| *** rh-jelabarre has quit IRC | 01:08 | |
| *** hongbin_ has joined #openstack-kuryr | 01:26 | |
| *** gkadam has joined #openstack-kuryr | 03:21 | |
| *** hongbin_ has quit IRC | 03:46 | |
| *** gkadam has quit IRC | 04:10 | |
| *** gkadam has joined #openstack-kuryr | 04:10 | |
| *** yboaron_ has joined #openstack-kuryr | 05:01 | |
| *** shachar is now known as snapiri | 05:16 | |
| *** itzikb has joined #openstack-kuryr | 05:23 | |
| *** openstackgerrit has joined #openstack-kuryr | 06:06 | |
| openstackgerrit | Itzik Brown proposed openstack/kuryr-tempest-plugin master: Add test_pod_pod_ping_different_host https://review.openstack.org/599314 | 06:06 |
|---|---|---|
| ltomasbo | yboaron_, regarding the error you mention, I'm not sure, but I know I saw this on the default local.conf at ovn: | 06:17 |
| ltomasbo | yboaron_, Specifying Q_AGENT is needed to allow devstack | 06:18 |
| ltomasbo | # to run various "ip link set" and "ovs-vsctl" commands for the provider | 06:18 |
| ltomasbo | # network setup. | 06:18 |
| ltomasbo | #Q_AGENT=openvswitch | 06:18 |
| ltomasbo | not sure if it could be related... | 06:18 |
| yboaron_ | ltomasbo, 10x for the info, so the devstack with ovn local.conf , also fails , right? | 06:19 |
| yboaron_ | ltomasbo, I just about trying fresh devstack with the ovn local.conf | 06:19 |
| ltomasbo | yboaron_, nop, it didn't fail for me | 06:19 |
| ltomasbo | yboaron_, only problem was with configuring the octavia healthmanager port | 06:20 |
| yboaron_ | ltomasbo, I hacked the octavia plugin.sh, to re-create the default route .. and devstack deployment succeeded | 06:20 |
| yboaron_ | ltomasbo, but kuryr-cni is failing | 06:21 |
| ltomasbo | yboaron_, I did the same (ip route replace via ...) | 06:21 |
| ltomasbo | yboaron_, let me see how it ended up for me... | 06:21 |
| ltomasbo | kubectl get pods --all-namespaces | 06:22 |
| ltomasbo | NAMESPACE NAME READY STATUS RESTARTS AGE | 06:22 |
| ltomasbo | kube-system kuryr-cni-ds-qhqjj 1/1 Running 0 14h | 06:22 |
| ltomasbo | kube-system kuryr-controller-7db7c944fb-4vhgk 1/1 Running 0 14h | 06:22 |
| ltomasbo | seems to be ok... | 06:22 |
| yboaron_ | ltomasbo, can u please try to deploy a pod ? | 06:23 |
| *** pcaruana has joined #openstack-kuryr | 06:23 | |
| ltomasbo | test demo-5995548848-psz2j 1/1 Running 0 52s 10.0.0.68 | 06:24 |
| ltomasbo | curl 10.0.0.68:8080 | 06:24 |
| ltomasbo | demo-5995548848-psz2j: HELLO! I AM ALIVE!!! | 06:24 |
| ltomasbo | seems to work... | 06:24 |
| openstackgerrit | Yash Gupta proposed openstack/kuryr-kubernetes master: nested_vif - Get trunk port IP from node label https://review.openstack.org/599930 | 06:24 |
| ltomasbo | yboaron_, though I'm trying without the ovn-octavia plugin... | 06:26 |
| yboaron_ | ltomasbo, I deployed kuryr non-containerized | 06:27 |
| ltomasbo | yboaron_, oh, ok. I always use daemon and containerized | 06:27 |
| ltomasbo | perhaps it is not plugged on the br-int in the proper way then... | 06:28 |
| yboaron_ | ltomasbo, I used daemon and non-containerized, I'll try the containerized .. | 06:29 |
| ltomasbo | yboaron_, ok! | 06:31 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Ensure ports pool are not duplicated with kuryr-controller HA https://review.openstack.org/593679 | 06:31 |
| yboaron_ | ltomasbo, 10x! , I"ll try to deploy the containerized, BTW this is the local.conf I used : https://pastebin.com/cFjGPq6H | 06:38 |
| ltomasbo | let me see the differences with mine | 06:39 |
| ltomasbo | yboaron_, http://paste.openstack.org/show/729474/ | 06:40 |
| ltomasbo | yboaron_, first time I see: enable_service disable_service | 06:41 |
| ltomasbo | yboaron_, seems to be pretty similar, with a couple of differences | 06:42 |
| dmellado | morning folks | 06:42 |
| ltomasbo | yboaron_, I compiled ovs from source: OVS_HOST_PATH=/usr/local/var/run/openvswitch | 06:42 |
| ltomasbo | yboaron_, and add some configuration for the providernet | 06:43 |
| ltomasbo | OVN_L3_CREATE_PUBLIC_NETWORK=True | 06:43 |
| ltomasbo | PHYSICAL_NETWORK=public | 06:43 |
| ltomasbo | Q_USE_PROVIDERNET_FOR_PUBLIC=True | 06:43 |
| ltomasbo | FIXED_RANGE=20.0.0.0/20 | 06:43 |
