| *** tonanhngo has joined #openstack-kuryr | 00:25 | |
| *** diogogmt has joined #openstack-kuryr | 00:27 | |
| *** tonanhngo has quit IRC | 00:29 | |
| *** irenab has joined #openstack-kuryr | 00:32 | |
| *** david-lyle_ is now known as david-lyle | 00:35 | |
| *** limao has joined #openstack-kuryr | 00:37 | |
| *** hongbin has quit IRC | 00:37 | |
| *** irenab has quit IRC | 00:37 | |
| *** tonanhngo has joined #openstack-kuryr | 00:45 | |
| *** tonanhngo has quit IRC | 00:49 | |
| *** irenab has joined #openstack-kuryr | 01:27 | |
| *** irenab has quit IRC | 01:32 | |
| limao | ping vikasc | 01:50 |
|---|---|---|
| *** irenab has joined #openstack-kuryr | 02:21 | |
| *** irenab has quit IRC | 02:26 | |
| *** tonanhngo has joined #openstack-kuryr | 02:34 | |
| *** tonanhngo has quit IRC | 02:38 | |
| vikasc | limao, pong | 03:07 |
| limao | vikasc: just want to confirm with you how did you do the devstack test? | 03:07 |
| limao | I tried without neutron_plugin in local, vagrant can't enable qos | 03:08 |
| limao | I also paste a doc there, please help to have a look when you have time | 03:09 |
| vikasc | limao, yes please share the doc | 03:09 |
| vikasc | limao, for my understanding i am digging into it | 03:10 |
| limao | https://www.openstack.org/assets/Uploads/QoS-Neutron-N00bie.pdf | 03:10 |
| limao | P32 | 03:10 |
| vikasc | limao, i tried yesterday without plugin but could not notice any error | 03:10 |
| limao | Yes, it will no error, but it will not enable qos | 03:10 |
| vikasc | limao, yes i can see, this document is mentioning | 03:10 |
| limao | can you check the /etc/neutron/neutron.conf | 03:11 |
| vikasc | ok | 03:11 |
| limao | did it enable qos service plugin? | 03:11 |
| * vikasc checkoing | 03:11 | |
| limao | here is correct service_plugin: | 03:12 |
| limao | service_plugins = qos,neutron.services.l3_router.l3_router_plugin.L3RouterPlugin | 03:12 |
| vikasc | limao, there is "#service_plugins =" no service plugin at all | 03:16 |
| *** irenab has joined #openstack-kuryr | 03:16 | |
| limao | ohh... I think at least this should be l3... | 03:16 |
| vikasc | limao, but in screens i can see that neutron services are running fine, like q-svc and all | 03:16 |
| limao | Did you use vagrant install? | 03:17 |
| vikasc | limao, everthing is commented in /etc/neutron/neutron.conf... | 03:17 |
| vikasc | limao, no vagrant | 03:17 |
| vikasc | limao, cloned devstack and copied localrc, and then stack.sh | 03:17 |
| limao | it should be same | 03:17 |
| limao | it is strange | 03:18 |
| vikasc | yeah, | 03:18 |
| vikasc | so you confirmed that without enabling neutron plugin, qos does not get added in neutron.conf? | 03:19 |
| limao | yes, I tried | 03:19 |
| vikasc | then it is ok to me.. i will remove -1, | 03:19 |
| limao | Well, you can try vagrant with my patch | 03:19 |
| limao | :) | 03:19 |
| vikasc | :) | 03:20 |
| limao | Thanks for your confirm, vikasc | 03:20 |
| vikasc | i just wanted to confirm | 03:20 |
| *** irenab has quit IRC | 03:20 | |
| vikasc | u confirmed so thank you!!! | 03:21 |
| *** yedongcan has joined #openstack-kuryr | 03:22 | |
| vikasc | limao, how devstack is running with neutron.conf totally commented still remains to be understood for me | 03:23 |
| limao | vikasc: yeah... at least need have mq, db info ;-) | 03:24 |
| vikasc | limao, yep, and on the nested_port patch of mine | 03:25 |
| vikasc | limao, that nested_port is vm_port and i dont think it is required in port_bind | 03:25 |
| openstackgerrit | Merged openstack/kuryr-libnetwork: Enable neutron qos in default vagrant devstack local.conf.sample https://review.openstack.org/397201 | 03:25 |
| vikasc | limao, do you see a problem in that change? | 03:26 |
| vikasc | limao, you removed your +1 :P | 03:26 |
| *** tonanhngo has joined #openstack-kuryr | 03:35 | |
| *** tonanhngo has quit IRC | 03:35 | |
| *** tonanhngo has joined #openstack-kuryr | 03:44 | |
| *** tonanhngo has quit IRC | 03:48 | |
| limao | vikasc: I just want to be clear about which "port param" is for container, and if we do not use vm port in port_binding, we may remove that param. that's why remove +1 there. | 03:53 |
