| *** hongbin has joined #openstack-kuryr | 02:49 | |
| *** hongbin has quit IRC | 03:41 | |
| *** janki has joined #openstack-kuryr | 03:57 | |
| *** irenab has joined #openstack-kuryr | 04:44 | |
| *** pcaruana has joined #openstack-kuryr | 06:38 | |
| *** pliu has joined #openstack-kuryr | 08:03 | |
| *** dulek has joined #openstack-kuryr | 08:19 | |
| *** aojea has joined #openstack-kuryr | 08:22 | |
| *** janki has quit IRC | 08:36 | |
| *** celebdor has joined #openstack-kuryr | 08:38 | |
| *** garyloug has joined #openstack-kuryr | 08:48 | |
| *** aojea has quit IRC | 08:54 | |
| *** janki has joined #openstack-kuryr | 09:15 | |
| *** neiljerram has joined #openstack-kuryr | 09:55 | |
| dulek | ltomasbo: Okay, for the first issue (race condition) - if you know what's causing the race, just check that condition and raise ResourceNotReady in the handler. | 10:03 |
|---|---|---|
| dulek | ltomasbo: Not sure really if that's solvable by this, but maybe? | 10:03 |
| ltomasbo | dulek, for the first issue I think I have an idea, yes | 10:03 |
| dulek | ltomasbo: In the Octavia SGs case - it's kuryr-controller raising the exceptions? | 10:04 |
| ltomasbo | I think it is because we are cleaning up pools, but the ports in used are left as they are being removed at the same time | 10:04 |
| ltomasbo | dulek, yes, the second one I'm more worried about | 10:04 |
| ltomasbo | dulek, and yes, I see an exception about security group XXX (which is the octavia generated one lb-xxxx) not found | 10:04 |
| ltomasbo | and it is not found because it belongs to admin instead of demo | 10:05 |
| ltomasbo | I wonder what is the differences, if I deploy it with devstack (baremetal) it works and it has no problems to update that sg | 10:05 |
| ltomasbo | but deploying nested with openshift-ansible raises that | 10:06 |
| celebdor | dulek: ltomasbo: dmellado: I'm considering changing how we do tempest access to the API | 10:08 |
| celebdor | and have two service accounts | 10:08 |
| dulek | celebdor: Which API? K8s? | 10:08 |
| celebdor | dulek: both | 10:08 |
| celebdor | k8s and OCP | 10:08 |
| celebdor | sorry, k8s and okd | 10:08 |
| celebdor | :P | 10:08 |
| dulek | celebdor: :) | 10:09 |
| celebdor | create a tempest-admin serviceaccount | 10:09 |
| dulek | ltomasbo: Are we able to give Kuryr's user access to Octavia's project? | 10:09 |
| dulek | ltomasbo: We should be. | 10:09 |
| dulek | ltomasbo: We may simply require it. | 10:09 |
| dulek | celebdor: Sure, sounds okay, but why? | 10:10 |
| celebdor | dulek: cause I'm very tempted to end up removing the openshift client from requirements | 10:11 |
| celebdor | (and possibly the k8s one) | 10:11 |
| celebdor | and just use requests | 10:11 |
| celebdor | and also, then we'll have tempest create service accounts for namespace specific tests | 10:11 |
| dulek | celebdor: Well, one of Tempest's tenets is to query APIs directly… | 10:13 |
| celebdor | dulek: exactly | 10:14 |
| celebdor | it bothers me so, so much | 10:14 |
| celebdor | that we use those clients | 10:14 |
| ltomasbo | dmellado, celebdor: this is finally green! https://review.openstack.org/#/c/589841 | 10:29 |
| ltomasbo | dulek, and this one https://review.openstack.org/#/c/580680/ seems to have time_out in a couple of them | 10:30 |
| ltomasbo | going to recheck experimental | 10:30 |
| celebdor | ltomasbo: you have a -1 from dmellado there | 10:33 |
| celebdor | :-) | 10:33 |
| ltomasbo | celebdor, because it was failing before | 10:33 |
