*** dougbtv|laptop has quit IRC | 00:24 | |
*** yuanying has joined #openstack-kuryr | 00:34 | |
*** hongbin has quit IRC | 00:43 | |
*** limao has joined #openstack-kuryr | 00:53 | |
openstackgerrit | Ruijing proposed openstack/kuryr-kubernetes master: Add ODL devstack example https://review.openstack.org/449309 | 01:27 |
---|---|---|
*** vikasc has quit IRC | 02:16 | |
*** vikasc has joined #openstack-kuryr | 02:33 | |
*** hongbin has joined #openstack-kuryr | 03:10 | |
openstackgerrit | Ruijing proposed openstack/kuryr-kubernetes master: Fix vagrant ssh issue for virtualbox bento https://review.openstack.org/453445 | 03:20 |
vikasc | ivc_, ping | 03:46 |
vikasc | ivc_, f target pod gets launched on a node which is not using kuryr-cni, then lbaas will get created in Neutron. And this lbaas does not gets deleted. I encountered this scenario on my local cluster. Now wondering does Kuryr support this type of cluster where some nodes can use non-kuryr cni. Thoughts? | 03:46 |
vikasc | s/f/if | 03:46 |
*** limao has quit IRC | 04:07 | |
*** limao has joined #openstack-kuryr | 04:08 | |
*** hongbin has quit IRC | 04:08 | |
*** limao has quit IRC | 04:12 | |
openstackgerrit | vikas choudhary proposed openstack/kuryr-kubernetes master: Add devstack support for https k8s api server https://review.openstack.org/453138 | 04:19 |
*** limao has joined #openstack-kuryr | 04:56 | |
*** limao has quit IRC | 05:01 | |
*** limao has joined #openstack-kuryr | 05:02 | |
*** aojea has joined #openstack-kuryr | 05:21 | |
*** aojea has quit IRC | 05:30 | |
*** pmannidi has quit IRC | 05:47 | |
*** janki has joined #openstack-kuryr | 05:48 | |
*** jchhatbar has joined #openstack-kuryr | 05:59 | |
*** janki has quit IRC | 06:02 | |
*** jchhatbar is now known as janki | 06:06 | |
*** pcaruana has joined #openstack-kuryr | 06:08 | |
*** pmannidi has joined #openstack-kuryr | 06:12 | |
*** kzaitsev_ws has joined #openstack-kuryr | 06:38 | |
*** ltomasbo|away is now known as ltomasbo | 06:58 | |
*** dimak_ has joined #openstack-kuryr | 07:10 | |
*** gsagie has joined #openstack-kuryr | 07:22 | |
*** aojea has joined #openstack-kuryr | 07:24 | |
*** vikasc is now known as vikasc|lunch | 07:31 | |
*** pmannidi has quit IRC | 07:47 | |
*** egonzalez has joined #openstack-kuryr | 07:48 | |
*** gsagie has quit IRC | 07:54 | |
apuimedo | vikasc|lunch: you encountered that problem with this patch series or with the original WIP? | 07:59 |
vikasc|lunch | apuimedo, with this patch series | 08:00 |
apuimedo | ok | 08:00 |
apuimedo | I would be fine with just a REVISIT note | 08:00 |
*** vikasc|lunch is now known as vikasc | 08:00 | |
vikasc | apuimedo, me too | 08:00 |
*** aojea_ has joined #openstack-kuryr | 08:01 | |
vikasc | apuimedo, and may be bug ticket for the sake of tracking "Kuryr service handling should not create lbaas in Neutron if all endpoints of service are pods on non-kuryr cni" | 08:02 |
*** aojea has quit IRC | 08:04 | |
*** kzaitsev_ws has quit IRC | 08:13 | |
openstackgerrit | vikas choudhary proposed openstack/kuryr-kubernetes master: Add devstack support for https k8s api server https://review.openstack.org/453138 | 08:16 |
apuimedo | vikasc: agreed | 08:24 |
dmellado | +1, sounds reasonable | 08:25 |
dmellado | unlike the spanish tax agency | 08:25 |
dmellado | xD | 08:25 |
apuimedo | xD | 08:27 |
*** kzaitsev_ws has joined #openstack-kuryr | 08:32 | |
*** neiljerram has joined #openstack-kuryr | 08:39 | |
ltomasbo | hi folks | 08:45 |
ltomasbo | apuimedo, ivc_, I'm testing the lbaas patch | 08:46 |
ltomasbo | and when a service is created I'm getting the next ResourceNotReady exception | 08:46 |
ltomasbo | ResourceNotReady: Resource not ready: LBaaSLoadBalancer(id=<?>,ip=172.30.122.141,name='default/docker-registry',project_id='af706450a3814e99853b1582f64facb2',subnet_id=f374cfce-e45c-435e-908b-0ae72db96747) | 08:46 |
