| *** kiennt has joined #openstack-kuryr | 00:49 | |
| *** yedongcan has joined #openstack-kuryr | 01:09 | |
| *** hongbin has joined #openstack-kuryr | 01:14 | |
| *** gouthamr has quit IRC | 01:25 | |
| *** caowei has joined #openstack-kuryr | 02:13 | |
| *** yamamoto has joined #openstack-kuryr | 02:17 | |
| *** zengchen has quit IRC | 02:18 | |
| *** gouthamr has joined #openstack-kuryr | 02:21 | |
| *** yamamoto has quit IRC | 02:38 | |
| *** yamamoto has joined #openstack-kuryr | 02:48 | |
| *** pmannidi has joined #openstack-kuryr | 03:05 | |
| *** mattmceuen has joined #openstack-kuryr | 03:29 | |
| *** hongbin has quit IRC | 03:36 | |
| *** mattmceuen has quit IRC | 03:43 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/fuxi master: Updated from global requirements https://review.openstack.org/483522 | 03:55 |
|---|---|---|
| *** yamamoto has quit IRC | 03:57 | |
| *** yamamoto has joined #openstack-kuryr | 03:59 | |
| *** yamamoto has quit IRC | 04:04 | |
| *** yamamoto has joined #openstack-kuryr | 04:05 | |
| *** gouthamr has quit IRC | 04:14 | |
| *** yamamoto has quit IRC | 04:14 | |
| *** reedip has quit IRC | 04:14 | |
| *** reedip has joined #openstack-kuryr | 04:15 | |
| *** yamamoto has joined #openstack-kuryr | 04:15 | |
| *** yamamoto has quit IRC | 04:21 | |
| *** janki has joined #openstack-kuryr | 04:33 | |
| *** gouthamr has joined #openstack-kuryr | 04:42 | |
| *** yamamoto has joined #openstack-kuryr | 04:44 | |
| *** janki has quit IRC | 04:50 | |
| *** gouthamr has quit IRC | 05:07 | |
| *** janki has joined #openstack-kuryr | 05:30 | |
| *** tonanhngo has quit IRC | 05:34 | |
| *** kiennt has quit IRC | 05:42 | |
| *** kiennt has joined #openstack-kuryr | 06:02 | |
| *** caowei_ has joined #openstack-kuryr | 06:16 | |
| *** caowei has quit IRC | 06:16 | |
| *** caowei_ is now known as caowei | 06:16 | |
| *** kiennt has quit IRC | 06:21 | |
| *** kzaitsev_ws has joined #openstack-kuryr | 06:22 | |
| irenab | kzaitsev_ws, will try to chat a bit later, a lot of meetings back to back. Will ping you | 06:30 |
| *** kzaitsev_ has joined #openstack-kuryr | 06:31 | |
| kzaitsev_ | irenab: thanks, I also am a bit busy this morning. :) | 06:33 |
| *** kzaitsev_ has quit IRC | 06:37 | |
| *** tonanhngo has joined #openstack-kuryr | 07:01 | |
| *** tonanhngo has quit IRC | 07:04 | |
| *** vikasc has quit IRC | 07:13 | |
| *** vikasc has joined #openstack-kuryr | 07:15 | |
| *** aojea has joined #openstack-kuryr | 07:20 | |
| apuimedo | dims: sorry. I was on PTO | 07:29 |
| apuimedo | dulek: any idea about that bonds failure in pyroute2? | 07:30 |
| apuimedo | I can talk with svinota if we have more info | 07:30 |
| *** kiennt has joined #openstack-kuryr | 07:40 | |
| dmellado | apuimedo: welcome back xD | 07:41 |
| apuimedo | thanks dmellado | 07:41 |
| apuimedo | and thanks for leading the irc meeting | 07:41 |
| apuimedo | I forgot to delegate | 07:42 |
| dmellado | apuimedo: np! I'll add that to the fuet debt backlog | 07:42 |
| apuimedo | :-) | 07:43 |
| *** egonzalez has joined #openstack-kuryr | 07:43 | |
| apuimedo | vikasc: ping | 07:46 |
| vikasc | apuimedo, pong | 07:46 |
| apuimedo | did you see dulek's comments to https://review.openstack.org/#/c/466675/ | 07:46 |
| apuimedo | ? | 07:46 |
| vikasc | apuimedo, yeah | 07:47 |
| apuimedo | tbf, I'm puzzled about the pyroute2 thingie | 07:47 |
| apuimedo | maybe the bond module was absent from his env | 07:47 |
| apuimedo | who knows | 07:47 |
| *** pmannidi has quit IRC | 07:48 | |
| vikasc | yeah, might be | 07:50 |
| vikasc | i need to check on ubuntu | 07:50 |
| vikasc | apuimedo, will update the patch and reply to his comments | 07:51 |
| apuimedo | I'm replying to some now | 07:51 |
| vikasc | sure | 07:52 |
| janonymous | vikasc: Also any hints on: http://paste.openstack.org/show/615794/ | 07:57 |
| dulek | apuimedo, vikasc: No idea and I cannot really dug deeply into it now. | 07:59 |
| dulek | However I've noticed that when the lib is used the same way on my bare 16.04 (not from Docker on that host) everything works fine. | 07:59 |
