| *** kiennt has joined #openstack-kuryr | 00:31 | |
| *** limao has joined #openstack-kuryr | 00:34 | |
| *** hongbin has joined #openstack-kuryr | 00:43 | |
| *** yedongcan has joined #openstack-kuryr | 01:07 | |
| *** deep-book-gk has joined #openstack-kuryr | 01:45 | |
| *** deep-book-gk has left #openstack-kuryr | 01:48 | |
| *** irenab has quit IRC | 01:56 | |
| *** irenab has joined #openstack-kuryr | 01:57 | |
| *** hyunsun has joined #openstack-kuryr | 02:01 | |
| *** gouthamr has quit IRC | 02:07 | |
| *** caowei has joined #openstack-kuryr | 02:08 | |
| *** hyunsun has quit IRC | 02:14 | |
| *** vikasc has joined #openstack-kuryr | 03:11 | |
| *** aojea has joined #openstack-kuryr | 03:34 | |
| *** aojea has quit IRC | 03:38 | |
| *** kiennt_ has joined #openstack-kuryr | 04:14 | |
| *** yboaron has joined #openstack-kuryr | 04:31 | |
| openstackgerrit | Hyunsun Moon proposed openstack/kuryr-libnetwork master: Allow new network ID pattern https://review.openstack.org/487802 | 04:46 |
|---|---|---|
| *** yboaron has quit IRC | 04:47 | |
| *** hongbin has quit IRC | 05:01 | |
| *** aojea has joined #openstack-kuryr | 05:08 | |
| *** janki has joined #openstack-kuryr | 05:10 | |
| *** caowei has quit IRC | 05:23 | |
| *** aojea has quit IRC | 05:47 | |
| *** aojea has joined #openstack-kuryr | 05:48 | |
| *** kiennt_ has quit IRC | 05:49 | |
| *** kiennt has quit IRC | 05:49 | |
| *** aojea has quit IRC | 05:52 | |
| *** aojea has joined #openstack-kuryr | 05:52 | |
| openstackgerrit | Berezovsky Irena proposed openstack/kuryr-kubernetes master: Add devstack support for multi-node deployment https://review.openstack.org/489098 | 06:00 |
| *** kiennt has joined #openstack-kuryr | 06:02 | |
| *** kiennt_ has joined #openstack-kuryr | 06:02 | |
| *** kiennt__ has joined #openstack-kuryr | 06:05 | |
| *** kiennt__ has quit IRC | 06:05 | |
| *** kiennt_ has quit IRC | 06:06 | |
| *** kiennt has quit IRC | 06:06 | |
| openstackgerrit | Berezovsky Irena proposed openstack/kuryr-kubernetes master: [WIP] Add devstack support for multi-node deployment https://review.openstack.org/489098 | 06:06 |
| *** kiennt has joined #openstack-kuryr | 06:06 | |
| *** caowei has joined #openstack-kuryr | 06:15 | |
| *** pcaruana has joined #openstack-kuryr | 06:19 | |
| *** kzaitsev_ws has joined #openstack-kuryr | 06:25 | |
| ltomasbo | Good morning irenab! Yes! It is useful for others! :D | 06:40 |
| irenab | ltomasbo, still need to be properly verified. I didn't try to run it for master yet | 06:40 |
| irenab | hence WIP | 06:41 |
| ltomasbo | sure, there may be some small problems, but in general it should work! I'll try to test it too | 06:42 |
| *** yboaron has joined #openstack-kuryr | 07:06 | |
| irenab | ltomasbo, thanks! | 07:10 |
| *** yboaron_ has joined #openstack-kuryr | 07:21 | |
| *** yboaron has quit IRC | 07:24 | |
| apuimedo | irenab: what does your comment to https://review.openstack.org/#/c/487935/ mean? | 07:25 |
| irenab | apuimedo, if you click the bug link in the commit messgae, it takes you to the launchpad, but the link to the patch is not there | 07:26 |
| apuimedo | oh | 07:27 |
| apuimedo | that's odd | 07:27 |
| irenab | apuimedo, maybe Fixes-Bug does not work, need Closes-Bug: ####### | 07:27 |
| irenab | Partial-Bug: ####### | 07:27 |
| irenab | Related-Bug: ####### | 07:27 |
| irenab | maybe not, just checked here https://review.openstack.org/#/c/489098/, it is also not on the launchpad .. | 07:29 |
| apuimedo | ok | 07:29 |
| janonymous | irenab: https://review.openstack.org/#/c/489098/ --> you need to assign bug in launchpad to update | 07:31 |
| irenab | janonymous, it is assigned | 07:32 |
| janonymous | irenab: ah no , to you | 07:32 |
| janonymous | irenab: it is assigned to daniel | 07:32 |
| irenab | janonymous, interesting | 07:33 |
| irenab | dmellado, ping | 07:35 |
| apuimedo | irenab: dmellado is on pto | 07:43 |
| apuimedo | feel free to assign it to yourself | 07:43 |
| *** egonzalez has joined #openstack-kuryr | 07:44 | |
| *** yboaron_ has quit IRC | 07:44 | |
| *** yboaron_ has joined #openstack-kuryr | 07:45 | |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: devstack: create host interface to the pod subnet https://review.openstack.org/487935 | 07:46 |
| openstackgerrit | Berezovsky Irena proposed openstack/kuryr-kubernetes master: [WIP] Add devstack support for multi-node deployment https://review.openstack.org/489098 | 07:47 |
| irenab | apuimedo, great, thanks. Will appreciate your review on this patch | 07:47 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: devstack: create host interface to the pod subnet https://review.openstack.org/487935 | 07:47 |