| openstackgerrit | Zhigang Li proposed openstack/kuryr: getting random string from neutron-lib helpers that have supported https://review.openstack.org/398732 | 03:55 |
| vikasc | limao, if you see veth_driver, this nested_port is not being used and veth driver is default driver..from this can we infer that 'port' param is container port. Because IP for veth is also being extracted from 'port'. | 03:55 |
| vikasc | limao, i would appreciate a '-1' or '+1' :) | 03:56 |
| limao | vikasc: yes, I think so, but toni's comments confused me "That was the intended naming indeed. nested port is the one for the container, port is the one the VM. " | 03:57 |
| vikasc | limao, i asked apuimedo yesterday, lets request again today to clarify. | 03:58 |
| limao | vikasc : Yeah, if here is clear, this patch +1 from my point of view. (maybe need another follow up patch to remove nested_port if no need to use in bind_port) | 03:59 |
| vikasc | limao, or if we all agrees that nested-port is not needed, i would update same patch to remove it | 04:01 |
| limao | vikasc: yeah, agree | 04:02 |
| vikasc | limao, lets discuss with apuimedo | 04:02 |
| limao | sure :-) | 04:02 |
| janonymous | Hello, can some core members can have a look on https://review.openstack.org/#/q/status:open+project:openstack/kuryr-libnetwork+branch:master+topic:bp/drop-mox , to avoid merge conflicts and rebases | 04:19 |
| openstackgerrit | Zhigang Li proposed openstack/kuryr: getting random string from neutron-lib helpers that have supported https://review.openstack.org/398732 | 04:24 |
| *** tonanhngo has joined #openstack-kuryr | 04:55 | |
| *** tonanhngo has quit IRC | 04:56 | |
| yedongcan | limao: ping | 05:06 |
| limao | yedongcan: hello | 05:06 |
| yedongcan | Do you still working on this patch: https://review.openstack.org/#/c/387835/ ? | 05:08 |
| *** irenab has joined #openstack-kuryr | 05:09 | |
| limao | yedongcan: yeah, stopped for some time... ut need to fix in my mind | 05:09 |
| limao | yedongcan: so you want to do something related? | 05:10 |
| yedongcan | limao: Ok, I look it just now. The cause it we're missing subnet_pool_name in Docker option | 05:11 |
| yedongcan | s/it/is | 05:12 |
| yedongcan | limao: So we're missing subnetpool_id in create subnet | 05:13 |
| limao | yedongcan : If we create a kuryr network, it will create a subnetpool and subnet, right? | 05:13 |
| *** irenab has quit IRC | 05:14 | |
| yedongcan | limao: Yes, but we can't get subnetpool_id, if we are not passing subnet pool name. | 05:14 |
| yedongcan | limao: I had a related patch can fix it, https://review.openstack.org/#/c/373977/4 | 05:15 |
| limao | yedongcan: cool , let me check there | 05:15 |
| yedongcan | limao: Thanks, but I had already forgot why I changed this at the end, I need to look in detail. | 05:17 |
| limao | yedongcan: if we do not assign pool_name in option, we can find the pool_id from pool_cidr . | 05:20 |
| yedongcan | limao: Yes | 05:21 |
| limao | yedongcan: In this case, the subnet can still get the right subnetpool id | 05:21 |
| yedongcan | limao: Right. | 05:22 |
| limao | yedongcan: thanks for your reminder, I will update the patch today or tomorrow | 05:23 |
| yedongcan | limao: OK, thanks. | 05:25 |
| *** irenab has joined #openstack-kuryr | 06:03 | |
| openstackgerrit | Dongcan Ye proposed openstack/kuryr-libnetwork: Update neutron-lib and import_exceptions https://review.openstack.org/398756 | 06:04 |
| *** limao has quit IRC | 06:06 | |
| *** irenab has quit IRC | 06:08 | |
| *** limao has joined #openstack-kuryr | 06:08 | |
| *** tonanhngo has joined #openstack-kuryr | 06:08 | |
| *** tonanhngo has quit IRC | 06:13 | |
| *** irenab has joined #openstack-kuryr | 06:31 | |
| *** yedongcan has quit IRC | 06:47 | |
| *** irenab_ has joined #openstack-kuryr | 06:58 | |
| *** irenab_ has quit IRC | 07:02 | |
| *** yedongcan has joined #openstack-kuryr | 07:04 | |
| openstackgerrit | Zhigang Li proposed openstack/kuryr: getting random string from neutron-lib helpers that have supported https://review.openstack.org/398732 | 07:12 |