| dulek | So it's combination of 16.04 host + CentOS (?) container that triggers the problem IMO. | 08:00 |
| apuimedo | dulek: it is not really being used from Docker though, is it? | 08:00 |
| apuimedo | it is called by kubelet, or does kubelet run containerized? | 08:00 |
| apuimedo | vikasc: I added quite a few comments | 08:01 |
| apuimedo | dulek: vikasc: About the CNi needing to copy some certs, it should use token instead | 08:01 |
| apuimedo | oh, shoot | 08:01 |
| apuimedo | I'm dumb | 08:01 |
| apuimedo | of course, if it doesn't run inside the container, it won't get the token | 08:01 |
| * apuimedo needs to wake up | 08:02 | |
| vikasc | :) | 08:02 |
| apuimedo | janonymous: what's that failure about? | 08:02 |
| dulek | apuimedo: Yup! Although kubelet needs certs to call the kube-api anyway, right? | 08:02 |
| apuimedo | dulek: right. And vikasc patch seems to use kubelet's certs | 08:02 |
| * dulek is having a coffee now, so might be slow too. | 08:02 | |
| apuimedo | so I wonder why you had to copy stuff | 08:03 |
| vikasc | apuimedo, +1 | 08:03 |
| dulek | Let me take a look on kubelet config. | 08:03 |
| *** yedongcan has quit IRC | 08:03 | |
| janonymous | vikasc: When it tries to add vif , it throws traceback on recv() call | 08:04 |
| janonymous | apuimedo: ^ | 08:04 |
| dulek | Ah, kubeadm pastes certs directly into /etc/kubernetes/kubelet.conf. | 08:04 |
| janonymous | apuimedo: welcome back :) | 08:04 |
| apuimedo | thanks | 08:05 |
| janonymous | apuimedo: it is related to https://review.openstack.org/#/c/480028 patch | 08:05 |
| *** yedongcan has joined #openstack-kuryr | 08:05 | |
| dulek | apuimedo: And you're right, my kubelet runs as a process, so it's even more puzzling. When I've installed the lib through pip it works fine. | 08:05 |
| dulek | But not if it was included into the binary. | 08:05 |
| apuimedo | mmm | 08:05 |
| apuimedo | dulek: oh, of course | 08:06 |
| apuimedo | pypi has 0.4.17 which I patched to add support | 08:06 |
| apuimedo | maybe when the binary is built it pulls something earlier | 08:06 |
| apuimedo | s/it pulls/it pulled/ | 08:06 |
| apuimedo | the error seems consistent with the previous behavior | 08:06 |
| apuimedo | dulek: vikasc: I'm tempted to have the CNI container just write the darned token to /etc/ | 08:07 |
| dulek | apuimedo: "pip3.5 install pyroute2" - looks like the latest version and I've built the Docker image myself, so it should be new. | 08:08 |
| dulek | I mean here: https://review.openstack.org/#/c/466675/14/cni_builder.Dockerfile@31 | 08:09 |
| apuimedo | vikasc: did the same happen to you? | 08:10 |
| dulek | apuimedo: Cert location is configurable in configmap, so I think that documenting that it needs to point to correct cert is enough. | 08:10 |
| vikasc | apuimedo, this pyroute2 problem never happened to me | 08:10 |
| apuimedo | ok | 08:11 |
| apuimedo | so then maybe it is bond not being loaded | 08:11 |
| apuimedo | I'll investigate | 08:11 |
| apuimedo | dulek: but which cert? | 08:11 |
| apuimedo | we can't read the .conf at present | 08:11 |
| vikasc | i will also investigate on my side | 08:12 |
| dulek | apuimedo: I mean that it's a fair requirement from the user/admin to set ssl_client_*_file options in kuryr.yml to point to certificates. | 08:13 |
| dulek | apuimedo: He needs to set some stuff there anyway, e.g. subnets ids. | 08:14 |
| apuimedo | dulek: but it would be much smoother to put the token file in /etc/kuryr/cni_token | 08:14 |
| apuimedo | no need to do anything from the admin side | 08:14 |
| dulek | apuimedo: At first glance there isn't anything wrong with that. But I normally get paranoid around auth stuff, so it requires more thinking. ;) | 08:16 |
| apuimedo | dulek: it would be root readable only 400 | 08:18 |
| apuimedo | so I think it should be fine, considering there's already the kubelet stronger certs in /etc | 08:18 |