| ltomasbo | FLOATING_RANGE=172.24.5.0/24 | 06:43 |
| ltomasbo | NETWORK_GATEWAY=20.0.0.1 | 06:43 |
| ltomasbo | PUBLIC_NETWORK_GATEWAY=172.24.5.1 | 06:43 |
| ltomasbo | (and did not enable tempest) | 06:43 |
| ltomasbo | and by the way, the fixed_range and network gateway are just wrong (don't do anything...) | 06:44 |
| yboaron_ | ltomasbo, OK!, thanks for the detailed info, I"ll customize my local.conf and retry | 06:44 |
| ltomasbo | yboaron_, hope it helps, not sure what could be. But I know this 3 were needed on our gate to make it work | 06:45 |
| ltomasbo | <ltomasbo> OVN_L3_CREATE_PUBLIC_NETWORK=True | 06:45 |
| ltomasbo | <ltomasbo> PHYSICAL_NETWORK=public | 06:45 |
| ltomasbo | <ltomasbo> Q_USE_PROVIDERNET_FOR_PUBLIC=True | 06:45 |
| ltomasbo | the others I took it for the generated localconf on one of our gates, but I think they are not needed | 06:46 |
| ltomasbo | good morning dmellado! | 06:46 |
| yboaron_ | ltomasbo, Yep!, I"ll give it a try | 06:46 |
| yboaron_ | dmellado, Morning | 06:46 |
| dmellado | yboaron_: ltomasbo could you take a look at https://review.openstack.org/#/c/583540/ | 06:53 |
| dmellado | I'd like to get this one merged asap | 06:53 |
| yboaron_ | dmellado, sure, | 06:56 |
| ltomasbo | dmellado, I'll make some reviews in a few minutes... | 06:56 |
| dmellado | ltomasbo: of course after the coffee | 07:01 |
| dmellado | xD | 07:01 |
| ltomasbo | xD | 07:01 |
| *** janki has joined #openstack-kuryr | 07:21 | |
| *** celebdor has joined #openstack-kuryr | 07:29 | |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-tempest-plugin master: Add Network Policy tests https://review.openstack.org/589521 | 07:31 |
| *** garyloug has joined #openstack-kuryr | 07:37 | |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-tempest-plugin master: Remove direct calls to os_client for SGs https://review.openstack.org/599948 | 07:40 |
| *** garyloug has quit IRC | 07:45 | |
| *** gkadam_ has joined #openstack-kuryr | 08:23 | |
| *** gkadam has quit IRC | 08:27 | |
| *** ccamposr has joined #openstack-kuryr | 08:36 | |
| *** ccamposr_ has joined #openstack-kuryr | 08:36 | |
| *** ccamposr_ has quit IRC | 08:37 | |
| ltomasbo | dulek, at https://review.openstack.org/#/c/593679, the pools-octavia failure seemed to be unrelated, now it is passing | 08:45 |
| *** pcaruana has quit IRC | 08:50 | |
| ltomasbo | dulek, and regarding your comment here: https://review.openstack.org/#/c/593679/5/kuryr_kubernetes/controller/drivers/vif_pool.py | 08:57 |
| ltomasbo | dulek, happy to add the lock, but not sure under what situation may that happen? | 08:58 |
| ltomasbo | dulek, specially given the case that if a controller lost the leadership, it will not work if it gets it again (as the watchers will not be restarted) | 08:59 |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: Implement NP SG create/delete actions https://review.openstack.org/583540 | 09:11 |
| dulek | ltomasbo: It's spawning _return_ports_to_pool thread unconditionally, so it's always running. | 09:29 |
| dulek | ltomasbo: So IMO we can hit a situation when _trigger_return_to_pool is running and meanwhile in main thread leadership transition occurs. | 09:30 |
| dulek | ltomasbo: Am I misunderstanding some logic there? It's a bit complicated with all those threads. :) | 09:30 |
| ltomasbo | but return_ports_to_pool is not related to sync_pools | 09:31 |
| ltomasbo | dulek, but yea, it could be some weird behaviour I suppose | 09:31 |
| dulek | ltomasbo: It's iterating over self._recyclable_ports, while sync_pools change value of that variable. | 09:32 |
| dulek | ltomasbo: I think it's potentially dangerous. | 09:32 |
| ltomasbo | dulek, yep, you are right | 09:32 |
| ltomasbo | dulek, though there will be nothing on recyclabpe_ports if the controller is not active | 09:33 |
| ltomasbo | and sync_pools will not add anything to recyclable_ports | 09:33 |
| ltomasbo | but good to have it anyway to be on the safe side | 09:34 |
| dulek | ltomasbo: What if it was active before? We're not cleaning that stuff up in on_revoke_leader. | 09:34 |