| openstackgerrit | Zhigang Li proposed openstack/kuryr: getting random string from neutron-lib helpers that have supported https://review.openstack.org/398732 | 07:29 |
| *** oanson has joined #openstack-kuryr | 07:33 | |
| *** tonanhngo has joined #openstack-kuryr | 07:35 | |
| *** tonanhngo has quit IRC | 07:39 | |
| openstackgerrit | Dongcan Ye proposed openstack/kuryr-libnetwork: Using import_exceptions from kuryr-lib https://review.openstack.org/398756 | 07:48 |
| *** irenab_ has joined #openstack-kuryr | 07:52 | |
| *** irenab_ has quit IRC | 07:56 | |
| ivc_ | irenab, vikasc, i've replied to your comments on https://review.openstack.org/#/c/398324/2 | 08:05 |
| vikasc | thanks ivc_ , will take a look | 08:07 |
| *** dimak has joined #openstack-kuryr | 08:12 | |
| ivc_ | irenab, do we really need a TODO for unit tests if we already have one for the tested module? imo that only adds confusion (i.e. consider that someone reads the note in test module and tries to move the unit tests without moving the module itself). and there is no way that we move the os_vif_utils without updating the tests accordingly as the tests would simply fail without the module itself | 08:12 |
| irenab | ivc_, I think you can also leave it as is, its just more handy if soemone else will update it later | 08:13 |
| ivc_ | irenab, its a trivial change and i don't mind adding a note tbh. just wanted to share my opinion that i consider it redundant :) | 08:16 |
| irenab | ivc_, can you please share your approach for the drivers granularity, I am not sure I got your idea correct | 08:17 |
| ivc_ | sure | 08:17 |
| ivc_ | we'll have ProjectDriver(s) to take care of multitenancy in the future. or we could implement a driver that would take a project_id from some source other than the config file (e.g. annotation or etcd) | 08:18 |
| ivc_ | then we have SubnetDriver that does the same for subnets | 08:18 |
| ivc_ | then we have SecurityGroupsDriver(s) | 08:19 |
| ivc_ | for the same reason | 08:19 |
| *** gsagie has joined #openstack-kuryr | 08:19 | |
| ivc_ | then there would be VIFDriver that would be slightly different and will deal with different deployment options (i.e. one for bare-metal and another for container-in-vm) | 08:20 |
| apuimedo | morning all | 08:20 |
| ivc_ | apuimedo, hey toni :) | 08:20 |
| apuimedo | I would need to know who would be able to attend the Atlanta PTG | 08:21 |
| apuimedo | ivc_: irenab: vikasc: janonymous: limao: yedongcan: ltomasbo: | 08:21 |
| apuimedo | to decide if we book a space or not | 08:21 |
| apuimedo | otherwise we'll try to make some virtual one | 08:22 |
| irenab | apuimedo, +1 on virtuyal one | 08:22 |
| limao | apuimedo: currently, I do not have plan | 08:22 |
| gsagie | irenab: :) you need a trip.. | 08:22 |
| * vikasc reading back | 08:22 | |
| ivc_ | as part of VIFDriver i'll probably make some sort of NeutronToVIFConverter (it would not be in controller.drivers) to deal with 'bridge', 'ovs' and other vif types in os_vif_utils (it would serve the same role as https://github.com/openstack/nova/blob/master/nova/network/os_vif_util.py#L362, but would make it possible to have out-of-tree projects to provide their own drivers, think 'midokura' or smth) | 08:22 |
| *** yuval has joined #openstack-kuryr | 08:23 | |
| irenab | gsagie, nice to hear from you | 08:23 |
| vikasc | neither me | 08:23 |
| ivc_ | apuimedo, +1 on virtual too | 08:23 |
| vikasc | +! for virtual one | 08:24 |
| irenab | gsagie, planning to go to PTG? | 08:24 |
| vikasc | +1 | 08:24 |
| gsagie | irenab: only if you come :) | 08:24 |
| apuimedo | alright then. I'll make it a virtual one | 08:24 |
| irenab | gsagie, join us on the virtual one | 08:24 |
| apuimedo | but there'll be presentations and schedule nonetheless | 08:24 |
| apuimedo | ;-) | 08:24 |
| irenab | apuimedo, I could consider BCN :-) | 08:24 |
| apuimedo | irenab: my doors are always open to kuryr contributors | 08:25 |
| apuimedo | I have some bunk beds | 08:25 |
| apuimedo | :P | 08:25 |
| vikasc | super | 08:25 |