| vikasc | apuimedo, are you suggesting that cni builder container should copy token as well(along with executable) to the host? | 08:19 |
| apuimedo | yeah | 08:20 |
| apuimedo | this way we don't depend on different deployment behavior on the k8s side | 08:20 |
| dulek | Well, certs are there in /etc/kubernetes/kubelet, so I can agree it's no different. | 08:20 |
| vikasc | apuimedo, yeah, as dulek said | 08:21 |
| dulek | s/kubelet/kubelet.conf/ | 08:21 |
| vikasc | apuimedo, user need to copy any certs etc | 08:21 |
| apuimedo | vikasc: so then I prefer the builder do it | 08:21 |
| vikasc | apuimedo, s/no need | 08:22 |
| vikasc | apuimedo, cni uses kubelet certs | 08:22 |
| vikasc | apuimedo, i will try to make to make it work with token as well. | 08:23 |
| *** yedongcan has quit IRC | 08:24 | |
| apuimedo | vikasc: the thing is, the certs may not be there | 08:24 |
| apuimedo | they may be only embedded in .conf which we don't read | 08:25 |
| *** yedongcan has joined #openstack-kuryr | 08:25 | |
| apuimedo | and in any case, if we run with kubelet certs it's also more perms than we need | 08:25 |
| vikasc | apuimedo, i will try out tokens | 08:25 |
| apuimedo | thanks | 08:27 |
| apuimedo | vikasc: IIRC in my openshift-ansible patches there's already the clusterrole and clusterrolebinding for CNI | 08:27 |
| apuimedo | so you can reuse that | 08:27 |
| apuimedo | or maybe not | 08:27 |
| apuimedo | maybe I only have it on some presentation | 08:27 |
| apuimedo | XD | 08:27 |
| apuimedo | I go on holidays 5 days and forget where I have things | 08:27 |
| vikasc | apuimedo, i will figure out :) | 08:28 |
| apuimedo | good. Thanks | 08:29 |
| vikasc | yw! | 08:29 |
| *** garyloug has joined #openstack-kuryr | 08:45 | |
| *** garyloug has quit IRC | 08:46 | |
| irenab | apuimedo, morning | 08:48 |
| *** openstackgerrit has quit IRC | 08:49 | |
| kzaitsev_ws | apuimedo: you probably saw kuryr 0.5.0 release | 08:51 |
| kzaitsev_ws | https://review.openstack.org/#/c/484420/ here's a requirements followup | 08:52 |
| apuimedo | irenab: morning | 08:53 |
| apuimedo | kzaitsev_ws: I saw :-) | 08:53 |
| apuimedo | yesterday night | 08:53 |
| kzaitsev_ws | sorry for pushing the things (= | 08:53 |
| apuimedo | kzaitsev_ws: no, no. Big thanks | 08:54 |
| apuimedo | we should also push for upping the constraints on pyroute2 | 08:54 |
| kzaitsev_ws | yeah, at some point I'll ask to be a release liaison for kuryr, if you folks don't mind =P | 08:55 |
| kzaitsev_ws | apuimedo: https://github.com/openstack/requirements/blob/master/global-requirements.txt#L218 it's >=0.4.17 | 08:55 |
| kzaitsev_ws | there's 0.4.18 on the pypi however | 08:56 |
| apuimedo | kzaitsev_ws: I'd be happy if you could do that | 09:01 |
| apuimedo | 0.4.17 is fine | 09:01 |
| dmellado | apuimedo: ping meeting | 09:01 |
| apuimedo | you don't see me? | 09:02 |
| apuimedo | I'm there | 09:02 |
| kzaitsev_ws | apuimedo: I can. I do that for murano anyways. I'll ask around on the next community meeting. | 09:02 |
| *** yamamoto has quit IRC | 09:06 | |
| *** garyloug has joined #openstack-kuryr | 09:06 | |
| irenab | ltomasbo, vikasc any chance you had nested environment recently? I am not able to make pods running in nested env | 09:10 |
| *** yamamoto has joined #openstack-kuryr | 09:10 | |
| *** yamamoto has quit IRC | 09:10 | |
| ltomasbo | hi, I didn't deploy any recently | 09:11 |
| ltomasbo | fighting with ODL now, and trunk ports is not working well there | 09:12 |
| ltomasbo | irenab, are they in ContainerCreating or Pending status? | 09:12 |
| irenab | yes | 09:12 |
| irenab | trying to figure out why, but do not see anything specific | 09:12 |
| ltomasbo | do you see any problem with kubelet? | 09:13 |
| irenab | perodically it fails for missing HostIP, but not always | 09:13 |
| ltomasbo | ummm, I remember seeing that at some point | 09:13 |
| ltomasbo | but they remain in ContainerCreating then? | 09:14 |
| irenab | comntext deadline exceeded | 09:14 |
| ltomasbo | not pending | 09:14 |
| *** yamamoto has joined #openstack-kuryr | 09:14 | |
