*** hongbin has joined #openstack-kuryr | 02:13 | |
*** hongbin has quit IRC | 02:46 | |
*** hongbin has joined #openstack-kuryr | 03:21 | |
*** hongbin has quit IRC | 03:30 | |
*** gkadam has joined #openstack-kuryr | 05:45 | |
*** ccamposr has joined #openstack-kuryr | 07:16 | |
*** pcaruana has joined #openstack-kuryr | 07:25 | |
*** maysams has joined #openstack-kuryr | 07:51 | |
*** aperevalov has joined #openstack-kuryr | 07:53 | |
*** yboaron_ has quit IRC | 08:12 | |
openstackgerrit | Maysa de Macedo Souza proposed openstack/kuryr-kubernetes master: Add support for matchExpressions https://review.openstack.org/620572 | 08:22 |
---|---|---|
*** celebdor has joined #openstack-kuryr | 08:30 | |
*** yboaron_ has joined #openstack-kuryr | 08:51 | |
*** yboaron_ has quit IRC | 08:56 | |
*** yboaron_ has joined #openstack-kuryr | 08:57 | |
*** celebdor has quit IRC | 09:28 | |
*** celebdor has joined #openstack-kuryr | 09:31 | |
*** garyloug has joined #openstack-kuryr | 09:45 | |
*** EricAdamsZNC has quit IRC | 10:04 | |
*** EricAdamsZNC has joined #openstack-kuryr | 10:07 | |
ltomasbo | celebdor, dmellado, dulek, irenab: please review when you have time: https://review.openstack.org/#/c/619632/ | 10:45 |
dmellado | ltomasbo: added to my queue, thanks! | 10:45 |
irenab | same | 10:45 |
celebdor | will do | 10:46 |
ltomasbo | dmellado, btw, has anything change on the gates lately? seems we are getting a lot of timeouts... | 10:46 |
dmellado | not that I'm aware | 10:46 |
ltomasbo | http://logs.openstack.org/61/616261/25/gate/kuryr-kubernetes-tempest-daemon-containerized-octavia-py36/211ff56/controller/logs/kubernetes/pod_logs/kube-system-kuryr-cni-ds-2lj5s-kuryr-cni.txt.gz | 10:46 |
ltomasbo | this one at least seems to have a collision on the kuryr-cni port... | 10:46 |
dmellado | but we'll need to check those timeouts with the infra folks | 10:46 |
celebdor | ltomasbo: dulek: maysams: what are the reasons we always bind to the same port in cni and in the controller? | 10:58 |
celebdor | Is it just for the health probes? | 10:58 |
dulek | celebdor: We bind probes to different ports. | 10:58 |
dulek | celebdor: At least by default. | 10:58 |
dulek | celebdor: Otherwise you wouldn't be able to run controller on nodes having CNI. | 10:59 |
celebdor | dulek: no, no. What I meant to say is, cni always binds to port X and controller to port Y | 10:59 |
ltomasbo | celebdor, yep, controller is 8082 and cni 8090 (IIRC) | 10:59 |
celebdor | unless otherwise stated in config | 10:59 |
dulek | celebdor: True. You'd like it to be random? | 10:59 |
ltomasbo | random is never a good choice... | 10:59 |
celebdor | dulek: I was wondering if, considering it is host networking and many things can bind in that namespace | 10:59 |
dulek | It's not that bad, but very hard to implement. | 10:59 |
celebdor | we could probably have a range | 10:59 |
dulek | celebdor: I don't think it can be done with probes definitions in DaemonSet and Deployment. | 11:00 |
dulek | celebdor: Those require a single port if I'm not unaware of some cool mechanism. | 11:00 |
celebdor | dulek: it pretty much forces you to switch to executable probes, you are right | 11:00 |
celebdor | you would execute curl inside the container then | 11:01 |
dulek | celebdor: Or kuryr-status. | 11:01 |
dulek | That can be done, sure. | 11:01 |
openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: Introduce test case document for SRIOV functionality https://review.openstack.org/600022 | 11:09 |
*** aperevalov has quit IRC | 11:22 | |
*** aperevalov has joined #openstack-kuryr | 11:24 | |
ltomasbo | ping irenab | 11:26 |
ltomasbo | irenab, I don't get your comment here: https://review.openstack.org/#/c/619632/12/kuryr_kubernetes/controller/drivers/network_policy.py | 11:27 |
ltomasbo | irenab, the call at line 247 is using the CIDR, why the one at line 252 is not (as it is meant to be open for all) | 11:29 |
ltomasbo | irenab, so, probably I'm missing something... | 11:29 |
*** celebdor has quit IRC | 11:38 | |
*** celebdor has joined #openstack-kuryr | 11:40 | |
irenab | ltomasbo, pong | 11:42 |
irenab | the only difference I see is that protocol is set differently, but the rest of the call is the same | 11:42 |
ltomasbo | irenab, hi! thanks for the review! I have a question thought! | 11:42 |
ltomasbo | irenab, which lines? | 11:42 |
irenab | at least I could not fine any more differences | 11:43 |
irenab | 247 and 252 | 11:43 |
ltomasbo | irenab, I see the protocol is set in the same way, but I'm adding cidr on 247, and not on 252 | 11:43 |
irenab | same method invocation, right? | 11:43 |
irenab | ltomasbo, let me revisit | 11:43 |