| vikasc | :) | 08:25 |
| gsagie | apuimedo: watch out with what you promise... ;) | 08:25 |
| gsagie | I really liked the casino in Barcelona | 08:25 |
| irenab | gsagie, apuimedo needs babysitters | 08:25 |
| vikasc | :D | 08:26 |
| vikasc | thats the hidden plan | 08:26 |
| apuimedo | gsagie: you'd still be 40min away from barcelona | 08:26 |
| apuimedo | irenab: you got that right | 08:26 |
| apuimedo | it's the fee | 08:26 |
| irenab | :-) | 08:27 |
| yuval | gsagie: galllllllll | 08:28 |
| gsagie | yuvalllllll | 08:29 |
| irenab | ivc_, reading the explanation | 08:29 |
| yuval | gsagie: :) | 08:29 |
| ivc_ | irenab, the idea is to make it possible to combine different drivers without tightly coupling them | 08:30 |
| limao | BTW, | 08:30 |
| limao | apuimedo : If you need a webex meeting for virtual meeting , I can offer (It can free call back audio) | 08:31 |
| irenab | ivc_, so your idea is that there is driver per neutron entity that should be managed? | 08:31 |
| ivc_ | irenab, sort of | 08:32 |
| irenab | ivc_, lets say there is Pod creation via k8s api, what will be the workflow to reflect it in neutron? | 08:33 |
| ivc_ | irenab, take PodSubnetDriver for example. it takes pod object. there will also be a ServiceSubnet driver that takes service | 08:33 |
| irenab | ivc_, so it k8s object_neutro object drivers | 08:34 |
| ivc_ | irenab, VIFHandler takes care of managing neutron resources for Pod creation | 08:34 |
| irenab | ivc_, sorry, still do not see the whole picture ... VIFHandler is for the local stuff, isn't it? | 08:35 |
| ivc_ | irenab, i'll update VIFHandler patch to use drivers instead of _* methods | 08:35 |
| ivc_ | irenab, what do you mean by 'local stuff'? | 08:35 |
| irenab | worker node bindings | 08:35 |
| *** reedip has quit IRC | 08:36 | |
| ivc_ | no VIFHandler is for the controller. it takes care of managing neutron resources and updates pod annotation with VIF object that is later used for binding on worker nodes | 08:36 |
| apuimedo | I was gonna use bluejeans | 08:36 |
| irenab | ivc_, maybe the name (VIF) is a bit misleading | 08:38 |
| janonymous | apuimedo: +1 for both options :) | 08:38 |
| *** reedip has joined #openstack-kuryr | 08:38 | |
| irenab | apuimedo, as long as no travel involved, any options is fine | 08:38 |
| ivc_ | irenab, but it is what it does - it defines the os-vif VIF object | 08:38 |
| ivc_ | irenab, and it is the controller-side part of VIF binding process | 08:39 |
| irenab | ivc_, what would you have for Service creation? | 08:39 |
| vikasc | ivc_, but it handles pod create events | 08:39 |
| vikasc | ivc_, cant it be named PodCreateHandler | 08:39 |
| vikasc | ivc_, or PodHandler | 08:39 |
| ivc_ | vikasc, no. because we can have some other PodHandler that does something completely different than managing VIFs/ports | 08:40 |
| ivc_ | vikasc, irenab, i originally named it PodPortHandler, but since for the outside world it creates the VIF annotation and Neutron port is just an implementation detail, i've renamed it to VIFHandler | 08:41 |
| irenab | ivc_, so what you have in mind is that the Handler named upon what it responsible to create? | 08:42 |
| ivc_ | irenab, aye | 08:43 |
| irenab | so the ServiceHandler will be actually LoadBalancerHandler? | 08:43 |
| ivc_ | i guess so | 08:43 |
| irenab | ivc_, sorry, just trying to understant the intention | 08:43 |
| ivc_ | irenab, well at some point we decided that it would be nice to have multiple handlers for the same event, right? | 08:44 |
| ivc_ | so e.g. for pod.ADDED we have one that takes care of networking | 08:44 |
| ivc_ | but then there were plans to do something for volumes | 08:45 |
| ivc_ | so naming it PodHandler does not feel right imo | 08:45 |
| *** irenab_ has joined #openstack-kuryr | 08:46 | |
| ivc_ | and we have OBJECT_KIND = 'Pod' that makes it clear what sort of object it handles | 08:46 |
| irenab | ivc_, I agree on the last statement, just not sure if the 'end result' naming is the way to go, it makes me more confused | 08:46 |
| irenab | ivc_, but I like the 'miscroservice' like apprpach | 08:47 |