apuimedo | ltomasbo: ivc_: isn't it just waiting for the LB? | 08:47 |
apuimedo | (to be active) | 08:47 |
ltomasbo | it makes sense that we get that error as the loadbalancer is not yet in place | 08:47 |
ltomasbo | but, the information in that message must be wrong | 08:47 |
ltomasbo | since it is pointing to the docker-registry | 08:47 |
ltomasbo | which it is running since long before | 08:47 |
ltomasbo | it should belong to the new-created lb, right? | 08:48 |
ltomasbo | something like 'name=default/ruby-hello-world' ... | 08:48 |
ltomasbo | apuimedo, ivc_: or am I missing something? | 08:49 |
*** garyloug has joined #openstack-kuryr | 08:52 | |
* dmellado having some network issues today | 08:53 | |
dmellado | might be offline to check | 08:53 |
*** neiljerram has quit IRC | 09:21 | |
*** neiljerram has joined #openstack-kuryr | 09:23 | |
*** limao has quit IRC | 09:25 | |
*** limao has joined #openstack-kuryr | 09:26 | |
apuimedo | ltomasbo: I'd bet a chickpea that it is because it saw teh docker registry service for the first time | 09:27 |
apuimedo | and tried to create an lb for it | 09:27 |
apuimedo | (so you may not need to manually create LBs for infra services with this patch | 09:28 |
apuimedo | ) | 09:28 |
ltomasbo | yes, it automatically created them | 09:28 |
ltomasbo | but that was also after that happens | 09:28 |
ltomasbo | and every now and then, without I get more of those warnings | 09:28 |
ltomasbo | without pods being created | 09:28 |
ltomasbo | but yes, it was nice to use the patch to create those lbaas, pools, members, instead of having to do that manually... | 09:29 |
*** limao has quit IRC | 09:30 | |
kzaitsev_ws | ltomasbo: I bet it's not even the new patch | 09:41 |
*** mchiappero has quit IRC | 09:42 | |
kzaitsev_ws | it doesn't throw ResourceNotReady anywhere in the code | 09:42 |
kzaitsev_ws | LBaaSSpecHandler does and it is already merged =) | 09:43 |
janonymous | Can i know the steps which could be used to test lbaas service patch? | 09:53 |
janonymous | kzaitsev_ws:What steps did you used? | 09:54 |
dmellado | apuimedo: re: ansible openshift | 09:54 |
dmellado | did you push to the repo? | 09:54 |
ltomasbo | kzaitsev_ws, could be that it is not due to this patch | 09:57 |
ltomasbo | kzaitsev_ws, the exception is generated at the lbaas driver, but still it is when the handlers (this patch) calls it | 09:58 |
kzaitsev_ws | ltomasbo: hm. exception happens in LBaaSSpecHandler.on_present handler. the patch only adds a new handler to the pipeline _after_ the SpecHandler, and it doesn't seem to interact with any existing code | 10:05 |
kzaitsev_ws | janonymous: you need to apply the patch, re-install kuryr-k8s to apply new setuptools endpoint and have lbaas configured ) | 10:06 |
janonymous | I used vagrant :P | 10:06 |
kzaitsev_ws | janonymous: I don't think it changes any of the steps I described =) | 10:07 |
dmellado | heh | 10:08 |
dmellado | janonymous: of course if you just re-do your setup that will be done, but if you just do what kzaitsev_ws said it'll be quicker! | 10:08 |
ltomasbo | kzaitsev_ws, the error happens at /home/centos/kuryr-kubernetes/kuryr_kubernetes/controller/handlers/lbaas.py", line 441, in _sync_lbaas_loadbalancer | 10:08 |
janonymous | kzaitsev_ws: correct. so once that is done how exactly the load balancer functionality could be tested? | 10:08 |
janonymous | dmellado:heh | 10:08 |
ltomasbo | so, the new code does interact with the lbaas2 existing driver | 10:09 |
dmellado | ltomasbo: when does that happen? | 10:09 |
ltomasbo | https://review.openstack.org/#/c/376045/20/kuryr_kubernetes/controller/handlers/lbaas.py | 10:10 |
ltomasbo | dmellado, here at the LBaaSBalancerHandler | 10:10 |
kzaitsev_ws | janonymous: kubectl expose deployment deployment-name | 10:14 |
