| vikasc | ivc_, pping | 00:00 |
|---|---|---|
| ivc_ | vikasc pong | 00:00 |
| vikasc | ivc_, hi | 00:01 |
| ivc_ | vikasc hey :) | 00:01 |
| vikasc | ivc_, This one is not 'noop'. I think it would be 'vlan_subport'. | 00:01 |
| ivc_ | vikasc right | 00:02 |
| ivc_ | vikasc you got noop there - one for os-vif plugin and another for vif_translators | 00:03 |
| ivc_ | vikasc but for vif_translators you need a real thing that will convert neutron port dict into the VIF object | 00:04 |
| vikasc | ivc_, yep in the noop translator, i will create vif object reading subport type from config file , like 'vlan_subport' | 00:04 |
| vikasc | ivc_, like in ovs translator, ovs is being set as vif.plugin, similarly here i am thinking to set 'noop' as vif.plugin | 00:05 |
| ivc_ | vikasc in ovs translator it's just a for convenience that vif_plugin in translator == vif.plugin. i already figured now that it might be confusing :) | 00:07 |
| vikasc | ivc_, "need a real thing" is needed for deciding vif type or something else? | 00:09 |
| ivc_ | vikasc i think you can skip 'translators' part for now and just do the translation in Nested VIFDriver itself | 00:10 |
| vikasc | ivc_, this translator name is noop because it will set vif-plugin as 'noop' | 00:10 |
| vikasc | ivc_, why to skip? | 00:11 |
| ivc_ | vikasc yeah, but vif_translator name is not vif.plugin; they are equal for 'ovs' but don't have to be | 00:11 |
| ivc_ | vikasc we can move the translating code from Nested VIFDriver to the translator in a follow up | 00:12 |
| vikasc | ivc_, so may be 'nested_vif'? | 00:12 |
| ivc_ | vikasc mostly because i have not researched what content is in sub-port dict returned by neutron client and if it will be enough to construct VIF object | 00:13 |
| ivc_ | ^ thats about 'why skip' | 00:13 |
| vikasc | ivc_, sub-port is same as neutron normal port | 00:14 |
| ivc_ | vikasc no additional 'vlan' info there? | 00:14 |
| vikasc | ivc_, no | 00:14 |
| vikasc | ivc_, vlan kuryr-k8s has to allocate | 00:14 |
| vikasc | ivc_, we have logic for that already in kuryr-lib | 00:15 |
| ivc_ | vikasc i mean that in VIF object you will have 'vlan_id' field | 00:15 |
| vikasc | ivc_, while requesting neutron to add this port as a subport, this sel allocated vlan_id will be passed to neutron | 00:15 |
| ivc_ | vikasc thats another reason why vif_translator won't work for you i guess. i can't say for sure now | 00:17 |
| ivc_ | vikasc thats why i suggested that you do everything in Nested VIFDriver. you can call utility functions from os_vif_util | 00:19 |
| ivc_ | vikasc then when we have all the parts working we'll see if we can move some of them to vif_translator or maybe update vif_translator to do more than what it does now | 00:19 |
| vikasc | ivc_, ok..i am fine with this also | 00:20 |
| ivc_ | vikasc vif_translator was designed to work with assumption that all the necessary information is present in port dict. i thought it would work for sub-ports because i expected sub-port dicts to have sub-port related info (i.e. vlan_id and parent port id), i thought they've implemented it as an extension | 00:22 |
| vikasc | ivc_, nope, parent port id will come from configuration and vlan_id kuryr has to allocate | 00:23 |
| vikasc | ivc_, cant we pass these parameters as nullables to translators | 00:23 |
| vikasc | ivc_, the one want to use, will use them | 00:24 |
| ivc_ | vikasc https://github.com/openstack/neutron/blob/master/neutron/services/trunk/plugin.py#L43 | 00:28 |
| ivc_ | vikasc that's what i was referring to about 'port dict extension' | 00:28 |
| ivc_ | vikasc the 'trunk_details' part of port | 00:29 |
| ivc_ | vikasc i need to check how it actually works and what's there in neutron-client response | 00:32 |