| ivc_ | irenab, thats probably because you got used to midokura's poc approach :) | 08:47 |
| irenab | ivc_, 'deeply involved' :-) | 08:47 |
| vikasc | ivc_, makes sense | 08:48 |
| ivc_ | vikasc, thnx :) | 08:48 |
| irenab | ivc_, but seriously, I am missing the overall design approach understanding | 08:48 |
| vikasc | ivc_, thanks to you !! | 08:48 |
| vikasc | :) | 08:48 |
| vikasc | irenab, ivc_ i am also not clear.. trying to undrstand just by reviewing patches | 08:49 |
| irenab | ivc_, maybe you can push devref with the general idea, so we can start explore for next constructs | 08:49 |
| irenab | and be more helpful on reviews | 08:49 |
| ivc_ | irenab, vikasc, well i'd be happy to clarify | 08:49 |
| irenab | lets say I am about to add support for k8s net policy | 08:50 |
| irenab | would it be SGHandler? | 08:50 |
| *** irenab_ has quit IRC | 08:50 | |
| irenab | or it will use SGDriver? | 08:50 |
| ivc_ | irenab, the problem is that this thing is evolving as it goes. think agile rather than waterfall :) i only came with that 'driver' approach like last week | 08:51 |
| ivc_ | irenab, SGDriver | 08:51 |
| ivc_ | it will be used by VIFHandler | 08:51 |
| vikasc | ivc_, we would be happy to maintain devref also in agile approach :) | 08:51 |
| irenab | ivc_, I am all into agile approach, but some fundamental principals should be in place :-), can be refactored of course | 08:51 |
| vikasc | ivc_, atleast with some minimum placeholders | 08:52 |
| vikasc | ivc_, and then everybody from the team will keep updating as per evolution and understanding and you will be there to clarify | 08:53 |
| vikasc | ivc_, s/placeholders/headings | 08:54 |
| ivc_ | irenab, vikasc, well i totally agree that we need a devref | 08:56 |
| ivc_ | i've even started one here: https://review.openstack.org/#/c/369105/ | 08:56 |
| ivc_ | but its outdated | 08:56 |
| ivc_ | and incomplete | 08:57 |
| ivc_ | and tbh i'm not sure what exactly needs to be explained (i.e. what questions do you have and what is not clear from code) | 08:58 |
| irenab | ivc_, can you please update there the Handler/Driver approach as you now taking? | 08:58 |
| vikasc | that may help some other new developers as well | 08:59 |
| irenab | vikasc, +1 | 08:59 |
| vikasc | beginners | 08:59 |
| vikasc | and then they will also be able to help | 08:59 |
| irenab | ivc_, lets say someone is about to add support for k8s net policy, so it will help to understand that pices to add or to integrate with | 09:00 |
| irenab | s/pices/parts | 09:00 |
| ivc_ | irenab, the problem is there's nothing to integrate with now | 09:00 |
| irenab | ivc_, design wise | 09:00 |
| ivc_ | irenab, can we turn it into some sort of Q&A? i honestly can't predict what needs to be explained | 09:02 |
| irenab | ivc_, sure, let take the agile approach and improve as we go | 09:03 |
| ivc_ | and then there's another problem that writing a devref takes tons of time (at least for me) and i sort of have a deadline to implement vif and lbaas (pod/service) handling | 09:03 |
| ivc_ | and that deadline is pretty close | 09:03 |
| vikasc | :) | 09:03 |
| vikasc | cool | 09:04 |
| *** yuval has left #openstack-kuryr | 09:04 | |
| irenab | ivc_, for me the sw workflow you have in mind for Pod/Service creation will be enough to be more helpful on reviewing | 09:05 |
| vikasc | ivc_, please keep updating patches | 09:05 |
| ivc_ | irenab, 'sw workflow'? | 09:06 |
| irenab | ivc_, sequence of calls | 09:06 |
| irenab | ivc_, I can help with devref if you wish | 09:07 |
| ivc_ | irenab, that'd be awesome :) | 09:07 |
| vikasc | great | 09:07 |
| irenab | ivc_, but I have to understand what you have in mind :-) | 09:07 |
| ivc_ | irenab, sure :) | 09:08 |
| *** garyloug has joined #openstack-kuryr | 09:33 | |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes: devstack: make cni paths configurable https://review.openstack.org/396113 | 09:35 |
| apuimedo | vikasc: ^^ | 09:36 |
| vikasc | apuimedo, can you please review my patches | 09:40 |
| *** irenab_ has joined #openstack-kuryr | 09:40 | |
