| *** jayahn has joined #openstack-kuryr | 00:28 | |
| *** jayahn has quit IRC | 00:56 | |
| *** jayahn has joined #openstack-kuryr | 00:58 | |
| *** tonanhngo has joined #openstack-kuryr | 01:00 | |
| *** tonanhngo has quit IRC | 01:01 | |
| *** yedongcan has joined #openstack-kuryr | 01:14 | |
| *** limao has joined #openstack-kuryr | 01:18 | |
| openstackgerrit | Hongbin Lu proposed openstack/kuryr-kubernetes master: [WIP] Spec for fuxi-kubernetes https://review.openstack.org/452554 | 01:26 |
|---|---|---|
| *** jayahn has quit IRC | 01:41 | |
| *** tonanhngo has joined #openstack-kuryr | 01:42 | |
| *** tonanhngo has quit IRC | 01:42 | |
| *** jayahn has joined #openstack-kuryr | 01:44 | |
| *** yamamoto has joined #openstack-kuryr | 01:45 | |
| *** jayahn has quit IRC | 02:06 | |
| *** yamamoto has quit IRC | 02:23 | |
| *** yamamoto has joined #openstack-kuryr | 02:30 | |
| *** tonanhngo has joined #openstack-kuryr | 02:44 | |
| *** shu-mutou-AWAY has joined #openstack-kuryr | 02:44 | |
| *** tonanhngo has quit IRC | 02:44 | |
| *** shu-mutou-AWAY is now known as shu-mutou | 02:44 | |
| *** hongbin has joined #openstack-kuryr | 02:52 | |
| *** yamamoto has quit IRC | 02:56 | |
| *** aojea has joined #openstack-kuryr | 03:01 | |
| *** aojea has quit IRC | 03:05 | |
| *** jayahn has joined #openstack-kuryr | 03:15 | |
| *** Guest32684 has joined #openstack-kuryr | 03:34 | |
| *** yamamoto has joined #openstack-kuryr | 03:45 | |
| *** tonanhngo has joined #openstack-kuryr | 03:45 | |
| *** tonanhngo has quit IRC | 03:46 | |
| *** Guest32684 has quit IRC | 04:09 | |
| *** hongbin has quit IRC | 04:20 | |
| *** shu-mutou has left #openstack-kuryr | 04:37 | |
| *** limao has quit IRC | 04:50 | |
| *** limao has joined #openstack-kuryr | 05:18 | |
| *** kzaitsev_ws has joined #openstack-kuryr | 06:08 | |
| *** janki has joined #openstack-kuryr | 06:17 | |
| *** jayahn has quit IRC | 06:26 | |
| kzaitsev_ws | ivc-: ping. May I ask you a few questions on future design of kuryr-k8s? | 07:21 |
| kzaitsev_ws | I mean you seem to have mentioned previous discussions in your comments =) and I have a couple of questions about them | 07:22 |
| *** vikasc has quit IRC | 08:35 | |
| *** vikasc has joined #openstack-kuryr | 08:37 | |
| *** jchhatbar has joined #openstack-kuryr | 08:38 | |
| *** janki has quit IRC | 08:41 | |
| *** yedongcan has quit IRC | 08:47 | |
| *** egonzalez has joined #openstack-kuryr | 08:47 | |
| *** aojea has joined #openstack-kuryr | 08:48 | |
| *** yedongcan has joined #openstack-kuryr | 08:50 | |
| *** aojea has quit IRC | 08:55 | |
| *** jchhatbar_ has joined #openstack-kuryr | 08:59 | |
| *** jchhatbar has quit IRC | 09:02 | |
| *** aojea has joined #openstack-kuryr | 09:07 | |
| *** kzaitsev_ws has quit IRC | 09:11 | |
| *** kzaitsev_ws has joined #openstack-kuryr | 09:11 | |
| *** limao has quit IRC | 09:25 | |
| *** limao has joined #openstack-kuryr | 09:26 | |
| *** limao has quit IRC | 09:30 | |
| *** aojea has quit IRC | 09:39 | |
| *** aojea has joined #openstack-kuryr | 09:40 | |
| *** aojea has quit IRC | 09:44 | |
| *** jchhatbar has joined #openstack-kuryr | 10:31 | |
| ivc- | kzaitsev_ws pong | 10:31 |
| *** aojea has joined #openstack-kuryr | 10:31 | |
| *** jchhatbar_ has quit IRC | 10:33 | |
| *** jchhatbar_ has joined #openstack-kuryr | 10:34 | |
| *** jchhatbar has quit IRC | 10:36 | |
| *** janki has joined #openstack-kuryr | 10:42 | |
| *** jchhatbar_ has quit IRC | 10:42 | |
| *** jchhatbar has joined #openstack-kuryr | 10:46 | |
| *** janki has quit IRC | 10:48 | |
| kzaitsev_ws | ivc-: oh, sry =) I wanted to ask — what are the plans for KuryrK8sService handler registration? | 11:11 |
| ivc- | kzaitsev_ws soz on a meeting now. will be back in about 30 minutes or so | 11:12 |
| kzaitsev_ws | sure np =) | 11:12 |