| ltomasbo | it is passing now | 10:33 |
| ltomasbo | it was a gate flakiness | 10:33 |
| ltomasbo | dmellado, remove the -1 here! https://review.openstack.org/#/c/589841/4 xD | 10:34 |
| dmellado | lemme see | 10:34 |
| dmellado | maybe you deserve it | 10:34 |
| dmellado | ltomasbo: is it better now? xD | 10:35 |
| ltomasbo | no | 10:35 |
| ltomasbo | dmellado, why -2? | 10:35 |
| dmellado | ltomasbo: I was just teasing you xD | 10:35 |
| ltomasbo | xD | 10:35 |
| celebdor | ltomasbo: you shouldn't have voted him | 10:35 |
| dmellado | because you still didn't get to sing | 10:35 |
| dmellado | celebdor: how about you? | 10:35 |
| celebdor | how about what? | 10:35 |
| dmellado | will we be getting some catalonian song? | 10:36 |
| dmellado | just like tzumainn did? | 10:36 |
| dmellado | I'm looking forward to it | 10:36 |
| dmellado | xD | 10:36 |
| ltomasbo | celebdor, I didn't vote him :D | 10:36 |
| ltomasbo | (I was on PTO...) | 10:36 |
| celebdor | dmellado: when/where? | 10:36 |
| celebdor | ltomasbo: even worse, abstencionista! | 10:36 |
| dmellado | celebdor: don't tell me you missed the TMNT song by mainn? | 10:36 |
| celebdor | going to the beach instead of doing your civic duty | 10:36 |
| celebdor | dmellado: I didn't | 10:36 |
| celebdor | I only sing in public in karaoke booths | 10:37 |
| ltomasbo | lol | 10:37 |
| celebdor | and when walking up and down the street making my daughter sleep | 10:37 |
| dmellado | since then I thought I'd only merge ltomasbo patches if he gets to sing hakuna matata | 10:37 |
| dmellado | xD | 10:37 |
| ltomasbo | xD | 10:37 |
| celebdor | dmellado: he should sing something more traditional | 10:38 |
| celebdor | like "Los fruitis" | 10:38 |
| dmellado | ltomasbo: do you know that one? | 10:38 |
| celebdor | dmellado: I have a question regarding kuryr-tempest-plugin | 10:38 |
| dmellado | shoot | 10:39 |
| ltomasbo | I know that one! | 10:39 |
| celebdor | which user does it run as in the gates | 10:39 |
| celebdor | ? | 10:39 |
| ltomasbo | I'll practise it next week! | 10:39 |
| celebdor | xD | 10:39 |
| dmellado | IIRC, tempest user | 10:39 |
| celebdor | oh, right | 10:39 |
| celebdor | and you are copying the kubeconfig to it | 10:39 |
| celebdor | aren't you? | 10:39 |
| dmellado | yep | 10:39 |
| celebdor | cause I'm thinking about getting rid of the clients | 10:40 |
| celebdor | and I was thinking which is the best way to pass the bearer token | 10:40 |
| dmellado | celebdor: -2 | 10:40 |
| dmellado | there's a lot of code already written over the clients | 10:40 |
| dmellado | if you want to do that | 10:40 |
| dmellado | I'll force you to rewrite all existing code | 10:40 |
| celebdor | dmellado: 34 lines | 10:40 |
| celebdor | git grep k8s_client\. | wc -l | 10:40 |
| dmellado | celebdor: also, the k8s client will be part of kuryr | 10:41 |
| dmellado | as it'll be needed for the network policies | 10:41 |
| dmellado | basically to unwrap and parse the spec | 10:41 |
| dmellado | but anyways, feel free to propose the patch and we'll get to see it | 10:42 |
| celebdor | dmellado: wait, what? | 10:42 |
| celebdor | why would it ever be necessary? | 10:42 |
| dmellado | it's just convenient | 10:42 |
| celebdor | man... | 10:42 |
| celebdor | that does not seem like a good reason at all | 10:43 |
| dmellado | lol | 10:43 |
| dmellado | that was part of the conversations I had with eyal before he left | 10:43 |
| dmellado | and it makes things just easier | 10:43 |
