| *** tonanhngo has joined #openstack-kuryr | 00:19 | |
| *** tonanhngo has quit IRC | 00:22 | |
| *** rubabsyed has joined #openstack-kuryr | 00:30 | |
| *** tonanhngo has joined #openstack-kuryr | 00:51 | |
| *** tonanhngo has quit IRC | 00:55 | |
| *** hongbin has quit IRC | 01:08 | |
| openstackgerrit | Frederick F. Kautz IV proposed openstack/kuryr-kubernetes master: Add *.log to gitignore to ignore vagrant logs https://review.openstack.org/448322 | 01:08 |
|---|---|---|
| *** yedongcan has joined #openstack-kuryr | 01:08 | |
| *** svinota has quit IRC | 01:19 | |
| *** tonanhngo has joined #openstack-kuryr | 01:25 | |
| *** tonanhngo has quit IRC | 01:26 | |
| *** tonanhngo has joined #openstack-kuryr | 01:57 | |
| *** tonanhngo has quit IRC | 01:58 | |
| *** hlo323 has joined #openstack-kuryr | 02:28 | |
| *** tonanhngo has joined #openstack-kuryr | 02:28 | |
| *** tonanhngo has quit IRC | 02:31 | |
| *** hongbin has joined #openstack-kuryr | 02:42 | |
| *** yuanying has quit IRC | 03:05 | |
| *** hongbin has quit IRC | 03:13 | |
| *** hongbin has joined #openstack-kuryr | 03:14 | |
| *** limao has joined #openstack-kuryr | 03:16 | |
| *** hongbin has quit IRC | 03:50 | |
| *** yedongcan has quit IRC | 03:50 | |
| *** yuanying has joined #openstack-kuryr | 04:18 | |
| *** hongbin has joined #openstack-kuryr | 04:58 | |
| *** hongbin has quit IRC | 05:08 | |
| *** xdreamer has joined #openstack-kuryr | 05:55 | |
| *** xdreamer has left #openstack-kuryr | 05:55 | |
| *** vikasc has quit IRC | 06:04 | |
| *** vikasc has joined #openstack-kuryr | 06:10 | |
| *** yedongcan has joined #openstack-kuryr | 06:16 | |
| *** ltomasbo|away is now known as ltomasbo | 06:17 | |
| *** vikasc has quit IRC | 06:24 | |
| *** vikasc has joined #openstack-kuryr | 06:27 | |
| *** svinota has joined #openstack-kuryr | 06:37 | |
| *** aojea has joined #openstack-kuryr | 06:42 | |
| *** aojea has quit IRC | 06:56 | |
| *** kzaitsev_ws has joined #openstack-kuryr | 07:23 | |
| *** pcaruana has joined #openstack-kuryr | 07:39 | |
| *** irenab_ has joined #openstack-kuryr | 08:00 | |
| apuimedo | limao: please review https://review.openstack.org/#/c/431966/ | 08:06 |
| *** aojea has joined #openstack-kuryr | 08:16 | |
| *** aojea has quit IRC | 08:16 | |
| *** aojea has joined #openstack-kuryr | 08:17 | |
| openstackgerrit | Merged openstack/kuryr-libnetwork master: Updated from global requirements https://review.openstack.org/431966 | 08:20 |
| openstackgerrit | Ruijing proposed openstack/kuryr-kubernetes master: update rc files https://review.openstack.org/447771 | 08:21 |
| *** irenab_ has quit IRC | 08:21 | |
| *** svinota has quit IRC | 08:32 | |
| apuimedo | yedongcan: I was thinking that maybe we should have a blueprint for having kuryr-libnetwork devstack use the swarm mode instead of the clustered mode with ETCD | 08:36 |
| *** egonzalez has joined #openstack-kuryr | 08:43 | |
| yedongcan | apuimedo: okay, sounds good. | 09:03 |
| apuimedo | I'll add the blueprint then | 09:04 |
| yedongcan | apuimedo: what's key/value store we use? | 09:04 |
| apuimedo | etcd | 09:05 |
| yedongcan | I mean in swarm, it seems support Consul, ZooKeeper and etcd | 09:08 |
| apuimedo | swarm mode now uses its own | 09:08 |
| *** mchiappe|timeoff is now known as mchiappero | 09:08 | |
| yedongcan | ok. | 09:11 |
| *** alraddarla has quit IRC | 09:13 | |
| irenab | apuimedo: regarding libnetwork features for pike, what about QoS stuff | 09:28 |
| yedongcan | irenab, apuimedo: +1 for QoS | 09:30 |
| *** garyloug has joined #openstack-kuryr | 09:30 | |
| *** yuanying has quit IRC | 09:31 | |
| dmellado | +1 | 09:34 |
| dmellado | apuimedo: so, then, do I owe you something? | 09:34 |
