| *** hongbin has joined #openstack-kuryr | 01:00 | |
| *** pmannidi has joined #openstack-kuryr | 01:13 | |
| *** maysams has quit IRC | 03:31 | |
| *** hongbin has quit IRC | 04:33 | |
| *** pcaruana has joined #openstack-kuryr | 05:41 | |
| *** irenab has joined #openstack-kuryr | 06:40 | |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Add HOWTO for SRIOV use case https://review.openstack.org/594125 | 06:58 |
|---|---|---|
| *** celebdor has joined #openstack-kuryr | 07:39 | |
| *** garyloug has joined #openstack-kuryr | 08:08 | |
| *** garyloug_ has joined #openstack-kuryr | 08:09 | |
| *** garyloug__ has joined #openstack-kuryr | 08:11 | |
| garyloug__ | dmellado, all, this is the os-vif patch as mentioned yesterday https://review.openstack.org/#/c/604992/ | 08:12 |
| dulek | ltomasbo: Whoops, you've merged the SR-IOV docs patch. The document linked there is still private. | 08:28 |
| dulek | I guess I should have gave +1 instead of +2 there. | 08:28 |
| ltomasbo | dulek, it is private? | 08:29 |
| ltomasbo | dulek, I tried and it opened for me... | 08:29 |
| dulek | ltomasbo: Maybe you've got permissions earlier, dunno. | 08:30 |
| danil | ltomasbo, dulek: hello. Could you please explain what wrong with https://review.openstack.org/#/c/600022/13 ?? It says that "Failed to read build info file: ValueError("build info file is broken: IndexError('list index out of range',)",)" | 08:32 |
| danil | but I can't find an issue | 08:32 |
| ltomasbo | dulek, maybe... sorry then, I just checked if I could open it, and as I open it, and saw that the first line said [public] | 08:33 |
| ltomasbo | dulek, i thought it was already fixed | 08:33 |
| dulek | ltomasbo: Sure, I understand that. Let's just make sure aperevalov makes it public. :) | 08:33 |
| dulek | danil: Looking… | 08:34 |
| danil | dulek: thanks | 08:34 |
| *** pmannidi has quit IRC | 08:35 | |
| dulek | danil: http://logs.openstack.org/22/600022/13/check/openstack-tox-docs/89a600b/job-output.txt.gz#_2018-09-21_07_08_45_600310 | 08:35 |
| dulek | danil: Seems like this is the root reason? | 08:35 |
| aperevalov | dulek: it wasn't my document, I asked original authors, it's Gary's colegues, they promissed to publish it yesterday, but they were busy... and they still busy | 08:36 |
| dulek | aperevalov: Sure, sure. Let's just make sure it'll become public one day. ;) | 08:36 |
| danil | dulek: hm, thanks, I'll check it . There is a yaml spec example | 08:37 |
| dulek | danil: I think you shouldn't use "::" followed by code blocks. | 08:39 |
| dulek | danil: But after fixing that it still doesn't build locally for me. | 08:39 |
| danil | dulek: hmmmm, it's strange | 08:40 |
| dulek | danil: Maybe it's some kind of bug on openstackdocstheme or even infra. | 08:45 |
| dulek | danil: Yup, master doesn't build for me as well. | 08:47 |
| danil | dulek: looks like that. I've removed the whole part of yaml description from file but I stil have the same error | 08:47 |
| dmellado | garyloug__: ack, thanks! | 08:59 |
| *** janki has joined #openstack-kuryr | 09:20 | |
| *** janki has quit IRC | 10:07 | |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: ignore Editor tags https://review.openstack.org/605031 | 10:19 |
| celebdor | dmellado: are you around? | 10:27 |
| dmellado | celebdor: what's up? | 10:27 |
| celebdor | dmellado: what did you say that you wanted done in ansible | 10:27 |
| celebdor | ? | 10:27 |
| celebdor | and why | 10:27 |