ltomasbo | I see in both of them protocol=port.get('protocol') | 11:44 |
irenab | ltomasbo, my bad, somehow missed it. maybe worth to split the line and get the cidr on the following line for the sake of readability | 11:45 |
ltomasbo | irenab, sure! I was simply not sure about what you meant | 11:46 |
ltomasbo | in line 247 basically we call it X times (one per each cidr), while in liine 257 is without cird (as it is open from everywhere) | 11:46 |
openstackgerrit | Daniel Mellado proposed openstack/kuryr-tempest-plugin master: Add network policy gate to experimental pipeline https://review.openstack.org/621550 | 11:47 |
ltomasbo | (in case that wa the part that it was not clear) | 11:47 |
ltomasbo | irenab, I'll update the patch asap! | 11:47 |
irenab | ltomasbo, yes, got it, thanks! | 11:47 |
ltomasbo | thank you for the review! | 11:48 |
dmellado | ltomasbo: irenab dulek https://review.openstack.org/621550 | 11:48 |
dmellado | it's a one-liner but we'll need it as well | 11:49 |
ltomasbo | dmellado, sure! Can I make one request? | 11:49 |
dmellado | shoot | 11:50 |
ltomasbo | as we are going to have network policies with namespace (handler) can you rename it (on both side) | 11:50 |
ltomasbo | kuryr-kubernetes-tempest-daemon-containerized-octavia-namespace-network-policy | 11:50 |
ltomasbo | dmellado, ^6 | 11:50 |
dmellado | ltomasbo: in order not to mess up with zuul | 11:51 |
dmellado | let's add this first and then I'll rename | 11:51 |
ltomasbo | fine with me!\ | 11:51 |
dmellado | otherwise it'll depend on another patch and I've seen odd things with this on another similar renamings | 11:51 |
dmellado | ltomasbo: /me goes finally to the sofa in order to die feverish death | 11:56 |
dmellado | ping me if you need anything | 11:56 |
ltomasbo | ok! get better! | 11:56 |
openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Add namespaceSelector support for NetworkPolicies https://review.openstack.org/619632 | 11:57 |
dulek | celebdor: Hi! Regarding your comment on https://review.openstack.org/#/c/620045/4/devstack/plugin.sh, you want me to put each parameter of the command into a bash array and just unpack it when needed? | 11:59 |
irenab | ltomasbo, thanks for the update and sorry for confusion | 12:01 |
ltomasbo | no problem! | 12:01 |
openstackgerrit | Merged openstack/kuryr-kubernetes master: Remove commented code that should be deleted https://review.openstack.org/601829 | 12:02 |
irenab | ltomasbo, regarding the verification. Do you plan to add some section in the installation guide similar to https://review.openstack.org/#/c/605150/28/doc/source/installation/network_policy.rst or tests? | 12:04 |
ltomasbo | irenab, ohh, nice catch! let me add info about the namespaceSelector there! | 12:05 |
irenab | cool, thanks | 12:06 |
openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Add namespaceSelector support for NetworkPolicies https://review.openstack.org/619632 | 12:18 |
openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Ensure pod relabeling is supported by the Network Policy https://review.openstack.org/618521 | 12:18 |
openstackgerrit | Maysa de Macedo Souza proposed openstack/kuryr-kubernetes master: Add support for matchExpressions https://review.openstack.org/620572 | 12:22 |
openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: cri-o support https://review.openstack.org/620045 | 12:35 |
openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Add experimental gate with cri-o https://review.openstack.org/620388 | 12:35 |
openstackgerrit | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Add possibility to build with podman/buildah https://review.openstack.org/620807 | 12:35 |
aperevalov | Hi, what kind of tempest credentials are you using in kuryr-tempest-test? Dynamic or pre-provisioned? | 12:58 |
openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: Introduce test case document for SRIOV functionality https://review.openstack.org/600022 | 13:05 |
*** rh-jelabarre has joined #openstack-kuryr | 13:06 | |
*** EricAdamsZNC has quit IRC | 13:25 | |
*** EricAdamsZNC has joined #openstack-kuryr | 13:28 | |
celebdor | dulek: yes, it's more correct to put all the cmd in an array and then just use it like "${myarray[@]}" | 13:33 |
dulek | aperevalov: What do you mean by what credentials? The DevStack's ones are used, IIUC. | 13:54 |
dulek | celebdor: Yep, I've updated the patch to do that, let's see if I haven't screwed this up. | 13:55 |
dulek | dmellado: Meeting time? | 14:01 |
aperevalov | dulek: I mean "Credential Provider Mechanisms", which one are you using, do you have use_dynamic_credentials = True in you [auth] section or such kind of predefined credentials in etc/account.yaml? | 14:01 |