| irenab | yes, stack at container creating | 09:14 |
| vikasc | irenab neither me | 09:15 |
| irenab | in the describe, the status is pending | 09:15 |
| irenab | in the events I see MissingClusterDNS | 09:16 |
| ltomasbo | does this only happen to you on the nested one? | 09:16 |
| irenab | yes | 09:17 |
| irenab | BM works fine | 09:17 |
| ltomasbo | I think I saw those errors in both, BM and Nested | 09:17 |
| ltomasbo | Warning MissingClusterDNS kubelet does not have ClusterDNS IP configured and cannot create Pod using "ClusterFirst" policy. Falling back to DNSDefault policy. | 09:18 |
| ltomasbo | I'm getting that one BM | 09:18 |
| ltomasbo | but the container starts anyway | 09:18 |
| irenab | but pods are running, right? | 09:19 |
| ltomasbo | yep | 09:19 |
| irenab | yea, I do not think this is the problem. I do see annotation with vif details on kubelet side | 09:19 |
| irenab | so problem seems to be the kubelet related | 09:19 |
| irenab | cannot figure out what exactly .. | 09:20 |
| *** caowei has quit IRC | 09:20 | |
| ltomasbo | umm | 09:20 |
| ltomasbo | we got some problems with etcd and raised to 3.0.12 | 09:20 |
| ltomasbo | but at the end we found problems somewhere else, so probably is was not related | 09:20 |
| ltomasbo | irenab, no errors at kubelet? | 09:28 |
| vikasc | irenab, just see if kuryr-cni is getting invoked in kubelet logs | 09:29 |
| vikasc | irenab, if annotation is there means controller is working fine | 09:30 |
| irenab | vikasc, yes. Agree. Just cannot see anything specific in the kubelet | 09:31 |
| dmellado | apuimedo: you looked like you were praying | 09:31 |
| dmellado | xD | 09:31 |
| vikasc | irenab, there is a specific log which you should search for in kubelet logs to ensure if cni is getting invoked | 09:32 |
| irenab | which one? | 09:32 |
| vikasc | irenab, i once told ltomasbo | 09:32 |
| vikasc | irenab, looking for it if i haved it somewhere | 09:32 |
| vikasc | :) | 09:32 |
| vikasc | s/have saved | 09:32 |
| vikasc | irenab, a moment pls | 09:33 |
| irenab | vikasc, interesting ... I prionted vif.active status and it is false | 09:35 |
| irenab | will dig into it | 09:35 |
| irenab | ping | 09:36 |
| *** yamamoto has quit IRC | 09:36 | |
| vikasc | irenab, means cni is not plugging | 09:36 |
| irenab | vikasc, I thought its comming from neutron | 09:36 |
| *** yamamoto has joined #openstack-kuryr | 09:36 | |
| ltomasbo | umm | 09:37 |
| vikasc | irenab, vif.status will become active after vif plugging is done, right? | 09:37 |
| ltomasbo | can you test one think irenab | 09:37 |
| vikasc | s/binding | 09:37 |
| ltomasbo | I'm fighting with one issue and thought it was related to ODL | 09:37 |
| ltomasbo | but perhaps something went wrong in neutron trunk ports support | 09:37 |
| ltomasbo | can you check if when adding a subport to the trunk | 09:38 |
| ltomasbo | it becomes active? | 09:38 |
| ltomasbo | the subport I mean | 09:38 |
| irenab | a sec | 09:38 |
| ltomasbo | it used to be, but I have it 'down' when using ODL and thought it was where the problem was coming from | 09:38 |
| irenab | ltomasbo, I am checking with DragonFlow, so we are in the similar boats :-) | 09:39 |
| ltomasbo | :D | 09:39 |
| irenab | Down | 09:39 |
| ltomasbo | there you are... | 09:39 |
| irenab | brb | 09:39 |
| ltomasbo | so, will this happen also with ml2/ovs? | 09:39 |
| ltomasbo | I'm going to deploy one and check | 09:47 |
| apuimedo | ltomasbo: shouldn't. You think it broke recently? | 09:47 |
| ltomasbo | I thought it was broken in ODL (or I don't know how to configure it) | 09:47 |
| ltomasbo | but perhaps it is broken on neutron side... | 09:48 |
| ltomasbo | as it is happening to irenab on DragonFlow too | 09:48 |
| ltomasbo | I'm creating a VM and I'll check | 09:48 |
| apuimedo | ltomasbo: the fact that both odl and df are just starting with trunk ports gets it equally likely to be on Neutron as in df and odl | 10:00 |
| apuimedo | janonymous: which eventlet version do you run? | 10:01 |