| irenab | apuimedo: do we have some etherpad for libnetwork items for pike or shall it go to the trello and then launchpad? | 09:40 |
| limao | apuimedo: yedongcan: I tested kuryr with swarn mode, looks like it do not support kuryr works in plugin v1, it should works plugin v2 mode(I have not tested but I see comments from mavenugo in https://github.com/docker/docker/issues/24569). | 10:02 |
| *** aojea_ has joined #openstack-kuryr | 10:05 | |
| limao | apuimedo: yedongcan: and I filed a bug before https://bugs.launchpad.net/kuryr-libnetwork/+bug/1668486 , if we use swarm mode , we may need to let kuryr-libnetwork works as docker-libnetwork plugin v2 first . | 10:06 |
| openstack | Launchpad bug 1668486 in kuryr-libnetwork "We can only create kuryr local network in docker swarm 1.13" [Undecided,New] - Assigned to Liping Mao (limao) | 10:06 |
| *** aojea has quit IRC | 10:08 | |
| *** yuanying has joined #openstack-kuryr | 10:11 | |
| *** limao has quit IRC | 10:16 | |
| irenab | limao: this looks quite high priority issue | 10:16 |
| *** alraddarla has joined #openstack-kuryr | 10:20 | |
| apuimedo | dmellado: yes you do | 10:25 |
| apuimedo | cause you had all the [kubernetes] section wrong man | 10:25 |
| apuimedo | xD | 10:25 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: doc: Add reminder to have https in api_root https://review.openstack.org/448490 | 10:28 |
| apuimedo | dmellado: ltomasbo: this patch is for you | 10:28 |
| apuimedo | irenab: please review | 10:28 |
| yedongcan | limao: ok, see it. | 10:29 |
| dmellado | apuimedo: I guess one action item could be to *dramatically* improve docs xD | 10:29 |
| irenab | apuimedo: commit messagecontains a lot of pain in its voice :-) | 10:30 |
| dmellado | irenab: sadly yes xD | 10:31 |
| ltomasbo | apuimedo, thanks! that would have saved us quite some time... | 10:32 |
| irenab | tiny nit, otherwise LGTM | 10:32 |
| apuimedo | irenab: indeed | 10:47 |
| apuimedo | irenab: link it how? | 10:48 |
| *** yedongcan has left #openstack-kuryr | 10:51 | |
| kzaitsev_ws | I was wondering if you folks plan on adding stable/* branches at some point in future | 11:18 |
| kzaitsev_ws | I know, that kuryr(& kuryr-k8s too I suppose) is independent | 11:20 |
| kzaitsev_ws | but the main question as usual would be managing dependencies; you seem to have requirements bot configured, so it's just a matter of time when master stops being compatible with, say stable/ocata requirements-wise | 11:22 |
| apuimedo | kzaitsev_ws: kuryr-k8s will most likely be release-model: cycle-with-intermediary | 11:23 |
| apuimedo | and have stable/ocata | 11:23 |
| apuimedo | kuryr-lib probably too | 11:23 |
| kzaitsev_ws | apuimedo: oh cool! | 11:23 |
| apuimedo | kuryr-libnetwork is up for discussion | 11:23 |
| kzaitsev_ws | just a friendly reminde, that you only have time untill p1 milestone to change/declare your release model | 11:23 |
| kzaitsev_ws | *reminder | 11:24 |
| apuimedo | kzaitsev_ws: do you have a source for that? | 11:27 |
| * apuimedo is not big on bureaucracy and tends to miss these things | 11:28 | |
| kzaitsev_ws | =) | 11:28 |
| kzaitsev_ws | I've messed it the same way 2 cycles ago, when we wnated to switch murano from milestones to intermediary | 11:29 |
| *** neiljerram has joined #openstack-kuryr | 11:29 | |
| kzaitsev_ws | I think it was in one of the latest letters from release team, let me dig it up | 11:29 |
| apuimedo | thanks kzaitsev_ws ! | 11:29 |
| kzaitsev_ws | apuimedo: http://lists.openstack.org/pipermail/openstack-dev/2017-March/114227.html | 11:30 |
| apuimedo | very well | 11:31 |
| apuimedo | kzaitsev_ws: thanks a lot | 11:31 |