| dmellado | I'll think about that after today's meeting about python-kubernetes | 10:43 |
| dmellado | oh, and don't forget we have python-openshift as well | 10:43 |
| celebdor | like what is easier? | 10:43 |
| dmellado | spec and ingress/egress parsing | 10:44 |
| dmellado | in any case I won't ue the client part for that, so no worries | 10:44 |
| ltomasbo | dulek, celebdor: there is one thing I don't quite get | 10:58 |
| celebdor | ltomasbo: which? | 10:58 |
| ltomasbo | dulek, celebdor: I have 2 devstack deployments | 10:58 |
| openstackgerrit | Genadi Chereshnya proposed openstack/kuryr-tempest-plugin master: Testing VM to loadBalancer service connectivity https://review.openstack.org/571968 | 10:59 |
| ltomasbo | and in both cases the loadbalancer belongs to the admin | 10:59 |
| ltomasbo | why in one it does complains about it, and not in the other? | 10:59 |
| ltomasbo | ohh, wait, what kind of users is k8s? is admin? | 10:59 |
| dulek | ltomasbo: Magic? :P | 11:00 |
| ltomasbo | the difference is that I'm using demo tenant in the one I'm running with openshift-ansible... | 11:00 |
| celebdor | the loadbalancer should not belong to the admin project | 11:00 |
| celebdor | didn't I make a patch fixing all the resources to belong to the right project? | 11:01 |
| ltomasbo | celebdor, sorry, I meant loadbalancer security group | 11:01 |
| openstackgerrit | Alexey Perevalov proposed openstack/kuryr-kubernetes master: Add SR-IOV capabilities to VIF handler https://review.openstack.org/512280 | 11:13 |
| openstackgerrit | Alexey Perevalov proposed openstack/kuryr-kubernetes master: Add SR-IOV binding driver to CNI https://review.openstack.org/512281 | 11:13 |
| openstackgerrit | Alexey Perevalov proposed openstack/kuryr-kubernetes master: Separate helper function for pod https://review.openstack.org/590245 | 11:13 |
| celebdor | ltomasbo: can I get more context? | 11:14 |
| ltomasbo | I think the problem is we use k8s project for devstack | 11:16 |
| ltomasbo | while I used demo project for openshift-ansible | 11:16 |
| ltomasbo | celebdor, ^^ | 11:16 |
| ltomasbo | so, I see demo has the next roles | 11:17 |
| ltomasbo | | member | demo@Default | | demo@Default | | | False | | 11:17 |
| ltomasbo | | load-balancer_member | demo@Default | | demo@Default | | | False | | 11:17 |
| ltomasbo | | anotherrole | demo@Default | | demo@Default | | | False | | 11:17 |
| ltomasbo | while k8s project does not have any | 11:17 |
| ltomasbo | and kuryr user has service and admin... | 11:17 |
| ltomasbo | that could explain why it is enabled, but we should use a project_id similar to the demo one, right? | 11:18 |
| celebdor | ltomasbo: I don't really know about that | 11:18 |
| celebdor | meaning, I don't know how real the demo project and user are | 11:19 |
| ltomasbo | I just looked for the differences | 11:19 |
| celebdor | but if they are more, let's make the kuryr user like the demo one | 11:19 |
| ltomasbo | to see why one can, and the other cannot | 11:19 |
| celebdor | ltomasbo: I still don't know what's being blocked | 11:19 |
| ltomasbo | I see openstack role assignment list --project k8s is empty | 11:19 |
| ltomasbo | ahh, ok, for services isolation with octavia, we need to modify the rules of the lbaas sg created for the amphora | 11:20 |
| ltomasbo | celebdor, ^^ | 11:20 |
| *** aojea has joined #openstack-kuryr | 11:20 | |
| ltomasbo | otherwise it will open port 80 (for example) from everywhere | 11:20 |