| janonymous | apuimedo: 0.20.0 | 10:03 |
| janonymous | apuimedo: it is related to eventlet, as when i moved it out of eventlet to run service it worked.. | 10:03 |
| apuimedo | yeah | 10:04 |
| apuimedo | janonymous: I wonder why openstack/requirements finds 0.20.0 acceptable but not 0.20.1 | 10:04 |
| janonymous | apuimedo: for now i guess moving daemon out of eventlet would work but have to find way to patch/make it work with eventlet as well | 10:07 |
| ltomasbo | apuimedo, maybe you can talk to amoralej or dalvarez about the eventlet version thing issue | 10:07 |
| ltomasbo | I believe there was some problem with some memory leaks | 10:08 |
| ltomasbo | and some problems with the dns | 10:08 |
| ltomasbo | but maybe it is not even the same version... | 10:08 |
| ltomasbo | yep, it was that | 10:10 |
| apuimedo | janonymous: we should really patch it to make it work with eventlet | 10:10 |
| ltomasbo | from ihar: Eventlet, tire fire. 0.20.0 takes 30s to recover from broken DNS with entry in /etc/hosts. 0.20.1 - 60s. 0.21 is ok, but has broken timers. | 10:10 |
| apuimedo | ltomasbo: yeah, yeah. I know | 10:10 |
| apuimedo | eventlet is tire fire | 10:10 |
| apuimedo | svinota never fails to mention it | 10:11 |
| ltomasbo | :D | 10:11 |
| apuimedo | ltomasbo: I know it was necessary, but you don't know how much I regret not being able to just use our old 3.5 asyncio code | 10:11 |
| apuimedo | I never liked eventlet | 10:11 |
| ltomasbo | :D | 10:11 |
| janonymous | apuimedo: yeah.. | 10:12 |
| janonymous | :) | 10:12 |
| *** atoth has quit IRC | 10:17 | |
| apuimedo | ltomasbo: https://blueprints.launchpad.net/kuryr-kubernetes/+spec/kuryr-manager-cli-tool | 10:17 |
| apuimedo | janonymous: do you think it would be very difficult to check if it works with https://cotyledon.readthedocs.io/en/latest/ ? | 10:21 |
| apuimedo | https://cotyledon.readthedocs.io/en/latest/oslo-service-migration.html | 10:22 |
| *** caowei has joined #openstack-kuryr | 10:23 | |
| apuimedo | octavia uses it as well | 10:24 |
| janonymous | apuimedo: will try to go through it | 10:28 |
| apuimedo | janonymous: I'm now checking it as well | 10:28 |
| janonymous | apuimedo: thanks | 10:28 |
| *** kiennt has quit IRC | 10:45 | |
| *** yedongcan has left #openstack-kuryr | 10:46 | |
| *** openstackgerrit has joined #openstack-kuryr | 10:48 | |
| openstackgerrit | Kirill Zaitsev proposed openstack/kuryr-kubernetes master: Allow requesting additional subnets via annotation https://review.openstack.org/482544 | 10:48 |
| irenab | apuimedo, https://review.openstack.org/#/c/484804/ | 10:50 |
| janonymous | apuimedo: i think https://review.openstack.org/#/c/480028/8/kuryr_kubernetes/cni/cni_daemon.py needs to be changed to oslo service model first for that... | 10:53 |
| dims | apuimedo : welcome back. we took the liberty of shipping kuryr-lib :) | 10:55 |
| apuimedo | dims: thanks for that | 10:55 |
| apuimedo | :-) | 10:55 |
| *** yamamoto has quit IRC | 10:55 | |
| apuimedo | dims: do you know off the top of your head which projects use etcd3 now that it is a base service? | 10:56 |
| dims | apuimedo : cinder requires it for locking, but it's through tooz. i am encouraging direct dependency on etcd3 for queens | 10:57 |
| apuimedo | interesting | 10:57 |
| apuimedo | dims: any reason for not using tooz? | 10:58 |
| dims | i mean if they choose to | 10:58 |
| apuimedo | dims: currently we're impacted by eventlet breakage with pyroute2. I'm considering very hard about leaving eventlet | 10:58 |
| apuimedo | but have to check how feasible it would be | 10:58 |
| dims | i'd encourage that very much apuimedo | 10:58 |
| *** neiljerram has joined #openstack-kuryr | 10:59 | |
| apuimedo | dims: with cotyledon or something else? | 10:59 |
| *** tonanhngo has joined #openstack-kuryr | 11:03 | |
| *** tonanhngo has quit IRC | 11:05 | |
| *** atoth has joined #openstack-kuryr | 11:05 | |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Fix k8s service subnet devstack setting https://review.openstack.org/484804 | 11:08 |