| dulek | ltomasbo: I may really misunderstand something there - it's complicated as I've said. :) | 09:34 |
| ltomasbo | dulek, yep, it is getting complex | 09:35 |
| ltomasbo | dulek, I thought about having a empty_pools method for the on_revoke | 09:36 |
| ltomasbo | but then I saw that current code does not allow to become active for a second time | 09:36 |
| ltomasbo | dulek: self.watcher = watcher.Watcher(pipeline, self.tg, exit_on_stop=True) | 09:37 |
| ltomasbo | dulek, that ^^ makes that ones you execute on_revoke, the watcher is lost | 09:37 |
| ltomasbo | you can not restarted with the self.watcher.start() | 09:37 |
| dulek | ltomasbo: Oh crap, I might have missed that. | 09:37 |
| * dulek looks. | 09:37 | |
| ltomasbo | dulek, at least, when I was testing this, if I make a few transitions, after a controller that was active, become active again, I didn't see the watching starting on the logs, adn the pods were stack forever | 09:38 |
| dulek | ltomasbo: But it haven't stopped the process? sys.exit(1) wasn't triggered? Strange… | 09:39 |
| ltomasbo | dulek, perhaps I missed something. I saw the leader elector was working fine | 09:42 |
| ltomasbo | but the watchers were not | 09:42 |
| *** pcaruana has joined #openstack-kuryr | 09:58 | |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: Implement NP SG create/delete actions https://review.openstack.org/583540 | 10:28 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Ensure ports pool are not duplicated with kuryr-controller HA https://review.openstack.org/593679 | 10:33 |
| *** rh-jelabarre has joined #openstack-kuryr | 11:11 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Ensure ports pool are not duplicated with kuryr-controller HA https://review.openstack.org/593679 | 11:35 |
| *** garyloug has joined #openstack-kuryr | 11:52 | |
| *** gkadam_ has quit IRC | 11:54 | |
| *** garyloug has quit IRC | 11:57 | |
| dulek | dmellado: https://review.openstack.org/#/q/project:openstack/kuryr-kubernetes+topic:python3-first - that looks okay. | 12:00 |
| dulek | dmellado: https://review.openstack.org/#/q/project:openstack/kuryr+topic:python3-first - this needs a few abandons. | 12:00 |
| dmellado | dulek: o/ | 12:00 |
| dmellado | hold on, I was referring to our old abandoned repos | 12:01 |
| dulek | dmellado: https://review.openstack.org/#/q/project:openstack/kuryr-libnetwork+topic:python3-first - and this as well. | 12:01 |
| dmellado | such as fuxi and so | 12:01 |
| dulek | Oh, fuxi. | 12:01 |
| dmellado | we'd need to address that for those as well | 12:01 |
| dmellado | yeah, I was wondering whether to address that | 12:01 |
| dmellado | or drop the project totally | 12:01 |
| dulek | dmellado: Yep, fuxi and fuxi-kubernetes look bad in those listings. | 12:01 |
| dmellado | http://lists.openstack.org/pipermail/openstack-dev/2018-September/134239.html | 12:01 |
| dmellado | see there | 12:02 |
| dmellado | dulek: | 12:02 |
| celebdor | fuxi-kubernetes can be dropped | 12:02 |
| dmellado | | kuryr | 14 | 24 | migration in progress | 12:02 |
| celebdor | for fuxi you should check with hongbin | 12:02 |
| dmellado | celebdor: I fuess fuxi-golang as well | 12:02 |
| dmellado | ? | 12:02 |
| dmellado | I'll assign all those reviews to him | 12:02 |
| celebdor | nothin happened on it, did it? | 12:02 |
| dmellado | celebdor: nothing that I can recall of | 12:02 |
| celebdor | so then... | 12:02 |
| dmellado | all right | 12:05 |
| dmellado | celebdor: dulek | 12:05 |
| dmellado | on kuryr we should be fine | 12:05 |
| dmellado | I've abandoned all the dups from Filippo | 12:05 |
| dulek | dmellado: Awesome. I don't have any access to repos besides kuryr-kubernetes, so it's hard to help with the rest for me. | 12:06 |
| dmellado | let's check fuxi/golang and fuxi/kubernetes | 12:06 |
| dmellado | https://review.openstack.org/#/q/project:openstack/fuxi-golang+topic:python3-first+status:open | 12:07 |
| dmellado | https://review.openstack.org/#/q/project:openstack/kuryr-libnetwork+topic:python3-first | 12:08 |
| dmellado | oh, some more abandons here | 12:08 |
| dmellado | celebdor: abandon this patch, pls https://review.openstack.org/#/c/595428/ | 12:10 |