| ltomasbo | and we want to only open it for an specific namespace subnet cidr | 11:21 |
| ltomasbo | so we were modifying those rules when the listeners where created | 11:21 |
| ltomasbo | celebdor, for instance, here: https://github.com/openstack/kuryr-kubernetes/blob/master/kuryr_kubernetes/controller/drivers/lbaasv2.py#L155-L171 | 11:22 |
| celebdor | ltomasbo: mmm | 11:23 |
| celebdor | is there something in octavia that would change this behavior? | 11:24 |
| ltomasbo | not that I'm aware of | 11:24 |
| *** snapiri has quit IRC | 11:25 | |
| ltomasbo | well, that code is doing that, changing that sg rules by these ones | 11:25 |
| celebdor | ok, let's see if I get it | 11:25 |
| *** snapiri has joined #openstack-kuryr | 11:25 | |
| celebdor | on ansible deployments the LB is owned by the project user | 11:25 |
| celebdor | and there we're not able to do this sg change | 11:25 |
| celebdor | but on devstack we are | 11:26 |
| ltomasbo | loadbalancer belongs to default user in both (in one k8s in another one demo) | 11:26 |
| ltomasbo | the lbaas security group is the one belonging to admin | 11:26 |
| ltomasbo | and it can be changed from the tenant we use on devstack for kuryr, but not for demo tenant | 11:27 |
| ltomasbo | the actual error that codes throws is security group not found | 11:28 |
| ltomasbo | going to have lunch... | 11:28 |
| *** aojea has quit IRC | 11:33 | |
| celebdor | mmm | 11:35 |
| celebdor | dulek: regarding your comment to https://review.openstack.org/#/c/578310/18/kuryr_tempest_plugin/tests/scenario/base.py | 11:35 |
| celebdor | did you see that in the official https://github.com/openshift/openshift-restclient-python | 11:35 |
| celebdor | sorry, https://github.com/openshift/openshift-restclient-python#dynamic-client-usage | 11:35 |
| celebdor | they are using it the same way | 11:35 |
| openstackgerrit | Merged openstack/kuryr-tempest-plugin master: Ensure create_namespace function waits for CRD annotation https://review.openstack.org/589841 | 11:36 |
| celebdor | except that he should be using the dynamic support instead of using the deprecared functions | 11:36 |
| *** aojea has joined #openstack-kuryr | 11:45 | |
| *** rh-jelabarre has joined #openstack-kuryr | 11:52 | |
| celebdor | damn | 11:54 |
| celebdor | It'll have to be updated to use that | 11:54 |
| celebdor | :/ | 11:54 |
| *** aojea has quit IRC | 11:58 | |
| dulek | celebdor: I don't have hard feelings about it, just giving ideas. My main issue is that tests are failing on it. | 12:11 |
| celebdor | dulek: I'll take over that patch | 12:13 |
| celebdor | but I really do prefer to keep it yaml | 12:14 |
| celebdor | it makes trying things out in the terminal easier | 12:14 |
| celebdor | and it's not like scenario tests are supposed to be so fast that avoiding yaml.load would matter | 12:14 |
| dulek | celebdor: Sure, keep the yaml! | 12:16 |
| ltomasbo | dulek, celebdor: so, it seems the loadbalancer issue is a real problem and changing the lbaas sg is not supported (unless user belongs to admin project) | 12:28 |
| celebdor | yeah | 12:30 |
| *** garyloug_ has joined #openstack-kuryr | 12:30 | |
| dulek | celebdor, ltomasbo: This seems like Octavia issue to me. Why SGs are in admin instead of Octavia project? | 12:31 |
| ltomasbo | dulek, yep, good question | 12:32 |
| *** garyloug has quit IRC | 12:33 | |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-tempest-plugin master: Add Network Policy tests https://review.openstack.org/589521 | 12:36 |
