| *** celebdor1 has quit IRC | 00:16 | |
| *** pc_m has joined #openstack-kuryr | 00:20 | |
| *** yamamoto has joined #openstack-kuryr | 00:47 | |
| *** yamamoto has quit IRC | 00:51 | |
| *** hongbin has joined #openstack-kuryr | 01:05 | |
| openstackgerrit | Lv Jiawei proposed openstack/kuryr-kubernetes master: Devstack install fail if docker cgroup driver is systemd https://review.openstack.org/575319 | 01:24 |
|---|---|---|
| openstackgerrit | Lv Jiawei proposed openstack/kuryr-kubernetes master: Devstack install fail if docker cgroup driver is systemd https://review.openstack.org/575319 | 01:27 |
| openstackgerrit | Lv Jiawei proposed openstack/kuryr-kubernetes master: Devstack install fail if docker cgroup driver is systemd https://review.openstack.org/575319 | 01:29 |
| openstackgerrit | Peng Liu proposed openstack/kuryr-kubernetes master: Implement multi-vif driver https://review.openstack.org/575629 | 01:32 |
| *** yamamoto has joined #openstack-kuryr | 01:47 | |
| *** yamamoto has quit IRC | 01:53 | |
| *** yamamoto has joined #openstack-kuryr | 02:49 | |
| *** yamamoto has quit IRC | 02:55 | |
| *** hongbin has quit IRC | 03:10 | |
| *** rh-jelabarre has quit IRC | 03:22 | |
| *** rh-jelabarre has joined #openstack-kuryr | 03:26 | |
| *** yamamoto has joined #openstack-kuryr | 03:51 | |
| *** yamamoto has quit IRC | 03:57 | |
| *** yamamoto has joined #openstack-kuryr | 04:01 | |
| *** hongbin has joined #openstack-kuryr | 04:05 | |
| *** hongbin has quit IRC | 04:14 | |
| *** rh-jelabarre has quit IRC | 04:37 | |
| *** gcheresh has joined #openstack-kuryr | 05:12 | |
| *** ygupta has joined #openstack-kuryr | 05:16 | |
| *** ygupta has quit IRC | 05:18 | |
| *** janki has joined #openstack-kuryr | 05:47 | |
| openstackgerrit | Li Zhouzhou proposed openstack/kuryr-kubernetes master: Bug Fix: Kuryr installation as a Kubernetes network addon in kuryr-kubernetes. https://review.openstack.org/576367 | 06:01 |
| *** celebdor1 has joined #openstack-kuryr | 07:24 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Kuryr Kubernetes Port and PortPool CRD design reference document https://review.openstack.org/556181 | 07:30 |
| *** pcaruana has joined #openstack-kuryr | 07:50 | |
| *** celebdor1 has quit IRC | 08:11 | |
| *** pmannidi has quit IRC | 08:13 | |
| *** yboaron has joined #openstack-kuryr | 08:17 | |
| *** celebdor1 has joined #openstack-kuryr | 08:31 | |
| celebdor1 | ltomasbo: check the comment to https://review.openstack.org/#/c/556181/4 | 09:01 |
| ltomasbo | celebdor1, looking | 09:02 |
| ltomasbo | celebdor1, why selfLink is nicer? | 09:02 |
| *** lan has joined #openstack-kuryr | 09:06 | |
| ltomasbo | celebdor1, I'm happy to change it, but I though having the port id will be useful to perform some actions without having to perform an extra call | 09:07 |
| ltomasbo | celebdor1, but that will depend on the SDN, so maybe it is still better not to skip it and have the selfLink anyway | 09:07 |
| ltomasbo | dulek, as you were already using CRDs for the ports, what do you think? ^^ | 09:08 |
| celebdor1 | ltomasbo: which extra call? | 09:09 |
| celebdor1 | which ops? | 09:09 |
| ltomasbo | didn't check yet, but I expected having the port id there will help to do actions on the ports (re-apply sg, delete them) if need be | 09:12 |
| ltomasbo | but perhaps that can be obtained from the selfLink anyway | 09:14 |