| apuimedo | ! | 11:31 |
| apuimedo | I'll send a patch | 11:32 |
| kzaitsev_ws | and p-1 is https://releases.openstack.org/pike/schedule.html Apr 10 - Apr 14 | 11:32 |
| kzaitsev_ws | so there's some time yet =) | 11:32 |
| apuimedo | better send it now | 11:33 |
| apuimedo | before I forget | 11:33 |
| apuimedo | :-) | 11:33 |
| kzaitsev_ws | I've spent a lot of time on %company_name%'s bureaucracy, but managed to setup devstack with kuryr-k8s in the midst of it. | 11:35 |
| kzaitsev_ws | os_vif.plug #1673832 hit me, but the workaround did it ) | 11:35 |
| kzaitsev_ws | now kubelet on my devstack complains, that it can't find a bridge for my pod =( | 11:37 |
| apuimedo | mmm | 11:38 |
| kzaitsev_ws | I guess thats because minikube configures itself with bridge cni | 11:38 |
| apuimedo | ah right | 11:38 |
| apuimedo | if you use minikube, it may be pointing to the bridge cni | 11:38 |
| kzaitsev_ws | so I now have 2 nodes, one with bridge and other with kuryr | 11:39 |
| kzaitsev_ws | yeah. is there a workaround for that (i.e. maybe I can tell k8s to only use specific CNI for specifically labeled pods) | 11:39 |
| apuimedo | kzaitsev_ws: you want an environment with both cni plugins? | 11:40 |
| kzaitsev_ws | apuimedo: btw I hope you're ok that I'm asking seemingly k8s-only questions here =) | 11:41 |
| kzaitsev_ws | apuimedo: well I want to know my options, tbh | 11:41 |
| apuimedo | kzaitsev_ws: more than fine | 11:41 |
| apuimedo | I like a busy channel | 11:42 |
| apuimedo | kzaitsev_ws: I've never tried to use both in the same environment | 11:42 |
| apuimedo | s/environment/cluster | 11:42 |
| apuimedo | / | 11:42 |
| kzaitsev_ws | so the default setup would be to have single cni driver on all the nodes then. | 11:43 |
| apuimedo | I suppose you'd have to have some subclass of the kuryr-k8s drivers/handlers that knows to ignore pods with some metadata | 11:43 |
| apuimedo | kzaitsev_ws: right, that's the base scenario we have | 11:43 |
| kzaitsev_ws | is it a kuryr-k8s limitation or k8s's own? | 11:44 |
| kzaitsev_ws | I mean I haven't digged k8s docs deep enough on the topic yet ) | 11:45 |
| apuimedo | kzaitsev_ws: which limitation? | 11:51 |
| kzaitsev_ws | to have all nodes in the cluster with the same CNI | 11:51 |
| apuimedo | no. That's just what we target for now | 11:55 |
| apuimedo | nobody has blueprinted multiple cni clusters | 11:55 |
| apuimedo | but it is something we'd like to consider | 11:55 |
| apuimedo | basically you need the vif handler to inhibit itself from some pods | 11:56 |
| apuimedo | and/or a scheduler extension probably | 11:56 |
| apuimedo | to make it easier | 11:56 |
| irenab | kzaitsev_ws: waht is your use case? | 11:58 |
| kzaitsev_ws | apuimedo: great. I really need to dig deeper into kuryr-k8s structure to understand how complex that kind of thing would be | 12:02 |
| dmellado | kzaitsev_ws: you're not alone on that dig xD | 12:04 |
| kzaitsev_ws | irenab: well currently it's mostly a POC & research of what I can do with k8s & openstack network-wise, so naturally I came to kuryr-kubernetes | 12:04 |
| kzaitsev_ws | and tried to wire it up with minikube =) | 12:05 |
| openstackgerrit | Marco Chiappero proposed openstack/kuryr-kubernetes master: [WIP] Add MACVLAN based interfaces for nested containers https://review.openstack.org/440669 | 12:06 |
| kzaitsev_ws | and multiple cni was the first (or maybe 2d =)) thing I've hit =) | 12:06 |
| kzaitsev_ws | probably would have to deloy k8s a bit differently then | 12:08 |
| irenab | kzaitsev_ws: nice. Multiple CNI is on the team todo list, but more for the multi-homed Pods case | 12:08 |