kzaitsev_ws | works for me ) | 10:14 |
janonymous | ohkay, got it | 10:14 |
janonymous | kzaitsev_ws: thanks | 10:15 |
apuimedo | dmellado: nope | 10:15 |
apuimedo | I will do so now | 10:15 |
dmellado | apuimedo: too bad! xD | 10:16 |
kzaitsev_ws | ltomasbo: oh, I see, the error is thrown from a different part then | 10:16 |
kzaitsev_ws | you're right | 10:16 |
*** pcaruana|afk| has joined #openstack-kuryr | 10:21 | |
openstackgerrit | Marco Chiappero proposed openstack/kuryr-kubernetes master: [WIP] Add MACVLAN based interfaces for nested containers https://review.openstack.org/440669 | 10:23 |
*** pcaruana|afk| has quit IRC | 10:23 | |
*** pcaruana has quit IRC | 10:24 | |
*** pcaruana has joined #openstack-kuryr | 10:25 | |
*** janki has quit IRC | 10:34 | |
*** jchhatbar has joined #openstack-kuryr | 10:34 | |
*** neiljerram has quit IRC | 11:13 | |
*** jchhatbar_ has joined #openstack-kuryr | 11:14 | |
*** neiljerram has joined #openstack-kuryr | 11:15 | |
*** jchhatbar_ is now known as janki | 11:15 | |
*** jchhatbar has quit IRC | 11:17 | |
*** mchiappero has joined #openstack-kuryr | 11:59 | |
apuimedo | ltomasbo: ping | 12:27 |
ltomasbo | apuimedo, pong | 12:27 |
ivc_ | vikasc ltomasbo i've addressed your comments | 12:42 |
ltomasbo | ivc_, regarding the comment about the if ... change = True | 12:43 |
ivc_ | ltomasbo yes? | 12:43 |
ltomasbo | why do you need all the ifs? | 12:43 |
ivc_ | because i need all 3 calls to be called independently and A() or B() or C() does not provide that | 12:44 |
ltomasbo | why is it different to do if (remove_unused_member or sync_lbaas or new_member) | 12:44 |
ivc_ | ltomasbo because in python (and pretty much any other language) B() would not get called if A() == True | 12:46 |
ltomasbo | ok | 12:47 |
ltomasbo | I just don't like all the ifs there | 12:47 |
ivc_ | it can be changed to any([A(), B(), C()]), but that does not look good either | 12:49 |
ivc_ | ltomasbo ^ | 12:49 |
ivc_ | more specifically, ifs are good for checking code coverage | 12:51 |
ltomasbo | and bad for performance | 12:52 |
ltomasbo | ivc_, to me any([A(), B(), C()]) feels like a better choice | 12:56 |
ltomasbo | it is pretty much reading what it is doing, if any of those, then true | 12:56 |
ivc_ | ltomasbo 'bad for performance' how? | 13:09 |
ltomasbo | it is not a big deal, but it impacts the performance as you (the cpu) cannot know for sure which one will be the next cpu instruction to execute | 13:12 |
ivc_ | ltomasbo we are talking about python here right? | 13:12 |
ltomasbo | so, it is usually recommended to limit the amount of ifs (when possible) | 13:12 |
ltomasbo | no no | 13:12 |
ltomasbo | no python, this is due to the way cpu instruction are executed, with segmentation and pre-feching and such | 13:13 |
ivc_ | man i know about cpu and all that | 13:13 |
ivc_ | but we are using python here, not c or asm | 13:13 |
ltomasbo | :D | 13:13 |
dmellado | pls don't tell me we'll be starting to see some JMP calls around kuryr | 13:15 |
dmellado | xD | 13:15 |
ivc_ | so 'if branching' overhead does not make much sense considering we are using interpreter | 13:15 |
ivc_ | ltomasbo also, i prefer those ifs because it highlights that all 3 calls are made regardless of others. and while any([]) is more compact, its more of a hack | 13:17 |
*** limao has joined #openstack-kuryr | 13:18 | |
ltomasbo | ivc_, to me it is easier to read it with the any, but that is my opinion. Nothing bad about the ifs approach. | 13:20 |
*** limao_ has joined #openstack-kuryr | 13:21 | |
ltomasbo | dmellado, so, no 'goto XXX' in the code?? xD | 13:23 |
dmellado | ltomasbo: that's a dangerous way you're going down xD | 13:23 |
*** jchhatbar has joined #openstack-kuryr | 13:23 | |
*** limao has quit IRC | 13:23 | |
ivc_ | dmellado hey goto is fine | 13:23 |