| celebdor1 | ltomasbo: isn't the port_uuid which is also the name, part of the selflink? | 09:17 |
| celebdor1 | ltomasbo: from your example: /apis/openstack.org/v1/kuryrports/18f968a5-c420-4318-92d7-941eb5f9e60e | 09:18 |
| ltomasbo | yes yes, it is there. Ok, I'll change it! xD | 09:18 |
| dulek | celebdor1, ltomasbo: My ISP isn't doing a great job today, sorry if I get disconnected from IRC occasionally. | 09:19 |
| dulek | (and of course it's the match day when this happens :D) | 09:19 |
| dulek | ltomasbo: I've had issues with selfLinks not being created correctly. | 09:19 |
| dulek | ltomasbo: I mean by K8s API. | 09:19 |
| dulek | ltomasbo: IIRC selfLink put into the KuryrPort resource was simply malformed (missing namespace probably?). | 09:20 |
| celebdor1 | dulek: are people watching the match through internet rather than dvb-t2 ? | 09:23 |
| celebdor1 | cause that would be significant load | 09:23 |
| celebdor1 | dulek: did you file a bug on K8s for that | 09:23 |
| celebdor1 | it sounds like a big CRD bug | 09:23 |
| dulek | celebdor1: I'm forced to do it streaming, my home is in a valley that's cutting off signal. And I don't really need cable TV besides sport. ;) | 09:25 |
| dulek | celebdor1: I'm not sure if it wasn't listed on their CRDs issues list already. | 09:25 |
| dulek | Okay, I see that it is not listed there. I'll file a bug if I'll reproduce it. | 09:30 |
| ltomasbo | dulek, that depends on how you create the CRD | 09:30 |
| dulek | ltomasbo: ? | 09:31 |
| ltomasbo | dulek, what I experienced is that you can create them cluster wide or per namespace | 09:31 |
| dulek | ltomasbo: Yeah, sure. | 09:31 |
| celebdor1 | thanks dulek | 09:32 |
| celebdor1 | dulek: In czech the dvb-t2 signal is pretty spotty outside of cities. I suspect it is a lobby by the cable providers | 09:32 |
| dulek | ltomasbo: So I see that mine's scoped for Cluster. I'll need to remind myself what exactly was the issue. | 09:32 |
| ltomasbo | xD then it should not include namespace info (I guess) | 09:33 |
| dulek | ltomasbo: Yes, yes, that's why I don't remember the issue exactly. | 09:33 |
| dulek | celebdor1: Plus I guess DVB-T infrastructure isn't exactly cheap. Fortunately Polish TV has official online stream that worked fine through the qualifications phase, so fingers crossed. ;) | 09:34 |
| dulek | celebdor1: Worst case - I have a neighbor with cable TV. | 09:35 |
| celebdor1 | :-) | 09:43 |
| celebdor1 | dulek: I'll try to watch too | 09:43 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Kuryr Kubernetes Port and PortPool CRD design reference document https://review.openstack.org/556181 | 09:45 |
| ltomasbo | celebdor1, done ^^ | 09:46 |
| dulek | celebdor1, ltomasbo, dmellado: https://review.openstack.org/#/c/576071 seems okay, can you take a look? | 09:56 |
| dulek | I've tried debugging LBaaS v2 tests there since yesterday, but now I'm pretty sure we simply have LBaaS v2 gates broken at the moment (timing out). | 09:57 |
| ltomasbo | dulek, it looks good to me and we don't need the token anymore, right? | 10:01 |
| dulek | ltomasbo: I've removed it from the script. And we still need it for contacting K8s API from inside the daemon. | 10:03 |
| celebdor1 | dulek: what are you using pprint for? | 10:03 |
| celebdor1 | I see it imported but unused | 10:03 |
| ltomasbo | that was my next question | 10:03 |