| kzaitsev_ws | I'll put my thoughts/questions here meanwhile | 11:13 |
| *** yedongcan has left #openstack-kuryr | 11:14 | |
| kzaitsev_ws | I think I saw you mention, that you already discussed making pipeline handlers configurable somewhere. | 11:14 |
| *** dougbtv_ is now known as dougbtv | 11:16 | |
| *** aojea has quit IRC | 11:18 | |
| *** aojea has joined #openstack-kuryr | 11:18 | |
| kzaitsev_ws | I want to allow Pods to request multiple additional ports from neutron. | 11:20 |
| kzaitsev_ws | So I was thinking about writing a separate handler to kuryr-k8s and then a separate cni (and use kuryr-cni and my onw via multus-cni) | 11:21 |
| *** aojea has quit IRC | 11:23 | |
| kzaitsev_ws | The CNI in turn would pass interfaces from host to Pods namespace. much like sriov-cni does =) | 11:24 |
| *** reedip is now known as reedip_afk | 11:26 | |
| kzaitsev_ws | I'm afraid this kind of change would be hard to upstream though. | 11:26 |
| kzaitsev_ws | I can imagine adding requests for multiple IPs (and possibly requesting specific IP at Pod creation time) added to kuryr-k8s | 11:27 |
| kzaitsev_ws | but interface passthrough feels out of scope of kuryr-cni | 11:28 |
| kzaitsev_ws | so I wanted to ask for opinions | 11:29 |
| kzaitsev_ws | if we make handlers configurable/dynamic (through stevedore for example) it would enable makeing rather custom configurations. And I would probably be able to share mine somewhere under /contrib | 11:30 |
| kzaitsev_ws | ivc-: would love to hear your thoughts on ^^^ =) | 11:31 |
| dmellado | kzaitsev_ws: that looks like an interesting idea that con be drafted on a spec ;) | 11:31 |
| *** jchhatbar has quit IRC | 11:32 | |
| *** jchhatbar has joined #openstack-kuryr | 11:33 | |
| kzaitsev_ws | dmellado: yeah =) but I'd like to get some preliminary opinions on design ) | 11:39 |
| ivc- | kzaitsev_ws multi-port support for pods has been discussed many times. there are some proposed changes to k8s itself that will add 'network' entity to api | 11:39 |
| dmellado | IIRC we discussed also the idea of having multiple CNIs | 11:39 |
| * dmellado just came back from PTO, so still unsure | 11:39 | |
| dmellado | xD | 11:39 |
| kzaitsev_ws | yep, I've see those, but they're in design stages, right? | 11:40 |
| kzaitsev_ws | so for now I'm thinking of using https://github.com/Intel-Corp/multus-cni | 11:40 |
| kzaitsev_ws | the only downside is that k8s would know nothing about those additional ips on a pod | 11:41 |
| ivc- | kzaitsev_ws also i'd be against using different handlers/cnis to do the same thing (i.e. single vs multi-port). i'd much prefer if the current model is extended to use a list (or dict) of VIFs in place of current single-VIF annotation | 11:41 |
| kzaitsev_ws | Any reason why it's a bad idea to allow multiple handlers? | 11:43 |
| ivc- | why have multiple handlers for the same thing? | 11:43 |
| kzaitsev_ws | in my scenario they would be configuring different interfaces differently | 11:45 |
| ivc- | so how do you expect to compose it with other parts? e.g. the vif pool? | 11:45 |
| ivc- | also are we talking about controller-side handler or a cni-side handler? | 11:47 |
| kzaitsev_ws | actually both. I want to request additional IP from neutron, assign it to a VF of a Network Card and pass this interface to Pods namespace | 11:49 |
| kzaitsev_ws | am I talking nonsence? =) | 11:50 |
| kzaitsev_ws | So I was looking for a way to make it in a non-disruptive manner, so that I would be able to contribute some parts of my solution back to kuryr-k8s | 11:55 |
| *** egonzalez has quit IRC | 11:58 | |
| kzaitsev_ws | I just think, that adding this kind of passthrough capabilities to kuryr-cni would be out of scope of its functionality. | 12:06 |