| *** rwallner has joined #openstack-kuryr | 11:10 | |
| apuimedo | irenab: ^^ was fast | 11:10 |
| irenab | apuimedo, thanks | 11:11 |
| *** roeyc has joined #openstack-kuryr | 11:11 | |
| irenab | ltomasbo, I eliminated the check for vif.status to be active and pods are running. Let me know if you reveaved any change in the neutron trunk | 11:12 |
| ltomasbo | irenab, I'm deploying a devstack, but it failed, so I'm re-deploying | 11:13 |
| dims | apuimedo : talk to keystone folks, they were the first to cleanup eventlet from what i remember | 11:13 |
| ltomasbo | irenab, I'll let you know as soon as I get that up and running and check if subports become active or not | 11:13 |
| irenab | apuimedo, for nested, we need to add lbaas stuff removal for unstack.sh, annoying to clean it manually when trying to restack ... | 11:13 |
| ltomasbo | :D | 11:14 |
| irenab | ltomasbo, thanks | 11:14 |
| ltomasbo | irenab, no problem, this is biting me too... | 11:14 |
| apuimedo | irenab: my devstack-heat-overcloud removes it thanks to heat | 11:14 |
| apuimedo | or you mean manual? | 11:15 |
| irenab | apuimedo, I didn't want to remove the VM, just restack inside | 11:15 |
| irenab | apuimedo, when in the overcloud VM and you need to restack | 11:15 |
| irenab | but agree, with heat you can just redeploy the stack | 11:16 |
| irenab | heat-stack | 11:16 |
| irenab | apuimedo, need to get the heat patch merged. | 11:17 |
| apuimedo | irenab: afaik it is | 11:17 |
| openstackgerrit | Kirill Zaitsev proposed openstack/kuryr-kubernetes master: Allow requesting additional subnets via annotation https://review.openstack.org/482544 | 11:17 |
| irenab | apuimedo, https://review.openstack.org/#/c/472763/ | 11:18 |
| ltomasbo | irenab, I do have is an ovn based one, and the same problem happens, the subport is not active either | 11:18 |
| irenab | ltomasbo, proably need to check for the neutron change | 11:19 |
| ltomasbo | yep | 11:19 |
| irenab | Maybe we need to skip this check for the trunk sub ports | 11:19 |
| irenab | I wonder if the port status related to it being subport | 11:20 |
| ltomasbo | irenab, if the port is not up | 11:25 |
| ltomasbo | I think we don't have connectivity | 11:25 |
| ltomasbo | can you try to ping it from outside | 11:25 |
| ltomasbo | ? | 11:25 |
| apuimedo | irenab: oh, I thought you meant the Heat patch that adds trunk port support | 11:27 |
| apuimedo | xD | 11:28 |
| *** yamamoto has joined #openstack-kuryr | 11:40 | |
| ltomasbo | irenab, I can confirm this is not happening with ml2/ovs | 11:41 |
| ltomasbo | the subport status becomes active when you attach it | 11:41 |
| ltomasbo | so, both ovn, odl and df have some problems with trunk support | 11:41 |
| irenab | ltomasbo, I do not have ping | 11:48 |
| irenab | and see all ports on pod-subnet in Down stat, including router interface | 11:48 |
| irenab | ltomasbo, with apuimedo heat patch I till had to run stack.sh inside the VM | 11:49 |
| irenab | s/till/still | 11:49 |
| ltomasbo | irenab, I believe it is not becoming active for a reason | 11:53 |
| *** aojea has quit IRC | 11:53 | |
| ltomasbo | perhaps there is some missing steps in the odl/df/ovn drivers when dealing with trunk ports | 11:53 |
| ltomasbo | irenab, ohh, I did not read the other sentence | 11:54 |
| ltomasbo | can you create a VM in that subnet? | 11:54 |
| irenab | ltomasbo, will try a bit later, another meeting ... | 11:54 |
| ltomasbo | sure! I'll keep digging... | 11:54 |
| *** aojea has joined #openstack-kuryr | 12:02 | |
| *** yamamoto has quit IRC | 12:03 | |
| *** aojea has quit IRC | 12:07 | |
| *** yamamoto has joined #openstack-kuryr | 12:21 | |
| *** aojea has joined #openstack-kuryr | 12:21 | |
| *** neiljerram has quit IRC | 12:23 | |
| *** aojea has quit IRC | 12:25 | |
| *** caowei has quit IRC | 12:26 | |
| *** kural has joined #openstack-kuryr | 12:33 | |
| irenab | ltomasbo, VM is created and its port is Active | 12:33 |
| apuimedo | irenab: can you create a subport manually? | 12:34 |