| dulek | celebdor1: :D | 10:03 |
| dulek | celebdor1: Good eye, it's leftover as I've copypasted your script. | 10:04 |
| celebdor1 | dulek: also, you should return after the first print, shouldn't you? | 10:04 |
| celebdor1 | Why to look at the rest of the potentially long list of containers? | 10:04 |
| dulek | celebdor1: Yeah, that makes sense as well. | 10:04 |
| celebdor1 | dulek: otherwise it looks good | 10:07 |
| celebdor1 | personally I'd probably just write to /run/kuryr_cni_cid | 10:07 |
| celebdor1 | and then read from there | 10:07 |
| celebdor1 | but that is more complicated | 10:08 |
| celebdor1 | we can do it if we see the get from docker is slow | 10:08 |
| dulek | celebdor1: Hm, I'm not sure if I don't actually remember Docker API being slow with high number of containers. | 10:12 |
| dulek | I'll look it up. | 10:12 |
| celebdor1 | dulek: I guess you can just do | 10:21 |
| dulek | celebdor1: Yeah, I think I've already have it. Testing. | 10:21 |
| celebdor1 | ah, ok | 10:22 |
| celebdor1 | test with 1000 containers | 10:22 |
| *** lan1 has joined #openstack-kuryr | 10:26 | |
| *** lan has quit IRC | 10:29 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-tempest-plugin master: Extend namespace tempest plugin coverage https://review.openstack.org/576476 | 10:39 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: DNM: Testing namespace coverage https://review.openstack.org/576478 | 10:42 |
| *** yamamoto has quit IRC | 10:50 | |
| *** yamamoto has joined #openstack-kuryr | 10:58 | |
| *** yamamoto has quit IRC | 10:58 | |
| *** lan2 has joined #openstack-kuryr | 11:12 | |
| *** lan1 has quit IRC | 11:15 | |
| *** lan1 has joined #openstack-kuryr | 11:16 | |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: Deprecate running Kuryr-Kubernetes with lbaasv2 https://review.openstack.org/576486 | 11:19 |
| *** lan2 has quit IRC | 11:19 | |
| celebdor1 | dulek: please review https://review.openstack.org/#/c/576183/1 | 11:26 |
| dmellado | so, folks, deprecation day has come | 11:28 |
| dmellado | lbaasv2, beware | 11:28 |
| dmellado | xD | 11:28 |
| dmellado | I shall put you out of your misery by asking about african or european swallows! | 11:29 |
| celebdor1 | dmellado: I -1ed your patch | 11:29 |
| dmellado | why? | 11:29 |
| celebdor1 | wrong wording | 11:29 |
| dmellado | is not of your favorite color? | 11:29 |
| celebdor1 | dmellado: and because I have here a bag of -1s I have to use | 11:30 |
| dmellado | well, I'll reword it | 11:30 |
| dmellado | but it'll contain refernces to the holy grail | 11:30 |
| dulek | Crap, we had a meeting about network policy today? Google calendar haven't created a reminder for me again. :( | 11:30 |
| dmellado | https://www.youtube.com/watch?v=pWS8Mg-JWSg | 11:30 |
| dmellado | dulek: yep, we had that this morning | 11:30 |
| dmellado | but I'll be happy as long as you keep reviewing patches ;) | 11:31 |
| celebdor1 | dulek: it's fine, we assigned you to write all the unit tests since you didn't show up | 11:32 |
| dulek | dmellado: Okay, sorry about missing it. If that ever happens don't hesitate to ping me on IRC. | 11:32 |
| dmellado | dulek: no worries, it'll be enough with the unit test thing | 11:32 |
| dmellado | xD | 11:32 |
| * dmellado goes for lunch | 11:32 | |
| dulek | celebdor1: Am I allowed to outsource that? :D | 11:32 |