| *** oanson has joined #openstack-kuryr | 12:42 | |
| *** garyloug_ has quit IRC | 12:46 | |
| *** garyloug_ has joined #openstack-kuryr | 12:46 | |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: Add a new gate for network policies https://review.openstack.org/590304 | 12:55 |
| ltomasbo | dulek, celebdor http://logs.openstack.org/80/580680/17/experimental/kuryr-kubernetes-tempest-daemon-containerized-openshift-octavia-serial/de27183/controller/logs/kubernetes/pod_logs/kube-system-kuryr-controller-7db7c944fb-2wnnp.txt.gz | 13:03 |
| ltomasbo | still not clean on the openshift-gates... | 13:03 |
| ltomasbo | seems some auth problem | 13:03 |
| ltomasbo | when using containerized | 13:04 |
| dulek | ltomasbo: Okay, this is because of service account. | 13:04 |
| dulek | ltomasbo: So this will need to be fixed in DevStack and openshift-ansible, I guess. | 13:05 |
| ltomasbo | I think that is working on openshift-ansible | 13:05 |
| ltomasbo | I think I already added something for that | 13:05 |
| ltomasbo | let me double check | 13:05 |
| ltomasbo | https://github.com/openshift/openshift-ansible/pull/8340/files | 13:05 |
| ltomasbo | dulek, ^^ | 13:05 |
| ltomasbo | dulek, what I don't understand is how/why it worked for gate: kuryr-kubernetes-tempest-daemon-openshift-octavia | 13:07 |
| dulek | ltomasbo: This change in main.yaml is neough? | 13:07 |
| dulek | ltomasbo: It's not using service account. | 13:07 |
| dulek | ltomasbo: SA is only used in case of containerized - through standard K8s mechanisms. | 13:07 |
| ltomasbo | ahh, true | 13:07 |
| ltomasbo | and yes, that default/main.yaml is working for me with openshift-ansible | 13:08 |
| ltomasbo | thatt clusterrole is the one associated to the sa, right? | 13:08 |
| ltomasbo | dulek, https://github.com/openshift/openshift-ansible/blob/master/roles/kuryr/tasks/serviceaccount.yaml | 13:10 |
| dulek | ltomasbo: I think so. | 13:11 |
| openstackgerrit | Peng Liu proposed openstack/kuryr-kubernetes master: Implement NPWG multi-vif driver https://review.openstack.org/578009 | 13:15 |
| ltomasbo | dulek, this is the problem: https://github.com/openstack/kuryr-kubernetes/blob/master/devstack/lib/kuryr_kubernetes#L484-L493 | 13:16 |
| ltomasbo | we don't have namespaces there | 13:16 |
| ltomasbo | I'll add it | 13:16 |
| dulek | ltomasbo: OH! :) | 13:16 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Ensure OpenShift gate uses the namespace subnet/sg drivers https://review.openstack.org/580680 | 13:20 |
| ltomasbo | dulek, ^^ lets see if this is the last try! xD | 13:22 |
| celebdor | dulek: I'm going with your recommendation to add a timeout to the route creation | 13:22 |
| openstackgerrit | Merged openstack/kuryr-tempest-plugin master: Use expected pod output as a constant https://review.openstack.org/589878 | 13:29 |
| *** garyloug__ has joined #openstack-kuryr | 13:36 | |
| *** jistr is now known as jistr|call | 13:39 | |
| *** garyloug_ has quit IRC | 13:39 | |
| *** jistr|call is now known as jistr | 13:59 | |
| celebdor | ltomasbo: did you consider implementing octavia.network.drivers | 14:16 |
| celebdor | for the Sg problem | 14:16 |
| celebdor | it may be the only workable thing as of now | 14:16 |
| ltomasbo | celebdor, I'm not even sure I know what is that... | 14:17 |
| ltomasbo | celebdor, you mean a new driver for octavia in kuryr? | 14:17 |
| celebdor | ah... That was what I meant | 14:17 |
| celebdor | but it would mean our package in the tripleo container | 14:17 |