| apuimedo | and check | 12:35 |
| apuimedo | then with manually created vlan wether it works as intended | 12:35 |
| irenab | apuimedo, will try | 12:36 |
| apuimedo | thanks | 12:37 |
| *** aojea has joined #openstack-kuryr | 12:39 | |
| irenab | created in down | 12:41 |
| *** aojea has quit IRC | 12:44 | |
| *** aojea has joined #openstack-kuryr | 12:48 | |
| ltomasbo | irenab, the parent port is Active, but not the subports, right? | 12:52 |
| irenab | yes | 12:53 |
| *** aojea has quit IRC | 12:53 | |
| apuimedo | it should automatically go active, shouldn't it? | 12:54 |
| ltomasbo | irenab, yes, the subport is down, right? | 12:55 |
| ltomasbo | apuimedo, it should go if the VM is already up | 12:56 |
| irenab | ltomasbo, parent is Active, subport is Down | 12:56 |
| ltomasbo | not sure how df does the trunks | 12:56 |
| apuimedo | irenab: what is DF waiting for to put it up? | 12:56 |
| apuimedo | is it reactive? | 12:56 |
| ltomasbo | but odl does then without the need of the extra ovs-trunk bridge | 12:56 |
| irenab | ltomasbo, I think its more of control plane stuff | 12:56 |
| ltomasbo | therefore, br-int only has one port (parent) | 12:56 |
| irenab | in DF subports are implemented by flows | 12:57 |
| ltomasbo | I'm betting that it one of the problem to make the port active | 12:57 |
| irenab | so also no extra bridges | 12:57 |
| ltomasbo | yep, no extra bridge for odl either | 12:57 |
| *** aojea has joined #openstack-kuryr | 12:57 | |
| irenab | Do you see where in ref implementation subport status is set to Active? | 12:58 |
| *** gouthamr has joined #openstack-kuryr | 12:58 | |
| ltomasbo | didn't look for it, but maybe it is a callback from nova when it is plugged to br-int | 12:59 |
| irenab | I bet its the callback from the ovs agent | 13:00 |
| ltomasbo | that could be, and it would be a reason for not having it active in odl... | 13:00 |
| irenab | same for DF. so problem is not with kuryr or neutron | 13:01 |
| irenab | Im my case no ping either even if I mask the check for vif being active | 13:01 |
| irenab | ltomasbo, need to go. Tomorrow will be another day ... | 13:01 |
| ltomasbo | :D | 13:02 |
| *** aojea has quit IRC | 13:02 | |
| ltomasbo | have fun! | 13:02 |
| *** aojea has joined #openstack-kuryr | 13:06 | |
| *** garyloug has quit IRC | 13:10 | |
| *** aojea has quit IRC | 13:11 | |
| *** garyloug has joined #openstack-kuryr | 13:14 | |
| *** aojea has joined #openstack-kuryr | 13:25 | |
| *** aojea has quit IRC | 13:30 | |
| *** kural has quit IRC | 13:43 | |
| *** kural has joined #openstack-kuryr | 13:47 | |
| *** lihi has quit IRC | 13:47 | |
| *** kural has quit IRC | 13:47 | |
| *** kural_ has joined #openstack-kuryr | 13:47 | |
| *** aojea has joined #openstack-kuryr | 13:48 | |
| *** lihi has joined #openstack-kuryr | 13:48 | |
| *** hongbin has joined #openstack-kuryr | 13:49 | |
| *** tonanhngo has joined #openstack-kuryr | 14:01 | |
| *** roeyc has quit IRC | 14:02 | |
| *** tonanhngo has quit IRC | 14:03 | |
| *** kural has joined #openstack-kuryr | 14:04 | |
| *** kural has quit IRC | 14:05 | |
| *** kural_ has quit IRC | 14:05 | |
| *** kural has joined #openstack-kuryr | 14:06 | |
| *** yamamoto has quit IRC | 14:09 | |
| *** roeyc has joined #openstack-kuryr | 14:32 | |
| *** janki has quit IRC | 15:03 | |
| *** yamamoto has joined #openstack-kuryr | 15:09 | |
| *** yamamoto has quit IRC | 15:19 | |
| *** rwallner has quit IRC | 15:22 | |
| *** rwallner has joined #openstack-kuryr | 15:23 | |
| *** kzaitsev_ws has quit IRC | 15:38 | |
| *** kzaitsev_ws has joined #openstack-kuryr | 15:52 | |
| *** sridhar_ram has quit IRC | 16:06 | |
| *** aojea has quit IRC | 16:13 | |
| *** aojea has joined #openstack-kuryr | 16:14 | |
| *** aojea has quit IRC | 16:18 | |
| *** roeyc has quit IRC | 16:34 | |
| *** aojea has joined #openstack-kuryr | 16:41 | |
| *** aojea has quit IRC | 16:45 | |
| *** aojea has joined #openstack-kuryr | 16:50 | |
| *** aojea has quit IRC | 16:54 | |
| *** garyloug has quit IRC | 17:01 | |