| dmellado | dulek: yep, to volkswagen provider | 11:33 |
| dmellado | xD | 11:33 |
| * dulek didn't understand the joke, but whatever. | 11:33 | |
| celebdor1 | dulek: yes, you can outsource it to dmellado | 11:34 |
| celebdor1 | :-) | 11:34 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: Really use the test handler https://review.openstack.org/575771 | 11:37 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: Really use the test handler https://review.openstack.org/575771 | 11:37 |
| celebdor1 | ltomasbo: it now points to both bugs ^^ | 11:38 |
| celebdor1 | dulek: you should review this one too | 11:38 |
| ltomasbo | thanks celebdor1! | 11:40 |
| ltomasbo | lgtm | 11:40 |
| celebdor1 | thanks to you! | 11:40 |
| celebdor1 | ltomasbo: I have to say that this bug mind fscked me quite a bit | 11:41 |
| celebdor1 | so to write the bug I had to go and remove the fix and see what happened with the fixed test | 11:41 |
| celebdor1 | :/ | 11:41 |
| ltomasbo | :/ | 11:41 |
| celebdor1 | it deserves more coverage | 11:41 |
| dulek | celebdor1: Sure, I'll do. | 11:42 |
| celebdor1 | thanks | 11:42 |
| celebdor1 | we should backport this this week | 11:42 |
| celebdor1 | it's a bit of a nasty bug | 11:42 |
| ltomasbo | celebdor1, I'm still not sure how you find that thing! it was really hidden... | 11:43 |
| dulek | celebdor1: Okay, I think I'll skip saving container ID to file now. I don't see major performance issues with docker API and saving is actually troublesome - I'd need to add another volume to the CNI DaemonSet as I would need to remove old CID file. | 11:43 |
| dulek | And I want to limit such changes if we want to backport it. | 11:43 |
| *** yamamoto has joined #openstack-kuryr | 11:48 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Fetch CONTAINERID from Docker API in kuryr-cni https://review.openstack.org/576071 | 11:50 |
| celebdor1 | ltomasbo: well, when I added my fix for actually quitting the whole damn controller/cnidaemon when you run out of watchers | 11:51 |
| celebdor1 | some test that should not fail started failing | 11:51 |
| celebdor1 | and after two hours or more of head scratching | 11:51 |
| celebdor1 | I noticed that the test was not even running the code it contained | 11:51 |
| celebdor1 | and proposed this fix | 11:51 |
| ltomasbo | celebdor1, yep, I can imagine that was a PITA to find | 11:52 |
| ltomasbo | for just a couple of line fix... | 11:52 |
| celebdor1 | ltomasbo: yup | 11:54 |
| celebdor1 | I am not too fond of the way the watcher works | 11:54 |
| celebdor1 | in python3 coroutines we had it much simpler | 11:54 |
| *** dougbtv has joined #openstack-kuryr | 11:56 | |
| *** dougbtv has quit IRC | 11:59 | |
| *** dougbtv has joined #openstack-kuryr | 12:00 | |
| *** salv-orlando has joined #openstack-kuryr | 12:06 | |
| *** rh-jelabarre has joined #openstack-kuryr | 12:08 | |
| *** salv-orlando has quit IRC | 12:10 | |
| *** yamamoto has quit IRC | 12:15 | |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: process to gracefully exit when last watcher exits https://review.openstack.org/575119 | 12:21 |
| celebdor1 | irenab: I know you started the patch, but I think you should still vote on https://review.openstack.org/#/c/556181/ | 12:23 |
| irenab | celebdor, I thought I did | 12:27 |