| apuimedo | vikasc: I will | 09:40 |
| vikasc | apuimedo, https://review.openstack.org/#/c/397057/ i dont this nested_port(which is vm-port IMO) is needed in port_bind. | 09:41 |
| apuimedo | vikasc: can you apply the change I ask in https://review.openstack.org/#/c/397057/2 | 09:42 |
| ivc_ | vikasc, apuimedo, wait! xD | 09:42 |
| apuimedo | vikasc: rename it to vmport | 09:42 |
| apuimedo | ivc_: ? | 09:42 |
| ivc_ | dont merge loopback.conf :) | 09:43 |
| ivc_ | i'm writing a comment | 09:43 |
| ivc_ | done :) | 09:43 |
| vikasc | apuimedo, why do we need vm-port in port-bind? | 09:43 |
| *** yedongcan has quit IRC | 09:44 | |
| *** irenab_ has quit IRC | 09:45 | |
| *** yedongcan has joined #openstack-kuryr | 09:45 | |
| apuimedo | vikasc: to, for instance, find the device which the port should be linked to | 09:47 |
| apuimedo | if you have the vmport info you can match by UP | 09:47 |
| apuimedo | *IP | 09:47 |
| vikasc | apuimedo, that info will be needed in kuryr-libnetwork or kuryr-kubernetes controller i think | 09:48 |
| *** openstackgerrit has quit IRC | 09:48 | |
| vikasc | apuimedo, which driver will use this? | 09:48 |
| *** openstackgerrit has joined #openstack-kuryr | 09:48 | |
| apuimedo | ivc_: that's an interesting comment. I saw some examples with multiple conf for multiple drivers. I'll check the code to make sure the documentation is up to date | 09:49 |
| vikasc | apuimedo, currently no driver seem to be using it | 09:49 |
| ivc_ | apuimedo, ok | 09:49 |
| *** yedongcan has quit IRC | 09:49 | |
| apuimedo | vikasc: no driver is using this because we hardcode or have it configurable in kuryr.conf (something like bindinglink | 09:49 |
| apuimedo | binding_link_device) | 09:49 |
| *** yedongcan has joined #openstack-kuryr | 09:50 | |
| apuimedo | ivc_: hyperkube official image does have two configs | 09:52 |
| apuimedo | 99-loopback.conf | 09:52 |
| apuimedo | and | 09:52 |
| apuimedo | 10-containernet.conf | 09:53 |
| ivc_ | apuimedo, i like that 10-/99- approach | 09:53 |
| ivc_ | even if it will never read loopback.conf, we make the order explicit | 09:53 |
| apuimedo | ivc_: and it does look at all the conf files | 09:54 |
| apuimedo | ah no, sorry | 09:55 |
| apuimedo | there is a return | 09:55 |
| apuimedo | this darned habit of putting returns instead of breaks... | 09:55 |
| ivc_ | apuimedo, then it'll need some selector to pick the right driver | 09:55 |
| ivc_ | which is not there (yet) | 09:55 |
| apuimedo | in pkg/kubelet/network/cni/cni.go:112 | 09:55 |
| apuimedo | it does return after the first valid entry | 09:55 |
| apuimedo | so I say we keep both, but with the 10- 99- approach | 09:56 |
| ivc_ | apuimedo, i think they'll eventually add an option like --cni-driver or something | 09:56 |
| ivc_ | aye | 09:56 |
| apuimedo | very well | 09:56 |
| apuimedo | I'll update the patch | 09:56 |
| *** lmdaly has joined #openstack-kuryr | 09:56 | |
| *** garyloug has quit IRC | 09:56 | |
| ivc_ | right now its just pure luck our project name starts with "K" :) | 09:56 |
| apuimedo | :P | 09:57 |
| apuimedo | it's not luck, it is destiny | 09:57 |
| ivc_ | haha, i guess you envisioned that problem when you were considering the name :) | 09:58 |
| apuimedo | xD | 09:58 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes: devstack: make cni paths configurable https://review.openstack.org/396113 | 10:02 |
| apuimedo | ivc_: there | 10:02 |
| apuimedo | vikasc: ^^ | 10:02 |
| vikasc | apuimedo, so intention to passing vmport in port_bind is that some coe might pass vm_port instead of mentioning in config file? | 10:02 |
| ivc_ | apuimedo, thnx | 10:02 |
| apuimedo | vikasc: right | 10:02 |
| apuimedo | it is a nullable parameter | 10:02 |
| vikasc | apuimedo, or even in kuryr-libnetwork eventually we will be able to skip config file part | 10:02 |
| apuimedo | exactly | 10:03 |
| apuimedo | if we can do without config, better | 10:03 |
| vikasc | apuimedo, yep, cool | 10:03 |
| ivc_ | apuimedo, https://review.openstack.org/#/c/396113/3/devstack/plugin.sh@275 | 10:04 |