| irenab | your case sounds like some sort of different network segments | 12:08 |
| kzaitsev_ws | doesn't it bring a chicken-egg problem though? I mean; to install k8s I would need kuryr-k8s and kuryr cni, but kuryr-k8s needs access to k8s-api. Am I missing smth? ) | 12:10 |
| kzaitsev_ws | I need to dig how devstack plugin installs hypercube | 12:12 |
| irenab | kzaitsev_ws: with devstack, k8s is deployed and setup to work with kuryr CNI while kuryr-k8s Controller is deployed as well and setup to access k8s-api | 12:29 |
| *** limao has joined #openstack-kuryr | 12:34 | |
| *** garyloug has quit IRC | 12:41 | |
| *** garyloug has joined #openstack-kuryr | 12:46 | |
| *** limao has quit IRC | 12:56 | |
| *** limao has joined #openstack-kuryr | 13:07 | |
| *** limao has quit IRC | 13:18 | |
| *** limao has joined #openstack-kuryr | 13:18 | |
| *** limao has quit IRC | 13:29 | |
| *** limao has joined #openstack-kuryr | 13:30 | |
| *** janki has joined #openstack-kuryr | 13:31 | |
| *** jchhatbar has joined #openstack-kuryr | 13:34 | |
| *** janki has quit IRC | 13:38 | |
| *** limao has quit IRC | 13:41 | |
| *** limao has joined #openstack-kuryr | 13:41 | |
| *** jchhatbar has quit IRC | 13:47 | |
| *** janki has joined #openstack-kuryr | 13:48 | |
| *** limao has quit IRC | 13:52 | |
| *** limao has joined #openstack-kuryr | 13:52 | |
| *** limao has quit IRC | 14:03 | |
| *** limao has joined #openstack-kuryr | 14:04 | |
| *** limao has quit IRC | 14:15 | |
| *** limao has joined #openstack-kuryr | 14:15 | |
| kzaitsev_ws | hm. I'm still getting the error even without using external k8s-api, just what devstack provides | 14:25 |
| kzaitsev_ws | here's the log from kubelet http://paste.openstack.org/show/603770/ | 14:25 |
| *** limao has quit IRC | 14:26 | |
| *** limao has joined #openstack-kuryr | 14:26 | |
| kzaitsev_ws | I'm just running the thing with 'kubectl run mysql --image mysql' from devstack itself | 14:26 |
| *** limao has quit IRC | 14:37 | |
| *** limao has joined #openstack-kuryr | 14:38 | |
| *** janki has quit IRC | 14:41 | |
| kzaitsev_ws | indeed there is no qbr2614d46d-78 in the ipdb, but there is a 'tap2614d46d-78'. not sure if those two are connected, but they look so | 14:45 |
| kzaitsev_ws | as per https://github.com/openstack/kuryr-kubernetes/blob/master/kuryr_kubernetes/os_vif_util.py#L191 | 14:46 |
| kzaitsev_ws | so I wonder if this has recently changed in neutron | 14:47 |
| *** hongbin has joined #openstack-kuryr | 14:47 | |
| *** limao has quit IRC | 14:48 | |
| *** limao has joined #openstack-kuryr | 14:49 | |
| *** limao has quit IRC | 15:00 | |
| *** limao has joined #openstack-kuryr | 15:00 | |
| *** svinota has joined #openstack-kuryr | 15:02 | |
| *** hongbin_ has joined #openstack-kuryr | 15:03 | |
| *** hongbin has quit IRC | 15:05 | |
| *** limao has quit IRC | 15:11 | |
| *** limao has joined #openstack-kuryr | 15:11 | |
| *** limao has quit IRC | 15:23 | |
| *** limao has joined #openstack-kuryr | 15:23 | |
| apuimedo | kzaitsev_ws: brctl show | 15:32 |
| apuimedo | kzaitsev_ws: oh, check if neutron is configured with ovs firewall | 15:33 |
| *** limao has quit IRC | 15:34 | |
| *** limao has joined #openstack-kuryr | 15:35 | |
| *** limao has quit IRC | 15:45 | |
| *** limao has joined #openstack-kuryr | 15:46 | |
| *** limao has quit IRC | 15:56 | |
| *** limao has joined #openstack-kuryr | 15:57 | |
| *** limao has quit IRC | 16:08 | |
| *** limao has joined #openstack-kuryr | 16:08 | |
| kzaitsev_ws | apuimedo: probably you're right, I was fiddling with the firewall in neutron | 16:14 |
| *** gsagie has joined #openstack-kuryr | 16:18 | |
| *** limao has quit IRC | 16:19 | |