| *** jistr is now known as jistr|mtg | 12:28 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Fetch CONTAINERID from Docker API in kuryr-cni https://review.openstack.org/576071 | 12:28 |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Start kuryr-kubernetes after creating l7 router https://review.openstack.org/576183 | 12:32 |
| celebdor1 | thanks irenab | 12:36 |
| celebdor1 | dulek: you can merge it https://review.openstack.org/#/c/556181/ | 12:36 |
| celebdor1 | (if you have no comment to address, that is) | 12:36 |
| dulek | celebdor1: In my queue, just a sec, performance test first. | 12:37 |
| celebdor1 | dulek: you have your priority queue well set I see | 12:37 |
| celebdor1 | :-) | 12:37 |
| dulek | celebdor1: Naaah, it's allowing too many preemptive tasks. :P | 12:38 |
| celebdor1 | dulek: I guess the big preemption is gonna be at 17:00 CEST :P | 12:41 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-tempest-plugin master: Extend namespace tempest plugin coverage https://review.openstack.org/576476 | 12:43 |
| dulek | celebdor1: Something around 16:45, need to fire up the grill and move the couch as well. ;) | 12:43 |
| celebdor1 | ah nice | 12:45 |
| celebdor1 | I can smell the lamb in the grill downstairs already | 12:45 |
| celebdor1 | I'm gonna go eat a few ribs in 10min | 12:45 |
| *** yamamoto has joined #openstack-kuryr | 12:46 | |
| *** salv-orlando has joined #openstack-kuryr | 13:07 | |
| *** maysamacedos has joined #openstack-kuryr | 13:11 | |
| *** salv-orlando has quit IRC | 13:11 | |
| dulek | celebdor1: On my env I'm getting 0.2 to 0.3 second of delay with 1000 containers. | 13:14 |
| dulek | celebdor1: I guess that's acceptable. | 13:14 |
| openstackgerrit | Peng Liu proposed openstack/kuryr-kubernetes master: Implement multi-vif driver https://review.openstack.org/575629 | 13:32 |
| *** yamamoto has quit IRC | 13:34 | |
| *** atoth has joined #openstack-kuryr | 13:35 | |
| *** jistr|mtg is now known as jistr | 13:42 | |
| celebdor1 | dulek: I wonder if that would change if they had some of the labels | 13:50 |
| celebdor1 | and by how much | 13:50 |
| celebdor1 | (more part of the IF executed) | 13:50 |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Create network policy handler and driver https://review.openstack.org/573616 | 13:52 |
| *** hongbin has joined #openstack-kuryr | 13:53 | |
| *** rh-jelabarre has quit IRC | 13:53 | |
| dulek | celebdor1: I was only worrying about Docker API. Python's json module should be fine. | 13:55 |
| *** yamamoto has joined #openstack-kuryr | 13:55 | |
| *** yamamoto has quit IRC | 13:55 | |
| dulek | celebdor1: Hm, though it's not a C module, it seems… | 13:56 |
| *** yamamoto has joined #openstack-kuryr | 13:57 | |
| *** janki has quit IRC | 13:59 | |
| *** rh-jelabarre has joined #openstack-kuryr | 14:01 | |
| celebdor1 | dulek: it's worth it to check out | 14:04 |
| celebdor1 | brb | 14:05 |
| *** celebdor1 has quit IRC | 14:06 | |
| dulek | celebdor1: Now when I look at the code - how many labels do you expect? 1000? That's still a number that hashmap implementation of a Python dict can handle. | 14:06 |
| *** salv-orlando has joined #openstack-kuryr | 14:07 | |
| *** salv-orlando has quit IRC | 14:11 | |
| *** yamamoto has quit IRC | 14:16 | |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: Fix the experimental multinode jobs https://review.openstack.org/558762 | 14:29 |
| *** yamamoto has joined #openstack-kuryr | 14:30 | |