| *** egonzalez has quit IRC | 17:02 | |
| *** kural has quit IRC | 17:04 | |
| *** kural has joined #openstack-kuryr | 17:04 | |
| *** kural_ has joined #openstack-kuryr | 17:05 | |
| *** kural has quit IRC | 17:05 | |
| *** aojea has joined #openstack-kuryr | 17:16 | |
| *** aojea has quit IRC | 17:22 | |
| *** tonanhngo has joined #openstack-kuryr | 17:22 | |
| *** atoth has quit IRC | 17:51 | |
| *** kural_ has quit IRC | 17:52 | |
| *** atoth has joined #openstack-kuryr | 17:57 | |
| *** atoth has quit IRC | 18:06 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr master: Updated from global requirements https://review.openstack.org/485310 | 18:35 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-libnetwork master: Updated from global requirements https://review.openstack.org/484555 | 18:35 |
| *** vikasc has quit IRC | 19:04 | |
| *** vikasc has joined #openstack-kuryr | 19:05 | |
| *** aojea has joined #openstack-kuryr | 19:05 | |
| *** aojea has quit IRC | 19:10 | |
| *** aojea has joined #openstack-kuryr | 19:15 | |
| *** aojea has quit IRC | 19:19 | |
| *** aojea has joined #openstack-kuryr | 19:24 | |
| *** aojea has quit IRC | 19:28 | |
| *** aojea has joined #openstack-kuryr | 19:33 | |
| *** aojea has quit IRC | 19:38 | |
| *** rwallner has quit IRC | 20:14 | |
| *** rwallner has joined #openstack-kuryr | 20:18 | |
| *** rwallner has quit IRC | 20:22 | |
| *** aojea has joined #openstack-kuryr | 20:36 | |
| *** aojea has quit IRC | 20:41 | |
| *** aojea has joined #openstack-kuryr | 20:46 | |
| *** aojea has quit IRC | 20:50 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/fuxi master: Updated from global requirements https://review.openstack.org/483522 | 20:52 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-libnetwork master: Updated from global requirements https://review.openstack.org/484555 | 20:54 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-kubernetes master: Updated from global requirements https://review.openstack.org/481186 | 20:54 |
| *** aojea has joined #openstack-kuryr | 20:55 | |
| *** aojea has quit IRC | 20:59 | |
| *** aojea has joined #openstack-kuryr | 21:04 | |
| *** aojea has quit IRC | 21:08 | |
| *** aojea has joined #openstack-kuryr | 21:13 | |
| *** aojea has quit IRC | 21:18 | |
| *** aojea has joined #openstack-kuryr | 21:22 | |
| *** aojea has quit IRC | 21:27 | |
| *** aojea has joined #openstack-kuryr | 21:31 | |
| *** aojea has quit IRC | 21:35 | |
| *** aojea has joined #openstack-kuryr | 21:40 | |
| *** aojea has quit IRC | 21:44 | |
| *** gouthamr has quit IRC | 21:49 | |
| *** aojea has joined #openstack-kuryr | 21:49 | |
| *** aojea has quit IRC | 21:53 | |
| *** rwallner has joined #openstack-kuryr | 21:54 | |
| *** rwallner has quit IRC | 21:55 | |
| *** rwallner has joined #openstack-kuryr | 21:55 | |
| *** aojea has joined #openstack-kuryr | 21:59 | |
| *** aojea has quit IRC | 22:03 | |
| *** aojea has joined #openstack-kuryr | 22:08 | |
| *** aojea has quit IRC | 22:12 | |
| *** aojea has joined #openstack-kuryr | 22:17 | |
| *** gouthamr has joined #openstack-kuryr | 22:18 | |
| *** aojea has quit IRC | 22:21 | |
| *** aojea has joined #openstack-kuryr | 22:26 | |
| *** aojea has quit IRC | 22:30 | |
| *** aojea has joined #openstack-kuryr | 22:35 | |
| *** aojea has quit IRC | 22:39 | |
| *** aojea has joined #openstack-kuryr | 23:02 | |
| *** rwallner has quit IRC | 23:04 | |
| *** yamamoto has joined #openstack-kuryr | 23:04 | |
| *** aojea has quit IRC | 23:07 | |
| *** aojea has joined #openstack-kuryr | 23:11 | |
| *** aojea has quit IRC | 23:16 | |
| *** yamamoto has quit IRC | 23:18 | |
| *** aojea has joined #openstack-kuryr | 23:29 | |
| *** hongbin has quit IRC | 23:30 | |
| *** aojea has quit IRC | 23:34 | |
| *** aojea has joined #openstack-kuryr | 23:38 | |
| *** aojea has quit IRC | 23:43 | |
| *** aojea has joined #openstack-kuryr | 23:47 | |
| *** aojea has quit IRC | 23:52 | |
| *** tonanhngo has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!