| ivc_ | does it work as you expected now without "/" between ${KURYR_HOME}${kubelet_plugin_dir} | 10:05 |
| apuimedo | heh. It won't | 10:05 |
| apuimedo | I was now restacking | 10:05 |
| apuimedo | xD | 10:05 |
| ivc_ | also | 10:05 |
| ivc_ | you state in commit message it is /opt/stack/cni/bin | 10:06 |
| ivc_ | but it is not :P | 10:06 |
| ivc_ | i was too fast about +1 xD | 10:06 |
| apuimedo | ivc_: the one we use in devstack is /opt/stack/cni/bin | 10:06 |
| ivc_ | s@/opt/stack/cni/bin@/opt/stack/cni/conf@ | 10:07 |
| apuimedo | CNI_CONF_DIR | 10:07 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes: devstack: make cni paths configurable https://review.openstack.org/396113 | 10:08 |
| ivc_ | right, my bad | 10:08 |
| apuimedo | ivc_: https://review.openstack.org/#/c/396113/3..4/devstack/plugin.sh | 10:08 |
| apuimedo | that was the mistake. Well spotted | 10:08 |
| * apuimedo restacking | 10:08 | |
| ivc_ | apuimedo, i wonder if we can have some tests to validate devstack setup on gates | 10:14 |
| ivc_ | not in this patch ofc | 10:14 |
| openstackgerrit | Ilya Chukhnakov proposed openstack/kuryr-kubernetes: Default pod subnet driver and os-vif utils https://review.openstack.org/398324 | 10:17 |
| ivc_ | vikasc, irenab, ^^ https://review.openstack.org/#/c/398324/2..3 updated commit message and todo for tests | 10:18 |
| *** yedongcan has quit IRC | 10:20 | |
| vikasc | ivc_, thanks !! | 10:21 |
| apuimedo | ivc_: yes. We should | 10:22 |
| *** limao has quit IRC | 10:24 | |
| *** yedongcan has joined #openstack-kuryr | 10:24 | |
| openstackgerrit | vikas choudhary proposed openstack/kuryr: Fix container port ipaddress setting in ipvlan/macvlan drivers https://review.openstack.org/397057 | 10:29 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes: devstack: make cni paths configurable https://review.openstack.org/396113 | 10:30 |
| *** yedongcan has left #openstack-kuryr | 10:38 | |
| ivc_ | vikasc, i've noticed your request about config.py tests in some of my patchsets. do you mind if we add unit tests for config.py in a later patch after VIFHandler? or do you want it asap (then we need another patch that would add the testing facility and rebase everything on it)? | 10:39 |
| vikasc | ivc_, i am fine either way.. may be we can open a bug for the sake of tracking | 10:41 |
| vikasc | ivc_, in case you prefer in a later patch | 10:41 |
| ivc_ | vikasc, i'd prefer a bug. it would be a low-hanging-fruit one and might attract some other contributors :) | 10:44 |
| ivc_ | vikasc, my point is that it is non-essential and if we add it as a dependency for the drivers/VIFHandler it does not help us merging those - and i want them done asap along with the CNI driver so ppl can actually run our devstack setup | 10:48 |
| vikasc | ivc_, sounds reasonable | 10:49 |
| ivc_ | https://bugs.launchpad.net/kuryr-kubernetes/+bug/1640446 - here's an example of the problem that we have now | 10:50 |
| openstack | Launchpad bug 1640446 in kuryr-kubernetes "Kuryr-kubernetes services throws Dummy exception while creating a POD" [Undecided,In progress] - Assigned to Ilya Chukhnakov (ichukhnakov) | 10:50 |
| *** gsagie has quit IRC | 11:14 | |
| *** lmdaly has quit IRC | 11:15 | |
| *** irenab_ has joined #openstack-kuryr | 11:28 | |
| *** irenab_ has quit IRC | 11:33 | |
| *** garyloug has joined #openstack-kuryr | 11:54 | |
| *** reedip has quit IRC | 12:19 | |
| *** irenab_ has joined #openstack-kuryr | 12:22 | |
| apuimedo | ivc_: we can mark it as closed, not a bug, since with dummy it is expected | 12:23 |
| ivc_ | apuimedo, i don't mind and it was my first intention to do so | 12:25 |
| apuimedo | ;-) | 12:25 |
| ivc_ | but it's getting fixed by the linked patch so i thought it would make sense to close it 'properly' :) | 12:27 |
| *** irenab_ has quit IRC | 12:27 | |
| * ivc_ hates 'not-a-bug' resolutions | 12:27 | |
| *** reedip has joined #openstack-kuryr | 12:31 | |
| openstackgerrit | Gary Loughnane proposed openstack/kuryr-libnetwork: Fix default subnetpool assignmens from config file https://review.openstack.org/397640 | 12:31 |
| *** garyloug has quit IRC | 13:16 | |