| vikasc | ivc_, i will check and in case it supports then using nested translator make more sense. | 00:32 |
| vikasc | ivc_, i will complete it and lets take others opinion as well on using or skipping translator. | 00:33 |
| vikasc | ivc_, when i will test it, things will get more clear | 00:34 |
| ivc_ | vikasc aye, so if you can't get enough info to build the VIF object from neutron-client response, just don't use the translator plugin system for now. i'll see what can be done about it | 00:34 |
| ivc_ | vikasc the fact that it is merged does not mean we won't change it :) | 00:35 |
| vikasc | ivc_, sure | 00:35 |
| vikasc | ivc_, :) | 00:35 |
| vikasc | ivc_, what issue do you see in passing 'vlan_id' and 'parent_port_id' as nullables to translator api? | 00:38 |
| ivc_ | vikasc you mean adding those as arguments? | 00:39 |
| vikasc | ivc_, yes | 00:39 |
| ivc_ | vikasc it's special-casing | 00:39 |
| ivc_ | i'd prefer if we avoid that | 00:39 |
| vikasc | ivc_, its needed for all nested | 00:40 |
| ivc_ | why do you need 'parrent_port_id' in VIF? | 00:41 |
| vikasc | ivc_, yeah, sorry not that | 00:41 |
| vikasc | ivc_, just segmentation_id | 00:41 |
| ivc_ | vikasc yeah. well because it's not needed for non-nested :) | 00:43 |
| vikasc | ivc_, or may be i can update the vif with vlan_id in VIFdriver after getting vif | 00:43 |
| ivc_ | vikasc that's an option | 00:43 |
| vikasc | ivc_, then no need to change translator api | 00:43 |
| vikasc | ivc_, i like this option | 00:44 |
| ivc_ | vikasc yes, but it's a bit of a hack. we need to see how the 'whole picture' looks like and then we can decompose, thats why i suggested you skip the translator part for now | 00:45 |
| ivc_ | vikasc i mean just leave either of 'TODO/FIXME/REVISIT' that says something along 'consider using vif_translators' | 00:46 |
| vikasc | ivc_, :D | 00:46 |
| ivc_ | vikasc that should help us get something working without caring too much. think of it as PoC | 00:46 |
| ivc_ | vikasc i normally try to follow 'make it work -> make it clean -> make it perfect' approach :) | 00:47 |
| *** roeyc has quit IRC | 00:48 | |
| vikasc | ivc_, i would prefer to do it with translator and then a 'TODO/FIXME/REVISIT' saying this is kind of a hack, consider decomposing | 00:48 |
| vikasc | ivc_, it will 'work', it is 'cleaner' than not using vif and later we will make it perfect | 00:49 |
| vikasc | s/vif/vif-translator | 00:49 |
| ivc_ | vikasc yeah sure. either way would work i think | 00:49 |
| vikasc | ivc_, thanks a lot Ilya!! | 00:49 |
| vikasc | ivc_, really appreciate | 00:49 |
| ivc_ | vikasc np :) | 00:50 |
| vikasc | today also you have not slept it seems | 00:50 |
| vikasc | :D | 00:50 |
| vikasc | ivc_, ^ | 00:50 |
| ivc_ | nah i'm normally up until early morning :) | 00:51 |
| vikasc | yeah, i was not expecting you online when i pined | 00:52 |
| vikasc | :) | 00:52 |
| vikasc | s/pinged | 00:52 |
| ivc_ | its 6:30AM for you now, right? :) | 00:53 |
| vikasc | yes | 00:53 |
| vikasc | i wake up early | 00:53 |
| *** tonanhngo has joined #openstack-kuryr | 00:53 | |
| ivc_ | and its saturday | 00:53 |
| vikasc | yes :P | 00:53 |
| *** tonanhngo has quit IRC | 00:54 | |
| ivc_ | :P | 00:54 |
| ivc_ | hmm for some reason i don't see any 'trunk'-related commands in my devstack's neutron client CLI | 00:56 |
| ivc_ | vikasc do we need to enable something to get 'vlan-aware-vms' trunk/subport in client? | 00:57 |
| vikasc | ivc_, you mean in devstack? | 01:04 |
| vikasc | ivc_, trunk service plugin is to be enabled | 01:04 |
| ivc_ | vikasc i'm trying to figure out how to manage trunks from neutron cli. there are 'create_trunk'/etc. in python api | 01:05 |
| vikasc | ivc_, to enable cli commands, plugin is to be enabled | 01:05 |