dmellado | another step, and we'll start numbering the code lines with | 13:24 |
dmellado | 10 | 13:24 |
dmellado | 20 | 13:24 |
dmellado | and so | 13:24 |
dmellado | xD | 13:24 |
*** jchhatbar_ has joined #openstack-kuryr | 13:24 | |
*** janki has quit IRC | 13:27 | |
apuimedo | dmellado: no. Kuryr is designed to run on Lisp machines only | 13:27 |
apuimedo | I'm working on a transpiler | 13:27 |
dmellado | apuimedo: how about a logo port? | 13:28 |
*** jchhatbar has quit IRC | 13:28 | |
*** jchhatbar has joined #openstack-kuryr | 13:32 | |
*** jchhatbar_ has quit IRC | 13:35 | |
*** janki has joined #openstack-kuryr | 13:35 | |
apuimedo | I had it at school | 13:36 |
*** jchhatbar has quit IRC | 13:38 | |
dmellado | me too | 13:40 |
* dmellado recalls the turtle | 13:40 | |
dmellado | rollcall, kuryr logo port | 13:40 |
dmellado | xd | 13:40 |
apuimedo | -1 | 13:42 |
dmellado | ohhhh | 13:44 |
*** jchhatbar has joined #openstack-kuryr | 13:46 | |
*** jchhatbar_ has joined #openstack-kuryr | 13:47 | |
*** janki has quit IRC | 13:49 | |
*** jchhatbar has quit IRC | 13:51 | |
*** hongbin has joined #openstack-kuryr | 14:06 | |
*** dimak_ has quit IRC | 14:08 | |
openstackgerrit | Marco Chiappero proposed openstack/kuryr-kubernetes master: Add MACVLAN based interfaces for nested containers https://review.openstack.org/440669 | 15:03 |
*** kzaitsev_ws has quit IRC | 15:12 | |
*** jchhatbar_ has quit IRC | 15:24 | |
*** rustlebee is now known as russellb | 15:29 | |
*** russellb has quit IRC | 15:29 | |
*** russellb has joined #openstack-kuryr | 15:29 | |
apuimedo | ltomasbo: this https://review.openstack.org/#/c/453215/1/kuryr_kubernetes/cni/api.py one liner is quite hacky | 15:31 |
apuimedo | to set an empty dict and access it all at once | 15:31 |
apuimedo | ltomasbo: dmellado: wouldn't it be better to just do result['dns'] = {'nameservers': nameservers} | 15:32 |
ltomasbo | umm, I got that recommendation in another patch, think it was ivc_ | 15:33 |
ltomasbo | waht if results['dns'] has something? | 15:33 |
mchiappero | folks, what should we do then with log messages and translations? pep8 complains... | 15:33 |
ltomasbo | it doesn't in this case | 15:34 |
ivc_ | ltomasbo why not get instead of setdefault? | 15:34 |
ivc_ | ltomasbo disregard that comment ^ | 15:35 |
ltomasbo | :D | 15:36 |
ivc_ | mchiappero we fix it so pep8 does not complain? *pokerface* | 15:37 |
ivc_ | ltomasbo for that setdefault, just do result['dns'] = {'nameservers': nameservers} | 15:39 |
ltomasbo | apuimedo, ivc_, I can do that (for now) as no one else is writing at dns in that method, but note that is not the correct way to do it (unless there is an if none then..) | 15:40 |
ltomasbo | as it will remove any data dns may have before that | 15:40 |
ivc_ | ltomasbo result has no 'dns' key | 15:40 |
ivc_ | and it is initialised right in that method | 15:41 |
apuimedo | ivc_: exactly my point | 15:41 |
mchiappero | ivc_: ? | 15:42 |
ivc_ | mchiappero LOG.warning(_LW(blablabla)) | 15:44 |
mchiappero | we had a chat some time ago and I've been told not to translate anymore | 15:45 |
ivc_ | uhm, why? | 15:45 |
mchiappero | I don't remember who was part of the conversation but the rationale was like making easier for people to look up solutions by having english only messages | 15:46 |
apuimedo | ivc_: it was dropped | 15:46 |
apuimedo | logs are not to be translated anymore in openstack | 15:47 |
ivc_ | then we need to update pep8 | 15:47 |
ivc_ | rules | 15:47 |
mchiappero | apuimedo: how to get the patches merged though since they fail on pep8? | 15:47 |
mchiappero | I was expecting the CI checks to be already up to date | 15:48 |
ivc_ | mchiappero just do _LW for now | 15:48 |
apuimedo | mmmm. DId we forget to merge the new hacking? | 15:48 |