| celebdor | so it's not an easy path as well | 14:17 |
| celebdor | :/ | 14:17 |
| celebdor | s/as well/either/ | 14:18 |
| *** hongbin has joined #openstack-kuryr | 14:41 | |
| *** aojea has joined #openstack-kuryr | 14:55 | |
| *** janki has quit IRC | 15:03 | |
| *** aojea has quit IRC | 15:06 | |
| dulek | dmellado, ltomasbo, celebdor: Okay, release time? What do we need to get in before RC is being cut? | 15:07 |
| dmellado | dulek: yep, release time | 15:07 |
| dulek | Because if there's something, then we'll need to do the release tomorrow? | 15:07 |
| *** aojea has joined #openstack-kuryr | 15:07 | |
| dmellado | so, ltomasbo are you ok with your namespaces patches | 15:07 |
| dmellado | ? | 15:07 |
| dmellado | dulek: ltomasbo celebdor | 15:07 |
| dmellado | let's do a quick roll call | 15:07 |
| ltomasbo | well, I'm missing one regarding the gates, so that should not really matter that much | 15:07 |
| ltomasbo | we don't have a solution for the octavia issue, so... | 15:08 |
| dmellado | how about the QE folks? | 15:08 |
| ltomasbo | and I need to take a look to the race when deleting ports/namespaces | 15:08 |
| *** celebdor has quit IRC | 15:08 | |
| dulek | dmellado: You mean tempest-plugin patches? Not a priority for today IMO. | 15:08 |
| dmellado | let's do a quick check on gcheresh/itzik's patches | 15:08 |
| dulek | dmellado: We'll get plenty of time to get those in before final release. | 15:08 |
| dmellado | oh, I thought they had some around for the main repo | 15:09 |
| dmellado | my bad | 15:09 |
| * dmellado keeps trying for them to come upstream xd | 15:09 | |
| dmellado | anyways, dulek apuimedo | 15:09 |
| dulek | I have those two regarding upgrades: https://review.openstack.org/#/c/584377/, https://review.openstack.org/#/c/584421/. | 15:09 |
| dmellado | anything from your side? | 15:09 |
| dmellado | let's see | 15:09 |
| dulek | dmellado: First one is a bit of a non-bug, so might be nice to get it in. | 15:10 |
| dulek | And also there's https://review.openstack.org/#/c/578009/ - NWPG multi vif support. | 15:10 |
| dmellado | Agreed | 15:10 |
| dmellado | celebdor[m]: | 15:10 |
| dmellado | please approve that one youtoo | 15:10 |
| dulek | I'll review that one now, who's going to be second core? | 15:11 |
| dmellado | I'll take a look too dulek | 15:11 |
| dmellado | it has a post-failure | 15:11 |
| dulek | Awesome! | 15:11 |
| dmellado | probably just infra issues | 15:11 |
| dmellado | but let's take a look | 15:11 |
| dulek | Okay, so I'm grabbing a cup of tea and start with it. | 15:12 |
| dmellado | dulek: I'll recheck | 15:12 |
| dmellado | seems like some zuul issue | 15:12 |
| dulek | dmellado: I've already did that. | 15:12 |
| dmellado | oh, cool! | 15:13 |
| *** celebdor has joined #openstack-kuryr | 15:32 | |
| *** phuoc has quit IRC | 15:32 | |
| dmellado | dulek: I need to drop for today, if you need any urgent approve ping me | 15:43 |
| dulek | dmellado: So release coming up tomorrow? | 15:43 |
| celebdor | ltomasbo: dulek: The 4g network dropped completely in my zone :/ | 15:43 |
| dulek | dmellado: Note that I'll be out until Tuesday. | 15:43 |
| dmellado | oh, crap | 15:43 |
| dmellado | me too | 15:43 |
| dulek | dmellado: BTW I haven't told you why I'm on PTO - I'm getting married over the weekend. ;) | 15:43 |
| ltomasbo | celebdor, we noticed... :D | 15:44 |
| dmellado | dulek: oh, CONGRATS dulek !!!! | 15:44 |