| *** celebdor1 has joined #openstack-kuryr | 14:31 | |
| *** yamamoto has quit IRC | 14:35 | |
| *** yamamoto has joined #openstack-kuryr | 14:46 | |
| *** yboaron has quit IRC | 14:50 | |
| *** yamamoto has quit IRC | 14:50 | |
| *** yamamoto has joined #openstack-kuryr | 15:01 | |
| *** yamamoto has quit IRC | 15:05 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-tempest-plugin master: Extend namespace tempest plugin coverage https://review.openstack.org/576476 | 15:10 |
| *** yamamoto has joined #openstack-kuryr | 15:16 | |
| *** yamamoto has quit IRC | 15:20 | |
| *** gcheresh has quit IRC | 15:23 | |
| *** janki has joined #openstack-kuryr | 15:25 | |
| *** pcaruana has quit IRC | 15:29 | |
| *** yamamoto has joined #openstack-kuryr | 15:31 | |
| *** yamamoto has quit IRC | 15:35 | |
| *** yamamoto has joined #openstack-kuryr | 15:46 | |
| *** lan2 has joined #openstack-kuryr | 15:46 | |
| *** lan1 has quit IRC | 15:49 | |
| *** yamamoto has quit IRC | 15:50 | |
| *** yamamoto has joined #openstack-kuryr | 16:15 | |
| *** yamamoto has quit IRC | 16:20 | |
| *** yamamoto has joined #openstack-kuryr | 16:30 | |
| *** yamamoto has quit IRC | 16:35 | |
| *** phuoc has quit IRC | 16:45 | |
| *** yamamoto has joined #openstack-kuryr | 16:45 | |
| *** yamamoto has quit IRC | 16:50 | |
| *** yamamoto has joined #openstack-kuryr | 17:01 | |
| *** pcaruana has joined #openstack-kuryr | 17:05 | |
| *** yamamoto has quit IRC | 17:05 | |
| *** yamamoto has joined #openstack-kuryr | 17:15 | |
| *** yamamoto has quit IRC | 17:20 | |
| *** yamamoto has joined #openstack-kuryr | 17:31 | |
| *** yamamoto has quit IRC | 17:35 | |
| *** yamamoto has joined #openstack-kuryr | 17:45 | |
| *** yamamoto has quit IRC | 17:50 | |
| *** hongbin has quit IRC | 17:56 | |
| *** yamamoto has joined #openstack-kuryr | 18:00 | |
| *** yamamoto has quit IRC | 18:05 | |
| *** yamamoto has joined #openstack-kuryr | 18:15 | |
| *** yamamoto has quit IRC | 18:20 | |
| *** yamamoto has joined #openstack-kuryr | 18:30 | |
| *** janki has quit IRC | 18:32 | |
| *** yamamoto has quit IRC | 18:35 | |
| *** yamamoto has joined #openstack-kuryr | 18:36 | |
| *** yamamoto has quit IRC | 18:36 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-tempest-plugin master: Extend namespace tempest plugin coverage https://review.openstack.org/576476 | 18:49 |
| *** atoth has quit IRC | 18:51 | |
| *** gcheresh has joined #openstack-kuryr | 19:08 | |
| *** yamamoto has joined #openstack-kuryr | 19:36 | |
| *** yamamoto has quit IRC | 19:42 | |
| *** aojea has joined #openstack-kuryr | 19:48 | |
| *** gcheresh has quit IRC | 20:07 | |
| *** pcaruana has quit IRC | 20:29 | |
| *** yamamoto has joined #openstack-kuryr | 20:38 | |
| *** yamamoto has quit IRC | 20:47 | |
| *** maysamacedos has quit IRC | 20:47 | |
| *** rh-jelabarre has quit IRC | 21:17 | |
| *** aojea has quit IRC | 21:41 | |
| *** yamamoto has joined #openstack-kuryr | 21:43 | |
| *** yamamoto has quit IRC | 21:48 | |
| *** kzaitsev_pi has quit IRC | 22:23 | |
| *** yamamoto has joined #openstack-kuryr | 22:44 | |
| *** yamamoto has quit IRC | 22:49 | |
| *** celebdor1 has quit IRC | 23:05 | |
| *** kzaitsev_pi has joined #openstack-kuryr | 23:12 | |
| *** yamamoto has joined #openstack-kuryr | 23:46 | |
| *** yamamoto has quit IRC | 23:51 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!