| dmellado | hmmm I did recall wanting to swamp you into something regarding ansible | 10:28 |
| dmellado | but as of lately my mind is a LIFO pipe | 10:28 |
| dmellado | so who knows! xD | 10:28 |
| *** maysams has joined #openstack-kuryr | 10:28 | |
| dmellado | I could've said that I want you to implement a platypus ascii for kuryr related roles | 10:28 |
| celebdor | dmellado: wasn't it something to do with redoing most of our devstack stuff in ansible? | 10:30 |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: Translate security group rules. https://review.openstack.org/592349 | 10:30 |
| dmellado | celebdor: oh, yeah | 10:30 |
| celebdor | can you give me more details? | 10:31 |
| dmellado | now I'm starting to remember | 10:31 |
| dmellado | so, basically, as a part of dulek's conversations | 10:31 |
| celebdor | we should make a blueprint and such | 10:31 |
| dmellado | on the PTG | 10:31 |
| dmellado | yeah, deff | 10:31 |
| dmellado | otherwise we'll deff forget as you're even worse than me | 10:31 |
| celebdor | right | 10:31 |
| celebdor | I am | 10:31 |
| celebdor | so | 10:32 |
| celebdor | what will it be | 10:32 |
| dmellado | btw, before I forget too | 10:32 |
| celebdor | one playbook that is called from the plugin.sh? | 10:32 |
| dmellado | ltomasbo: celebdor | 10:32 |
| dmellado | could you please +2+A that patch? | 10:32 |
| dmellado | I've just gone over a trivial change | 10:32 |
| celebdor | which? | 10:32 |
| ltomasbo | dmellado, which one? | 10:33 |
| dmellado | https://review.openstack.org/592349 | 10:33 |
| dmellado | I've removed the handler unit test | 10:33 |
| dmellado | and will put that on a diff patch | 10:33 |
| dmellado | as we agreed | 10:33 |
| dmellado | otherwise, mainly the same | 10:33 |
| ltomasbo | dmellado, great! thanks! | 10:33 |
| dmellado | celebdor: yeah, basically a rewrite of our devstack plugin | 10:33 |
| dmellado | in ansible | 10:33 |
| dmellado | for the multinode thing | 10:33 |
| dmellado | dulek: could you ellaborate that? | 10:33 |
| * dulek reads… | 10:34 | |
| dulek | dmellado: More like for multinode and nested stuff. To correctly implement it we need some kind of synchronization between what's running on different nodes. devstack-gate is unable to provide us with this. | 10:35 |
| dmellado | yeah, so we wanted that to be done in ansible | 10:35 |
| dmellado | so we could control that node-like | 10:35 |
| dmellado | now it gets back | 10:35 |
| dmellado | lol | 10:35 |
| dulek | So I propose implementing Kubernetes/OpenShift installation as Ansible playbooks to be able to do that. | 10:35 |
| dulek | Moreover it'll be easier to make it run on VMs created on DevStack, instead of DevStack itself. | 10:36 |
| celebdor | how will ansible help with that? | 10:36 |
| dulek | celebdor: The "remote" execution or the synchronization? | 10:37 |
| celebdor | both | 10:37 |
| dulek | celebdor: Well, in case of sync in Ansible you can generate certificates and .kubeconfig and then copy it onto the nodes. devstack-gate runs are simply independent on both nodes. | 10:39 |
| dulek | Of course in case of "non-nested" deployment we could just deploy from controller's plugin.sh. | 10:39 |
| dulek | I mean nested. :P | 10:39 |
| celebdor | dulek: how do you know about the nodes? | 10:40 |
| celebdor | they are independent, right? | 10:40 |
| dulek | celebdor: In nested case nodes are VMs you spawn, so it's pretty easy. | 10:40 |
| celebdor | dulek: I was thinking of gates | 10:40 |