| dulek | dmellado: I'd really like to get https://review.openstack.org/#/c/578009 in. | 15:44 |
| dmellado | lemme see | 15:44 |
| dulek | dmellado: I'm finishing my review on it. | 15:44 |
| dmellado | dulek: IMHO overall looks good | 15:44 |
| dmellado | he addressed quite some concerns on the PS | 15:44 |
| celebdor | dulek: congratulations!!! | 15:44 |
| celebdor | I'm already 6 years on | 15:45 |
| dmellado | dulek: I guess it's the year! | 15:45 |
| dulek | dmellado: You too? ;) | 15:45 |
| dmellado | this year a lot of folks are either getting married | 15:45 |
| dmellado | or having kids | 15:45 |
| dmellado | I guess for now I'm on the kids' side | 15:45 |
| dmellado | but we were wondering about just getting married for the sake of paperworks xD | 15:45 |
| celebdor | getting married is no change... Getting kids on the other hand... | 15:45 |
| dulek | I'm still not ready for the latter. ;) | 15:45 |
| dulek | celebdor: That's what I expect. :D | 15:46 |
| celebdor | dmellado: for the sake of easier paperwork i'd give up a ball | 15:46 |
| dmellado | celebdor: lol | 15:46 |
| celebdor | fucking hell, how I hate paperwork | 15:46 |
| openstackgerrit | Vladimir Kuramshin proposed openstack/kuryr-kubernetes master: Add SR-IOV binding driver to CNI https://review.openstack.org/512281 | 15:46 |
| celebdor | dulek: you failed tox https://review.openstack.org/#/c/578009 | 15:46 |
| celebdor | *shame* | 15:46 |
| celebdor | xD | 15:46 |
| dmellado | celebdor: that's infra-related | 15:47 |
| dmellado | shame on infra | 15:47 |
| dmellado | xd | 15:47 |
| celebdor | best kind of shame | 15:47 |
| dulek | celebdor: Hey, it's not even me! :P | 15:47 |
| dulek | pliu: Looks like your patch is getting in. :) | 15:48 |
| dmellado | celebdor: could you go and check https://review.openstack.org/#/c/584377/ | 15:48 |
| openstackgerrit | Vladimir Kuramshin proposed openstack/kuryr-kubernetes master: Add SR-IOV capabilities to VIF handler https://review.openstack.org/512280 | 15:48 |
| celebdor | dmellado: dulek: https://review.openstack.org/#/c/584377 is one scary patch | 15:50 |
| celebdor | :P | 15:50 |
| celebdor | dulek: aren't you dropping old format support on https://review.openstack.org/#/c/584377/3/contrib/kubectl_plugins/kuryr/kuryr ? | 15:51 |
| ltomasbo | man, still authentication issues on the openshift side... | 15:52 |
| dulek | celebdor: We have 3 formats actually. :D | 15:53 |
| dulek | celebdor: 1. single vif (stable/queens) | 15:53 |
| celebdor | right | 15:53 |
| celebdor | I'm not counting the one that didn't make it a whole cycle | 15:53 |
| celebdor | :P | 15:53 |
| dulek | celebdor: 2. Intermediary format that multi-vif patch introduced. | 15:53 |
| dulek | celebdor: Oh yes, okay. | 15:53 |
| dulek | celebdor: Then this snippet in lines 135-137 is compatibility shim. | 15:54 |
| dulek | celebdor: If annotation is PodState o.vo, then extract default_vif from it. | 15:54 |
| dulek | celebdor: Otherwise treat the annotation as vif. | 15:55 |
| celebdor | right | 15:55 |
| celebdor | that's true | 15:55 |
| celebdor | perfect! | 15:55 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Ensure OpenShift gate uses the namespace subnet/sg drivers https://review.openstack.org/580680 | 15:56 |
| celebdor | dulek: what is this pod_state that started appearing everywhere? | 15:56 |
| dulek | celebdor: The o.vo that I introduce. Same as LBaaSState. | 15:56 |
| dulek | celebdor: I mean same concept. | 15:56 |
| celebdor | ok | 15:57 |