mchiappero | ok, if you're okay with updating later I'll do | 15:48 |
*** aojea_ has quit IRC | 15:49 | |
ivc_ | we have those in other places anyway | 15:49 |
mchiappero | I mean, I'm ok with any option, just let me know what's the preferred approach currently :) | 15:49 |
mchiappero | there was a patch under review some time ago, since I wasn't sure which one would have been merged first I decided not to translate | 15:50 |
apuimedo | I'll check why it still wants it | 15:50 |
mchiappero | but no problem, I can revert them right away | 15:50 |
mchiappero | ok, let me know :) | 15:51 |
mchiappero | apuimedo: would it be ok to translate in order to get it merged and remove the translation later? We have both now :) | 16:10 |
apuimedo | yes | 16:10 |
apuimedo | it would be okay | 16:10 |
mchiappero | ok, so, don't forget to review the latest code | 16:10 |
mchiappero | we plan to follow up with some improvements on the parent detection, but if possible accept this one meanwhile | 16:11 |
mchiappero | thanks! | 16:11 |
mchiappero | :) | 16:12 |
apuimedo | very well | 16:14 |
ivc_ | vikasc ping | 16:27 |
ivc_ | mchiappero apuimedo btw we got https://review.openstack.org/#/c/450501/ for log i18n already :) | 16:42 |
*** limao_ has quit IRC | 16:48 | |
*** limao has joined #openstack-kuryr | 16:48 | |
openstackgerrit | Merged openstack/kuryr-kubernetes master: Fix vagrant ssh issue for virtualbox bento https://review.openstack.org/453445 | 16:49 |
mchiappero | ivc_: yeah I know, but 1) it doesn't include the new code of course 2) I'm not sure with how that doesn't fail | 16:52 |
ivc_ | mchiappero check tox.ini in that ps | 16:53 |
*** limao has quit IRC | 16:54 | |
mchiappero | oh yeah, sorry | 16:54 |
mchiappero | unless that gets +2 & workflown NOW, we will push translated logs | 16:55 |
mchiappero | and follow up later | 16:55 |
ivc_ | mchiappero you can probably also use that tox.ini change in you patch | 16:55 |
mchiappero | :D | 16:55 |
mchiappero | well it's not related to the patch itself so... | 16:56 |
ivc_ | up to you :) | 16:56 |
mchiappero | np, thanks for the help :) | 16:57 |
*** egonzalez has quit IRC | 17:03 | |
*** saneax has quit IRC | 17:04 | |
*** saneax has joined #openstack-kuryr | 17:11 | |
openstackgerrit | Gary Loughnane proposed openstack/kuryr-kubernetes master: Add MACVLAN based interfaces for nested containers https://review.openstack.org/440669 | 17:13 |
*** garyloug has quit IRC | 17:24 | |
*** tonanhngo has joined #openstack-kuryr | 17:47 | |
*** tonanhngo_ has joined #openstack-kuryr | 17:50 | |
*** tonanhngo has quit IRC | 17:51 | |
*** tonanhngo_ has quit IRC | 17:54 | |
*** tonanhngo has joined #openstack-kuryr | 17:55 | |
*** ltomasbo is now known as ltomasbo|away | 18:09 | |
*** dimak_ has joined #openstack-kuryr | 18:39 | |
*** dougbtv|laptop has joined #openstack-kuryr | 19:02 | |
*** dimak_ has quit IRC | 19:11 | |
*** dougbtv|laptop has quit IRC | 19:30 | |
*** oanson has quit IRC | 19:47 | |
*** irenab has quit IRC | 19:53 | |
*** lihi has quit IRC | 19:53 | |
*** oanson has joined #openstack-kuryr | 19:55 | |
*** lihi has joined #openstack-kuryr | 19:55 | |
*** irenab has joined #openstack-kuryr | 19:56 | |
*** pcaruana has quit IRC | 20:01 | |
*** aojea has joined #openstack-kuryr | 20:11 | |
*** lihi has quit IRC | 20:33 | |
*** lihi has joined #openstack-kuryr | 20:34 | |
apuimedo | ivc_: I think we can take https://review.openstack.org/#/c/450501/2 | 21:21 |
openstackgerrit | Merged openstack/kuryr-kubernetes master: Add devstack support for https k8s api server https://review.openstack.org/453138 | 21:33 |
*** aojea has quit IRC | 21:53 | |
*** pmannidi has joined #openstack-kuryr | 22:28 | |
*** pmannidi has quit IRC | 22:40 | |
*** pmannidi has joined #openstack-kuryr | 22:45 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!