dmellado | hey dulek, I'm not feeling well and feverish | 14:01 |
dmellado | I asked celebdor to run the meeting today | 14:02 |
* dmellado goes back to the sofa | 14:02 | |
dmellado | dulek: if celebdor forgers, could you start the meeting? | 14:02 |
dmellado | thanks! | 14:02 |
dulek | dmellado: Doing that now. ;) | 14:02 |
*** yboaron_ has quit IRC | 14:05 | |
*** yboaron_ has joined #openstack-kuryr | 14:05 | |
*** garyloug_ has joined #openstack-kuryr | 14:32 | |
aperevalov | dulek: The new user is because I'm using dynamic credentials provider, as said in tempest documentation, this approach will create new user. And it creates. So I'm interesting, are you using the same way on gate server? Or you using predefined account.yaml? Maybe I'll describe my setup. I used local.conf.sample from kuryr-kubernets/devstack. But after I'v done my installation, I didn't see kuryr_tempest_test in tempest list-plugin output. So | 14:35 |
aperevalov | pip install in venv. | 14:35 |
*** garyloug has quit IRC | 14:36 | |
dulek | aperevalov: We're running from admin account if I recall correctly… | 14:37 |
dulek | aperevalov: Uh, I'm not sure if local.conf.sample has the plugin… | 14:38 |
aperevalov | dulek: it has this one enable_plugin kuryr-kubernetes ... | 14:39 |
dulek | aperevalov: Hm, but not kuryr-tempest-plugin, right? | 14:39 |
aperevalov | or you mean kuryr-tempest-test? | 14:39 |
dulek | aperevalov: You'll need `enable_service tempest` and enable the plugin so it will get installed. | 14:40 |
aperevalov | yes, I can't find it there. | 14:40 |
dulek | aperevalov: You can always try installing it by `pip install -e .`. | 14:40 |
dulek | aperevalov: I mean the plugin. As long as you have had the tempest. | 14:40 |
aperevalov | tempest's documentation promises to work with every correct openstack installation ) | 14:42 |
openstackgerrit | Genadi Chereshnya proposed openstack/kuryr-tempest-plugin master: Working with noop for vif_pool_driver https://review.openstack.org/621411 | 14:55 |
dulek | aperevalov: Sure, sure, it should work fine. | 14:55 |
dulek | aperevalov: But if you want tempest.conf to be automatically filled up, then you need enable_service tempest in local.conf | 14:55 |
celebdor | dulek: I'll split the move to master dir in https://review.openstack.org/#/c/619018/2 to a separate change | 14:58 |
dulek | celebdor: If it's related, then I don't care too much. | 14:58 |
dulek | celebdor: It just seemed odd. :P | 14:59 |
celebdor | well, they are all part of a branch in which I'm adding support for 3.11 | 14:59 |
celebdor | so it doesn't matter to split it | 14:59 |
openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: cri-o support https://review.openstack.org/620045 | 15:01 |
openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Add experimental gate with cri-o https://review.openstack.org/620388 | 15:01 |
openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: devstack: parametrize openshift version fetching https://review.openstack.org/619018 | 15:03 |
openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: devstack: move master config to separate dir https://review.openstack.org/621597 | 15:03 |
celebdor | dulek: there you go | 15:03 |
aperevalov | garyloug_: hi, are you agree if I'l rebase you patch (https://review.openstack.org/#/c/559363/)? | 15:21 |
garyloug_ | Hi aperevalov, sure no problem :-) | 15:24 |
garyloug_ | Thank you | 15:24 |
openstackgerrit | Merged openstack/kuryr-kubernetes master: Ensure existing pods use the right network policy https://review.openstack.org/616261 | 15:25 |
*** garyloug__ has joined #openstack-kuryr | 15:26 | |
*** garyloug_ has quit IRC | 15:30 | |
*** maysams has quit IRC | 15:51 | |
*** yboaron_ has quit IRC | 15:54 | |
*** janki has joined #openstack-kuryr | 15:58 | |
*** janki has quit IRC | 16:26 | |
*** ccamposr has quit IRC | 16:46 | |
*** garyloug__ has quit IRC | 17:49 | |
openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: cri-o support https://review.openstack.org/620045 | 18:31 |
openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Add experimental gate with cri-o https://review.openstack.org/620388 | 18:31 |
*** aojea has joined #openstack-kuryr | 19:29 | |
*** pcaruana has quit IRC | 22:18 | |
*** irclogbot_1 has quit IRC | 22:47 | |
*** aojea has quit IRC | 22:48 | |
*** rh-jelabarre has quit IRC | 22:54 | |
*** rh-jelabarre has joined #openstack-kuryr | 22:57 | |
*** rh-jelabarre has quit IRC | 23:02 | |
*** irclogbot_1 has joined #openstack-kuryr | 23:09 | |
*** gkadam has quit IRC | 23:48 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!