| kzaitsev_ws | Also I'm intending to use OPI in k8s to bookkeep the amount of free VFs on a node, so it complicates things a little tad further =) | 12:12 |
| kzaitsev_ws | *OIR (don't know why I keep using OPI) | 12:13 |
| *** aojea has joined #openstack-kuryr | 12:54 | |
| kzaitsev_ws | ivc-: also. I think I'll add a dpendant commit with doc-strings added to functions or https://review.openstack.org/#/c/376045 would never get merged ))) | 13:00 |
| *** jayahn has joined #openstack-kuryr | 13:49 | |
| *** limao has joined #openstack-kuryr | 13:51 | |
| *** yedongcan has joined #openstack-kuryr | 13:56 | |
| dmellado | kzaitsev_ws: huh, pls do that and hopefully it'll unlock that | 13:57 |
| dmellado | you've got my +1, in case it helps ;) | 13:57 |
| *** hongbin has joined #openstack-kuryr | 14:03 | |
| *** yedongcan has left #openstack-kuryr | 14:35 | |
| *** aojea has quit IRC | 14:44 | |
| *** aojea has joined #openstack-kuryr | 14:45 | |
| *** aojea has quit IRC | 14:49 | |
| *** jayahn has quit IRC | 14:50 | |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: K8s Services support: LoadBalancerHandler https://review.openstack.org/376045 | 14:52 |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Implement Guru meditation reports https://review.openstack.org/447942 | 14:52 |
| *** jchhatbar has quit IRC | 15:01 | |
| hongbin | kzaitsev_ws: there? | 15:26 |
| hongbin | kzaitsev_ws: left some comments on https://review.openstack.org/#/c/456586/ | 15:26 |
| kzaitsev_ws | sure would address those a bit later today | 15:35 |
| kzaitsev_ws | thanks for pointin that out | 15:36 |
| *** limao has quit IRC | 15:41 | |
| *** limao has joined #openstack-kuryr | 15:42 | |
| *** kzaitsev_ws has quit IRC | 15:44 | |
| *** limao has quit IRC | 15:46 | |
| *** aojea has joined #openstack-kuryr | 16:18 | |
| *** aojea has quit IRC | 16:37 | |
| *** aojea has joined #openstack-kuryr | 16:38 | |
| *** aojea has quit IRC | 16:42 | |
| openstackgerrit | Merged openstack/kuryr-libnetwork master: Remove subunit-trace fork https://review.openstack.org/454295 | 16:47 |
| openstackgerrit | Merged openstack/kuryr master: Replace http with https https://review.openstack.org/454430 | 16:47 |
| openstackgerrit | Merged openstack/kuryr master: Remove log translations https://review.openstack.org/447835 | 16:56 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-kubernetes master: Updated from global requirements https://review.openstack.org/455931 | 17:14 |
| *** tonanhngo has joined #openstack-kuryr | 17:22 | |
| *** tonanhngo has quit IRC | 17:26 | |
| *** tonanhngo has joined #openstack-kuryr | 17:27 | |
| *** vikasc has quit IRC | 17:30 | |
| *** vikasc has joined #openstack-kuryr | 17:45 | |
| *** alraddarla_ has joined #openstack-kuryr | 18:07 | |
| *** alraddarla_ has quit IRC | 18:12 | |
| *** vikasc has quit IRC | 18:37 | |
| *** vikasc has joined #openstack-kuryr | 18:40 | |
| *** aojea has joined #openstack-kuryr | 19:10 | |
| openstackgerrit | Hongbin Lu proposed openstack/kuryr-libnetwork master: [WIP] Avoid running more than one etcd process https://review.openstack.org/457361 | 21:23 |
| openstackgerrit | Hongbin Lu proposed openstack/fuxi master: [WIP] Avoid running more than one etcd process https://review.openstack.org/457363 | 21:25 |
| *** aojea has quit IRC | 21:37 | |
| *** alraddarla_ has joined #openstack-kuryr | 21:45 | |
| openstackgerrit | Hongbin Lu proposed openstack/fuxi master: [WIP] Avoid running more than one etcd process https://review.openstack.org/457363 | 21:49 |
| openstackgerrit | Hongbin Lu proposed openstack/kuryr-libnetwork master: [WIP] Avoid running more than one etcd process https://review.openstack.org/457361 | 21:50 |
| *** alraddarla_ has quit IRC | 21:51 | |
| *** pmannidi has joined #openstack-kuryr | 23:19 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!