| *** hongbin_ has quit IRC | 16:22 | |
| *** svinota has quit IRC | 16:42 | |
| dmellado | again? that privsep error? | 16:55 |
| *** egonzalez has quit IRC | 17:01 | |
| apuimedo | dmellado: no, he already solved that | 17:04 |
| kzaitsev_ws | oh, btw I decided to go with a clean devstack installation =) | 17:07 |
| apuimedo | :-) | 17:07 |
| kzaitsev_ws | I've decided, that I fiddled too much with stack/unstack | 17:07 |
| kzaitsev_ws | meanwhile I have some time to familiarise myself with the code | 17:08 |
| kzaitsev_ws | btw. is local.conf.sample, that you have in the devstack folder enough to get me running? | 17:09 |
| kzaitsev_ws | nah, didn't help. guess default devstack neutron config is not enough | 17:15 |
| kzaitsev_ws | apuimedo: what should I set in firewall_driver then? | 17:18 |
| apuimedo | depends if you want hybrid networking (ovs + linux bridge for sgs) or ovs native | 17:18 |
| apuimedo | gotta go for a while | 17:19 |
| dmellado | kzaitsev_ws: that's always a wise choice | 17:21 |
| dmellado | xd | 17:21 |
| dmellado | kzaitsev_ws: if you haven't done that before, I guess it's time to write your own 'deploy devstack script' | 17:22 |
| dmellado | while destroying/recreating the vm ;) | 17:22 |
| kzaitsev_ws | dmellado: well I don't really care at the moment (and I expect the vm to be torn donw/recreated multiple times) | 17:28 |
| kzaitsev_ws | dmellado: so what would you suggest me to start with? =) | 17:29 |
| dmellado | kzaitsev_ws: well, I do usually user a vagrant/libvirt + ansible to recreate my whole devstack env | 17:29 |
| dmellado | but up to you ;) | 17:29 |
| kzaitsev_ws | oh, you meant the devstack thing. yeah I have a snapshot to rollback to. hope it'll be enough for a while ) | 17:31 |
| kzaitsev_ws | dmellado: can you help me out with the ovs firewall option? | 17:32 |
| kzaitsev_ws | or maybe does http://paste.openstack.org/show/603770/ look familiar to you? | 17:33 |
| dmellado | kzaitsev_ws: you mean this part? | 17:33 |
| dmellado | let me check | 17:33 |
| dmellado | kzaitsev_ws: are you using ovs firewall? | 17:34 |
| kzaitsev_ws | well, no. I'm just trying to deploy a default devstack installation with kuryr-k8s, and it sets 'iptables_hybrid' | 17:37 |
| kzaitsev_ws | ugh, I need to setup notifications for irssi, to get pings faster about mentions | 17:38 |
| *** aojea_ has quit IRC | 17:46 | |
| *** tonanhngo has joined #openstack-kuryr | 17:49 | |
| *** tonanhngo has quit IRC | 17:54 | |
| *** tonanhngo has joined #openstack-kuryr | 17:54 | |
| kzaitsev_ws | ok, seem that it worked ) | 17:56 |
| kzaitsev_ws | I mean I just changed firewall option to openvswitch and restarted all things neutron %) | 17:56 |
| *** ltomasbo is now known as ltomasbo|away | 17:59 | |
| *** kzaitsev_ws has quit IRC | 18:04 | |
| *** garyloug has quit IRC | 18:14 | |
| *** ltomasbo|away is now known as ltomasbo | 18:26 | |
| *** svinota has joined #openstack-kuryr | 18:40 | |
| *** ltomasbo is now known as ltomasbo|away | 18:55 | |
| *** gsagie has quit IRC | 19:06 | |
| *** hongbin has joined #openstack-kuryr | 19:54 | |
| *** svinota has quit IRC | 20:49 | |
| *** svinota has joined #openstack-kuryr | 21:05 | |
| *** pcaruana has quit IRC | 21:47 | |
| *** irenab has left #openstack-kuryr | 21:51 | |
| *** pmannidi has quit IRC | 22:29 | |
| *** pmannidi has joined #openstack-kuryr | 22:51 | |
| *** pmannidi has quit IRC | 22:51 | |
| *** pmannidi has joined #openstack-kuryr | 22:52 | |
| *** pmannidi has quit IRC | 22:52 | |
| *** pmannidi has joined #openstack-kuryr | 22:52 | |
| *** hongbin has quit IRC | 23:19 | |
| *** svinota has quit IRC | 23:46 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!