| dulek | celebdor: Check out it's definition to understand it better. | 15:57 |
| celebdor | dulek: I don't like the implication that it can change | 15:57 |
| dulek | celebdor: Change? | 15:57 |
| celebdor | since pods can't have hot plugging | 15:57 |
| dulek | celebdor: Sure, but it's just an annotation, so it's no different than how it is today. | 15:59 |
| celebdor | I know, I know | 16:00 |
| celebdor | dulek: https://review.openstack.org/#/c/584421/4 should also be merged, right? | 16:00 |
| dulek | celebdor: It's more of a bug fix, so I guess including it in RC2 is also fine. | 16:01 |
| dmellado | damn timed out amphoraes | 16:01 |
| dmellado | yep | 16:01 |
| dmellado | let's put that to RC2 | 16:01 |
| dmellado | anyways, celebdor | 16:03 |
| dmellado | +A that patch so we can cut the RC | 16:03 |
| dmellado | I gotta leave | 16:03 |
| dmellado | xD | 16:03 |
| celebdor | ok | 16:03 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: drop deprecated oadm https://review.openstack.org/588303 | 16:09 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: Allow musl based distros to resolve using search https://review.openstack.org/580701 | 16:09 |
| dmellado | so then folks, I'll be doing the patch | 16:15 |
| dmellado | and cutting kuryr-kubernetes 0.5.0rc1 | 16:15 |
| dmellado | just will hold until the patches are merged so we can get the hash | 16:15 |
| dmellado | and then will add that one | 16:15 |
| dmellado | celebdor: dulek FYI | 16:15 |
| celebdor | ok | 16:16 |
| dulek | dmellado: Sure! It's your ACK which is needed. | 16:16 |
| dmellado | I'll be proposing the patch and amending the hash later on tonight | 16:16 |
| celebdor | dmellado: https://review.openstack.org/#/c/585345/2 | 16:17 |
| celebdor | please approve this | 16:17 |
| celebdor | and once dulek's patch is merged I'll w+1 it | 16:17 |
| dmellado | it should be good if it depends on that | 16:18 |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Refactor options definitons https://review.openstack.org/590415 | 16:18 |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-tempest-plugin master: WiP: A/P HA tests https://review.openstack.org/590416 | 16:18 |
| dmellado | celebdor: ack, waiting on your w+1 | 16:18 |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: WIP: Add HA gate https://review.openstack.org/588223 | 16:20 |
| dmellado | all right fellas | 16:21 |
| dmellado | off I go, will amend the hash later on today | 16:21 |
| dmellado | g'evening! | 16:21 |
| dulek | I'm out as well! See you on Tuesday! | 16:22 |
| *** hongbin has quit IRC | 16:37 | |
| *** aojea has quit IRC | 16:51 | |
| *** garyloug__ has quit IRC | 17:01 | |
| *** phuoc has joined #openstack-kuryr | 17:01 | |
| openstackgerrit | Vladimir Kuramshin proposed openstack/kuryr-kubernetes master: Add SR-IOV binding driver to CNI https://review.openstack.org/512281 | 17:21 |
| *** lihi has quit IRC | 17:44 | |
| *** lihi has joined #openstack-kuryr | 17:45 | |
| *** celebdor has quit IRC | 18:14 | |
| *** pcaruana has quit IRC | 19:33 | |
| *** aojea has joined #openstack-kuryr | 19:47 | |
| *** neiljerram has quit IRC | 22:08 | |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Implement NPWG multi-vif driver https://review.openstack.org/578009 | 22:09 |
| openstackgerrit | Merged openstack/kuryr-tempest-plugin master: Adapt plugin to new annotation format https://review.openstack.org/585345 | 22:31 |
| *** aojea has quit IRC | 22:46 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!