| dulek | In non-nested case… Hm, ansible should have them in inventory. | 10:41 |
| celebdor | dynamic inventory does htat | 10:41 |
| celebdor | but how can we get the list of nodes in the current job | 10:41 |
| celebdor | and credentials to access them | 10:41 |
| celebdor | dmellado: ^^ | 10:41 |
| dmellado | celebdor: we'd need to check how the default job does that | 10:42 |
| dmellado | I'll add that to my research topics | 10:42 |
| dulek | celebdor: http://logs.openstack.org/29/601829/1/check/kuryr-kubernetes-tempest-multinode-octavia/536b5c1/ara-report/ | 10:42 |
| dulek | celebdor: Seems like OpenStack's Ansible has it in the inventory already. | 10:42 |
| dulek | celebdor: I mean gate's Ansible. | 10:42 |
| celebdor | dulek: does that run in the node? | 10:43 |
| dulek | celebdor: What's "node"? | 10:44 |
| celebdor | nevermind | 10:44 |
| celebdor | so we have the controller label for a host | 10:45 |
| dmellado | yep | 10:46 |
| celebdor | and the other one is labeled 'compute'? Or what? | 10:46 |
| dmellado | celebdor: check even our own gate definitions | 10:46 |
| dmellado | you have subnodes and so | 10:46 |
| dmellado | you lazy emerit TPL xD | 10:46 |
| dmellado | seriously, would you like to have a quick bluejeans | 10:46 |
| dmellado | so I explain this to you? | 10:46 |
| celebdor | dmellado: I'd rather just know which label to use for the worker nodes | 10:46 |
| celebdor | xD | 10:46 |
| dmellado | lol | 10:47 |
| dmellado | celebdor: i.e. https://review.openstack.org/#/c/599608/1/.zuul.d/base.yaml | 10:47 |
| dulek | celebdor: "controller" and "compute1"? | 10:48 |
| celebdor | ok | 10:48 |
| celebdor | thanks | 10:48 |
| *** janki has joined #openstack-kuryr | 10:50 | |
| dmellado | dulek: btw, as I said before, could you merge https://review.openstack.org/592349 | 10:53 |
| dmellado | I'd like to move on to the next PS after it | 10:53 |
| celebdor | dmellado: let's say we do the ansible thing | 10:55 |
| celebdor | how would we deal with the devstack stages? | 10:55 |
| dmellado | we'll implement whatever necessary on playbooks | 10:55 |
| dmellado | and we can call that from our devstack plugin | 10:55 |
| dmellado | in any case we need to do a blueprint | 10:55 |
| celebdor | so split it into multiple different playbooks? | 10:55 |
| celebdor | what a drag | 10:56 |
| dmellado | exactly | 10:56 |
| dmellado | that's why I wanted you to do it | 10:56 |
| dmellado | xD | 10:56 |
| dulek | dmellado: You know that you can develop next commits locally without previous being merged? :D | 10:56 |
| dmellado | dulek: I knwo | 10:56 |
| dmellado | and I'm already doing that | 10:56 |
| dmellado | but the less I rebase, the better | 10:56 |
| dmellado | xD | 10:56 |
| dmellado | so c'mon! xD | 10:56 |
| dmellado | see? was it so hard? xD | 10:56 |
| dulek | dmellado: It's not that bad actually if you use `git rebase -i HEAD~10`, and then move commits around with "fixup" action. | 10:57 |
| celebdor | dulek: preach! | 10:57 |
| dmellado | LOL | 10:57 |
| celebdor | dulek: it seems the lazy PTL is him | 10:57 |
| dmellado | dulek: do you know that I once had an intern who confused merge with rebase and did squashes? | 10:57 |
| celebdor | xD | 10:57 |
| dmellado | xD | 10:57 |
| dmellado | celebdor: we both are | 10:57 |
| dmellado | xD | 10:57 |
| dmellado | now that I think about it, that was even worse than the pip thing on superfluidity | 10:58 |
| * dmellado shivers | 10:58 | |
| celebdor | lol | 10:58 |
