*** yamamoto has joined #openstack-kuryr | 22:32 | |
*** yamamoto has quit IRC | 22:38 | |
*** hongbin has quit IRC | 22:50 | |
*** salv-orl_ has quit IRC | 22:53 | |
*** salv-orlando has joined #openstack-kuryr | 22:54 | |
*** salv-orlando has quit IRC | 22:59 | |
*** celebdor has joined #openstack-kuryr | 23:04 | |
*** yamamoto has joined #openstack-kuryr | 23:27 | |
*** salv-orlando has joined #openstack-kuryr | 23:54 | |
*** salv-orlando has quit IRC | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!