| *** irenab_ has joined #openstack-kuryr | 13:17 | |
| *** lmdaly has joined #openstack-kuryr | 13:20 | |
| *** irenab_ has quit IRC | 13:21 | |
| *** lmdaly has quit IRC | 13:21 | |
| *** lmdaly has joined #openstack-kuryr | 13:21 | |
| apuimedo | ivc_: agreed | 13:26 |
| *** lmdaly has quit IRC | 13:26 | |
| *** lmdaly has joined #openstack-kuryr | 13:27 | |
| openstackgerrit | vikas choudhary proposed openstack/kuryr: Fix container port ipaddress setting in ipvlan/macvlan drivers https://review.openstack.org/397057 | 13:30 |
| *** limao has joined #openstack-kuryr | 13:56 | |
| *** diogogmt has quit IRC | 13:58 | |
| *** irenab has quit IRC | 14:09 | |
| *** irenab has joined #openstack-kuryr | 14:11 | |
| *** oanson has quit IRC | 14:13 | |
| *** irenab has quit IRC | 14:15 | |
| *** janonymous has quit IRC | 14:34 | |
| *** limao has quit IRC | 14:47 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-libnetwork: Updated from global requirements https://review.openstack.org/351976 | 15:01 |
| *** yamamoto has joined #openstack-kuryr | 15:03 | |
| *** irenab has joined #openstack-kuryr | 15:05 | |
| *** hongbin has joined #openstack-kuryr | 15:08 | |
| *** irenab has quit IRC | 15:09 | |
| *** garyloug has joined #openstack-kuryr | 15:12 | |
| *** janki has joined #openstack-kuryr | 15:22 | |
| *** diogogmt has joined #openstack-kuryr | 15:46 | |
| *** janonymous has joined #openstack-kuryr | 16:07 | |
| *** yamamoto has quit IRC | 16:34 | |
| *** yamamoto has joined #openstack-kuryr | 16:43 | |
| *** tonanhngo has joined #openstack-kuryr | 16:46 | |
| *** dimak has quit IRC | 16:46 | |
| *** yamamoto has quit IRC | 16:48 | |
| *** garyloug has quit IRC | 17:03 | |
| *** janki has quit IRC | 17:04 | |
| *** tonanhngo has quit IRC | 17:27 | |
| *** lmdaly has quit IRC | 17:46 | |
| *** tonanhngo has joined #openstack-kuryr | 18:07 | |
| *** tonanhngo_ has joined #openstack-kuryr | 18:10 | |
| *** tonanhngo_ has quit IRC | 18:11 | |
| *** tonanhngo has quit IRC | 18:11 | |
| *** tonanhngo has joined #openstack-kuryr | 18:11 | |
| *** yamamoto has joined #openstack-kuryr | 18:13 | |
| *** tonanhngo has quit IRC | 18:13 | |
| *** tonanhngo has joined #openstack-kuryr | 18:15 | |
| *** tonanhngo has quit IRC | 18:17 | |
| *** yamamoto has quit IRC | 18:19 | |
| *** irenab has joined #openstack-kuryr | 18:25 | |
| *** irenab has quit IRC | 18:29 | |
| *** diogogmt has quit IRC | 18:55 | |
| *** diogogmt has joined #openstack-kuryr | 18:57 | |
| *** irenab has joined #openstack-kuryr | 18:58 | |
| *** diogogmt has quit IRC | 19:02 | |
| *** irenab has quit IRC | 19:03 | |
| *** dimak has joined #openstack-kuryr | 19:19 | |
| *** tonanhngo has joined #openstack-kuryr | 19:19 | |
| *** dimak_ has joined #openstack-kuryr | 19:37 | |
| *** dimak_ has quit IRC | 19:39 | |
| *** openstackgerrit has quit IRC | 19:48 | |
| *** openstackgerrit has joined #openstack-kuryr | 19:48 | |
| *** irenab has joined #openstack-kuryr | 19:52 | |
| *** irenab has quit IRC | 19:57 | |
| *** irenab has joined #openstack-kuryr | 20:03 | |
| *** irenab has quit IRC | 20:07 | |
| *** irenab has joined #openstack-kuryr | 20:48 | |
| *** irenab has quit IRC | 20:52 | |
| *** dimak has quit IRC | 21:30 | |
| *** irenab has joined #openstack-kuryr | 21:41 | |
| *** irenab has quit IRC | 21:46 | |
| *** irenab has joined #openstack-kuryr | 22:36 | |
| *** irenab has quit IRC | 22:40 | |
| openstackgerrit | Hongbin Lu proposed openstack/fuxi: Fix the .gitignore file https://review.openstack.org/399282 | 22:41 |
| *** irenab has joined #openstack-kuryr | 23:30 | |
| *** irenab has quit IRC | 23:34 | |
| *** irenab has joined #openstack-kuryr | 23:35 | |
| *** irenab has quit IRC | 23:40 | |
| openstackgerrit | Hongbin Lu proposed openstack/fuxi: Add cffi to requirement.txt https://review.openstack.org/399295 | 23:42 |
| openstackgerrit | Hongbin Lu proposed openstack/fuxi: [WIP] Fix the installation guide https://review.openstack.org/399296 | 23:45 |
| *** hongbin has quit IRC | 23:46 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!