| dulek | dmellado: We've once worked with a team of folks that shared the SSH key and all the commits were attributed to their team leader. ;) | 10:58 |
| dmellado | LOL | 10:58 |
| dmellado | really? xD | 10:58 |
| dulek | dmellado: Yup, we were quite surprised about how quick the guy is producing the code. :D | 10:59 |
| dulek | dmellado: My point is - we've all seen *things*. | 10:59 |
| dmellado | lol | 10:59 |
| dmellado | let's call them 'things' yeah | 10:59 |
| dmellado | such as the guy who refused to use any kind of control version | 11:00 |
| dmellado | and sent his changes over email | 11:00 |
| dmellado | to all the team, daily | 11:00 |
| dmellado | xD | 11:00 |
| * dmellado off for lunch | 11:01 | |
| *** janki has quit IRC | 11:07 | |
| *** pcaruana has quit IRC | 11:15 | |
| *** maysams has quit IRC | 11:47 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: svc namespace isolation support for different octavia drivers https://review.openstack.org/604717 | 11:49 |
| *** aojea has joined #openstack-kuryr | 12:00 | |
| *** rh-jelabarre has joined #openstack-kuryr | 12:05 | |
| *** aojea has quit IRC | 12:51 | |
| *** maysams has joined #openstack-kuryr | 13:10 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Move Python 3.6 jobs to check and gate as voting https://review.openstack.org/605068 | 13:22 |
| *** aojea has joined #openstack-kuryr | 13:26 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-kubernetes master: Remove kuryr-kubernetes-tempest-lbaasv2 job https://review.openstack.org/605070 | 13:27 |
| *** aojea has quit IRC | 13:30 | |
| openstackgerrit | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Sync up with kuryr-kubernetes gates https://review.openstack.org/605072 | 13:32 |
| dulek | celebdor, dmellado, ltomasbo, irenab: ^ This is my try to reorganize our gates a bit. | 13:33 |
| ltomasbo | dulek, thanks! reviewing... | 13:33 |
| dmellado | yep, reviewing | 13:34 |
| dmellado | I was about to push the 1) | 13:34 |
| dmellado | so thanks for being faster xD | 13:34 |
| dmellado | dulek: in any case, note | 13:34 |
| dmellado | you should do the same and sync for the tempest plugin gates | 13:34 |
| dmellado | otherwise we'll differ and chaos will reign | 13:34 |
| dmellado | xD | 13:34 |
| dmellado | (and ltomasbo's bad jokes, which are better than mine, but that's another topic) | 13:34 |
| ltomasbo | dmellado, ???? xD | 13:35 |
| dmellado | xD | 13:35 |
| ltomasbo | dmellado, actually, dulek has already sync the tempest side too | 13:35 |
| ltomasbo | dmellado, look at the last patch | 13:35 |
| dmellado | oh, cool | 13:36 |
| dmellado | then chaos will not reign | 13:36 |
| dmellado | just bad jokes | 13:36 |
| dmellado | xD | 13:36 |
| ltomasbo | xD | 13:36 |
| dulek | :) | 13:36 |
| dulek | At the moment tempest plugin seems to be compatible with Python 3, but I've noticed one of gcheresh patches in review isn't. | 13:37 |
| dmellado | dulek: there you go, -1 | 13:37 |
| dmellado | xD | 13:37 |
| dulek | So I want to add py36 gate there, because I don't think we can notice incompatibilities manually. | 13:37 |
| dmellado | nope, don't let me have nightmares about manual testing, please | 13:38 |
| dmellado | already been thru that too many times already | 13:38 |
| dulek | It's a bit of an issue that there are some tests running only with certain configs. E.g. gcheresh patch is one of those, so this py36 gate won't help here anyway. | 13:38 |
| dulek | I'm starting to think about "consolidating" some of the gates. | 13:39 |