| celebdor | dmellado: done | 12:13 |
| dmellado | all right | 12:14 |
| dmellado | I've cleared up almost everything | 12:15 |
| dmellado | and assigned hongbin to reviewer | 12:15 |
| dmellado | will be sending an email to the ML and if there's no reply we'll remove those projects from under out umbrella | 12:15 |
| *** rh-jelabarre has quit IRC | 12:26 | |
| *** rh-jelabarre has joined #openstack-kuryr | 12:26 | |
| openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: Introduce proper diver mapping in VIF handler https://review.openstack.org/591352 | 12:28 |
| openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: Add SR-IOV pod vif driver https://review.openstack.org/512280 | 12:28 |
| openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: Add SR-IOV binding driver to CNI https://review.openstack.org/512281 | 12:28 |
| openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: Add HOWTO for SRIOV use case https://review.openstack.org/594125 | 12:28 |
| openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: Introduce test case document for SRIOV functionality https://review.openstack.org/600022 | 12:28 |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: Implement NP SG create/delete actions https://review.openstack.org/583540 | 12:33 |
| danil | folks, please review https://review.openstack.org/591352 | 12:57 |
| *** itzikb has quit IRC | 13:12 | |
| *** janki has quit IRC | 14:01 | |
| *** garyloug has joined #openstack-kuryr | 14:12 | |
| *** garyloug_ has joined #openstack-kuryr | 14:13 | |
| *** garyloug has quit IRC | 14:17 | |
| openstackgerrit | Merged openstack/kuryr-libnetwork master: Switch to using stestr https://review.openstack.org/584347 | 14:46 |
| *** yboaron_ has quit IRC | 15:06 | |
| *** takamatsu has quit IRC | 15:10 | |
| *** takamatsu has joined #openstack-kuryr | 15:12 | |
| *** pcaruana has quit IRC | 15:25 | |
| *** itzikb has joined #openstack-kuryr | 15:39 | |
| dulek | celebdor: Seems like you've found first Python 3 incompatibility. Congrats! | 15:55 |
| dulek | celebdor: http://logs.openstack.org/57/599457/7/experimental/kuryr-kubernetes-tempest-daemon-containerized-octavia-py36/33cc194/controller/logs/kubernetes/pod_logs/kube-system-kuryr-cni-ds-dwrfb.txt.gz#_2018-09-04_15_24_01_888 | 15:55 |
| celebdor | dulek: I don't know what you think, but I think that right there deserves my patch merging already :-) | 15:56 |
| dulek | celebdor: Hah, I was thinking to propose the fix first, but you're right. | 15:57 |
| celebdor | :-) | 15:59 |
| *** garyloug_ has quit IRC | 16:11 | |
| *** premsankar has joined #openstack-kuryr | 16:11 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Fix Python 3 compatiblity in K8sClient.watch() https://review.openstack.org/600116 | 16:20 |
| *** ccamposr has quit IRC | 16:37 | |
| *** itzikb has quit IRC | 16:42 | |
| openstackgerrit | Merged openstack/kuryr-tempest-plugin master: NPWG multi-vif driver tempest tests https://review.openstack.org/590643 | 17:09 |
| dulek | I'm officially declaring watcher.py spaghetti code. We need to refactor it baaaadly. | 17:24 |
| dulek | Just spent an hour until I've realized that we're doing hard .stop() on a thread, not giving it a chance to clean up. | 17:24 |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Clean up watch resources after watcher.stop() https://review.openstack.org/600142 | 17:35 |
| dulek | ltomasbo: ^ | 17:35 |
| *** yboaron_ has joined #openstack-kuryr | 17:47 | |
| *** pcaruana has joined #openstack-kuryr | 17:52 | |
| *** openstackgerrit has quit IRC | 18:00 | |
| *** yboaron_ has quit IRC | 18:16 | |
| *** yboaron_ has joined #openstack-kuryr | 18:16 | |
| *** openstackgerrit has joined #openstack-kuryr | 19:08 | |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: Implement NP SG create/delete actions https://review.openstack.org/583540 | 19:08 |
| *** yboaron_ has quit IRC | 20:00 | |
| *** pcaruana has quit IRC | 20:40 | |
| *** celebdor has quit IRC | 21:14 | |
| *** takamatsu has quit IRC | 21:42 | |
| *** rh-jelabarre has quit IRC | 22:22 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!