| ivc_ | ah ok | 01:05 |
| ivc_ | vikasc you got a local.conf for that? | 01:06 |
| vikasc | ivc_, looking for it | 01:06 |
| vikasc | ivc_, 1 sec | 01:07 |
| vikasc | ivc_, in neutron.conf, service_plugins = neutron.services.l3_router.l3_router_plugin.L3RouterPlugin,neutron.services.trunk.plugin.TrunkPlugin | 01:07 |
| vikasc | ivc_, and restart q-svc | 01:08 |
| ivc_ | aye | 01:08 |
| vikasc | ivc_, going for a walk with my dog | 01:08 |
| ivc_ | vikasc ok, ttyl | 01:09 |
| *** yamamoto has joined #openstack-kuryr | 02:00 | |
| *** tonanhngo has joined #openstack-kuryr | 03:26 | |
| *** tonanhngo has quit IRC | 03:27 | |
| *** saneax is now known as saneax-_-|AFK | 03:32 | |
| *** reedip has quit IRC | 03:46 | |
| *** reedip has joined #openstack-kuryr | 04:00 | |
| *** yamamoto has quit IRC | 04:13 | |
| *** dougbtv_ has joined #openstack-kuryr | 04:19 | |
| *** yamamoto has joined #openstack-kuryr | 05:14 | |
| *** yamamoto has quit IRC | 05:18 | |
| *** tonanhngo has joined #openstack-kuryr | 06:53 | |
| *** tonanhngo has quit IRC | 06:54 | |
| *** tonanhngo has joined #openstack-kuryr | 06:54 | |
| *** tonanhngo has quit IRC | 06:55 | |
| *** tonanhngo has joined #openstack-kuryr | 06:55 | |
| *** tonanhngo has quit IRC | 07:00 | |
| *** dougbtv has quit IRC | 07:17 | |
| *** dougbtv_ has quit IRC | 07:18 | |
| *** saneax-_-|AFK is now known as saneax | 07:22 | |
| *** yamamoto has joined #openstack-kuryr | 09:08 | |
| *** yamamoto has quit IRC | 09:13 | |
| *** yamamoto has joined #openstack-kuryr | 09:15 | |
| *** dougbtv has joined #openstack-kuryr | 09:15 | |
| *** dougbtv_ has joined #openstack-kuryr | 09:17 | |
| *** apuimedo has quit IRC | 09:29 | |
| *** apuimedo has joined #openstack-kuryr | 09:30 | |
| *** yamamoto has quit IRC | 10:02 | |
| openstackgerrit | vikas choudhary proposed openstack/kuryr-kubernetes: [WIP] Add support for nested pods with Vlan trunk port https://review.openstack.org/410578 | 10:13 |
| *** yamamoto has joined #openstack-kuryr | 11:03 | |
| *** yamamoto has quit IRC | 11:11 | |
| *** yamamoto has joined #openstack-kuryr | 11:17 | |
| *** yamamoto has quit IRC | 12:50 | |
| *** yamamoto has joined #openstack-kuryr | 13:06 | |
| *** saneax is now known as saneax-_-|AFK | 13:06 | |
| *** yamamoto has quit IRC | 13:29 | |
| *** yamamoto has joined #openstack-kuryr | 13:59 | |
| *** dimak has joined #openstack-kuryr | 14:34 | |
| *** dimak has quit IRC | 14:39 | |
| *** dimak has joined #openstack-kuryr | 14:52 | |
| *** dimak has quit IRC | 15:01 | |
| *** saneax-_-|AFK is now known as saneax | 15:09 | |
| *** dimak has joined #openstack-kuryr | 15:15 | |
| *** saneax is now known as saneax-_-|AFK | 15:17 | |
| *** dimak has quit IRC | 15:21 | |
| *** saneax-_-|AFK is now known as saneax | 15:29 | |
| *** dimak has joined #openstack-kuryr | 15:41 | |
| *** yamamoto has quit IRC | 16:11 | |
| *** yamamoto has joined #openstack-kuryr | 17:12 | |
| *** dimak has quit IRC | 17:16 | |
| *** yamamoto has quit IRC | 17:20 | |
| *** saneax is now known as saneax-_-|AFK | 17:44 | |
| *** david-lyle_ has joined #openstack-kuryr | 17:50 | |
| *** david-lyle has quit IRC | 17:51 | |
| *** dimak has joined #openstack-kuryr | 19:41 | |
| *** dimak has quit IRC | 21:00 | |
| *** hongbin has joined #openstack-kuryr | 21:34 | |
| openstackgerrit | Hongbin Lu proposed openstack/fuxi: Separate unit tests from fullstack tests https://review.openstack.org/403931 | 23:11 |
| openstackgerrit | Hongbin Lu proposed openstack/fuxi: Add basic fullstack volume tests https://review.openstack.org/403941 | 23:11 |
| *** saneax-_-|AFK is now known as saneax | 23:16 | |
| *** hongbin has quit IRC | 23:34 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!