| dmellado | what we could do is just port all gates to py3 | 13:39 |
| dmellado | yep | 13:39 |
| dulek | E.g. I don't really see a point in running separated openshift, openshift with ingress and openshift with namespaces gates. This could be joined into a single one. | 13:39 |
| dmellado | dulek: makes sense, as this started growing bit by bit | 13:40 |
| dmellado | now we'd be able to just set the defaults and if there's any odd leftover that can't be left out | 13:41 |
| dmellado | that'll be put out to experimental | 13:41 |
| dulek | dmellado: Yup. I haven't put out patches doing that, because we might want to discuss this on Monday's meeting. I might be missing some dependencies. | 13:42 |
| dmellado | +1 | 13:42 |
| dmellado | let's discuss that as I'd also like for gecheresh and the folks to be around | 13:43 |
| ltomasbo | dulek, I saw your comment about no-daemon removal on stein | 14:41 |
| ltomasbo | dulek, aren't we already on stein? | 14:42 |
| dulek | ltomasbo: Yes, we are. I was thinking that it should be delayed to T(rain) to follow deprecation policy (which we don't officially follow, BTW), but looks like it's not. | 14:45 |
| dulek | ltomasbo: I guess that's another topic for Monday's meeting? | 14:45 |
| ltomasbo | I don't mind having it there until T if that is the standard policy | 14:45 |
| dulek | ltomasbo: Nah, turns out it's just next release *and* at least 3 months. So we're past that, I've deprecated it 6 months ago. | 14:46 |
| ltomasbo | dulek, ok, for that, should we also remove the non-daemonize code? | 14:47 |
| dulek | ltomasbo: I would love to do that. :D | 14:47 |
| ltomasbo | +1 from my side | 14:47 |
| ltomasbo | it will make it simple to maintain | 14:47 |
| *** celebdor has quit IRC | 15:01 | |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: ignore Editor tags https://review.openstack.org/605031 | 15:12 |
| *** celebdor has joined #openstack-kuryr | 15:15 | |
| *** celebdor has quit IRC | 15:19 | |
| *** rh-jelabarre has quit IRC | 15:39 | |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Translate security group rules. https://review.openstack.org/592349 | 15:41 |
| *** rh-jelabarre has joined #openstack-kuryr | 15:42 | |
| *** aperevalov has quit IRC | 15:55 | |
| *** pcaruana has joined #openstack-kuryr | 16:07 | |
| *** garyloug_ has quit IRC | 16:23 | |
| *** garyloug has quit IRC | 16:23 | |
| *** garyloug__ has quit IRC | 16:23 | |
| dmellado | let's remove it | 16:50 |
| dmellado | it'll make things much simpler | 16:50 |
| dmellado | and I'd like to avoid having a big pipeline there | 16:50 |
| openstackgerrit | Daniel Mellado proposed openstack/kuryr-kubernetes master: [WIP] Support network policy update https://review.openstack.org/605150 | 17:01 |
| *** celebdor has joined #openstack-kuryr | 17:23 | |
| *** aojea has joined #openstack-kuryr | 18:13 | |
| *** maysams has quit IRC | 18:26 | |
| *** celebdor has quit IRC | 18:34 | |
| *** dmellado has quit IRC | 18:53 | |
| *** dmellado has joined #openstack-kuryr | 20:07 | |
| *** celebdor has joined #openstack-kuryr | 20:10 | |
| *** janki has joined #openstack-kuryr | 20:12 | |
| *** pcaruana has quit IRC | 20:43 | |
| *** janki has quit IRC | 21:08 | |
| *** aojea has quit IRC | 21:58 | |
| *** aojea has joined #openstack-kuryr | 23:31 | |
| *** aojea has quit IRC | 23:35 | |
| *** aojea has joined #openstack-kuryr | 23:54 | |
| *** aojea has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!