| apuimedo | irenab: very well | 07:48 |
| apuimedo | irenab: I have octavia working but due to the conversation with longfei_zhang here on IRC last week I want to make it configurable whether to use L2 connectivity between the LB and the members or L# | 07:49 |
| apuimedo | s/L#/L3/ | 07:49 |
| apuimedo | currently, as we do it, it's L2 (which is what he wanted) | 07:49 |
| apuimedo | but that means that for every load balancer, we take an extra pod subnet IP | 07:49 |
| irenab | apuimedo, want to split to iterations? | 07:49 |
| apuimedo | irenab: maybe I should | 07:49 |
| irenab | consuming pods' IPs are bad | 07:50 |
| irenab | but can be first step | 07:51 |
| apuimedo | irenab: it's a tradeoff | 07:51 |
| openstackgerrit | Berezovsky Irena proposed openstack/kuryr-kubernetes master: [WIP] Add devstack support for multi-node deployment https://review.openstack.org/489098 | 07:51 |
| apuimedo | for the ref impl, it gets you services -> pods without going through the router | 07:51 |
| apuimedo | which is a big win | 07:51 |
| apuimedo | you just resize the pod subnet accordingly | 07:52 |
| apuimedo | for df and ovn you don't gain anything though | 07:52 |
| irenab | apuimedo, yea. Agree that is should be configurable, but actualy I would go for the L3 as default | 07:52 |
| apuimedo | irenab: vikasc: can we get https://review.openstack.org/#/c/487935/ in? | 07:52 |
| apuimedo | irenab: that is my intention | 07:53 |
| apuimedo | to add an lbaas handler config option | 07:53 |
| apuimedo | [lbaas] | 07:53 |
| irenab | apuimedo, will check asap | 07:53 |
| apuimedo | member_connectivity = "L3" | 07:53 |
| apuimedo | that you can change to L2 | 07:53 |
| irenab | apuimedo, I rebased your overcloud heat patch | 07:53 |
| irenab | it worth to get it in, since it gets confliced every few days | 07:54 |
| kzaitsev_ws | janonymous: apuimedo: I thought we've fixed the infra bot not updating the bugs on the lp thing | 07:59 |
| apuimedo | irenab: :-) | 07:59 |
| apuimedo | kzaitsev_ws: apparently not | 07:59 |
| apuimedo | :/ | 07:59 |
| kzaitsev_ws | apuimedo: aha, you need to add openstack-infra to this group https://launchpad.net/~kuryr-bugs/+members (this group is the bug supervisor) | 08:00 |
| janonymous | kzaitsev_ws: ohh | 08:00 |
| kzaitsev_ws | let me fetch the bot's mail | 08:00 |
| janonymous | irenab: it worked right? | 08:01 |
| irenab | janonymous, didn't change it yet | 08:01 |
| kzaitsev_ws | apuimedo: https://launchpad.net/~hudson-openstack this one I think. | 08:02 |
| irenab | but really surprised that is is strict | 08:02 |
| janonymous | irenab: ah ohky | 08:02 |
| *** yamamoto has quit IRC | 08:02 | |
| irenab | apuimedo, done. Now your turn to review my patch | 08:03 |
| apuimedo | cool | 08:03 |
| openstackgerrit | Hyunsun Moon proposed openstack/kuryr-libnetwork master: Allow new network ID pattern https://review.openstack.org/487802 | 08:04 |
| kzaitsev_ws | btw. sorry for not tending to my patches. I'm suddenly made responsible for a bit of integration between the teems around here, so I'm fighting fires. Would get back to them when the fires are gone. | 08:09 |
| irenab | kzaitsev_ws, do not get burned :-) | 08:10 |
| kzaitsev_ws | irenab: I'll try my best ) | 08:11 |
| kzaitsev_ws | irenab: apuimedo: https://bugs.launchpad.net/kuryr-kubernetes/+bug/1706576 looks like the bot is working now ;) | 08:11 |
| openstack | Launchpad bug 1706576 in kuryr-kubernetes "devstack does not support multi node scenario" [High,In progress] - Assigned to Irena Berezovsky (irenab) | 08:11 |
| irenab | kzaitsev_ws, do not see link to the patch there | 08:13 |
| kzaitsev_ws | irenab: the bot has moved the bug to in-progress though | 08:14 |
| irenab | kzaitsev_ws, bot developed its own judgment? | 08:14 |
| kzaitsev_ws | I believe it triggers on a new-patchset. maybe only a brand new patch triggers a message with "fix proposed + link" and a new changeset to existing patch only triggers state transition? | 08:16 |
| irenab | kzaitsev_ws, sounds reasonable | 08:17 |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: Add ODL devstack example https://review.openstack.org/449309 | 08:20 |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: devstack: create host interface to the pod subnet https://review.openstack.org/487935 | 08:40 |
| *** garyloug has joined #openstack-kuryr | 08:43 | |
| apuimedo | irenab: the multi node overcloud one looks good to me :-) | 08:57 |
| irenab | apuimedo, great. Hope you will have time to verify (just in case) and we can merge it | 09:01 |
| apuimedo | irenab: I'll try to get to that today | 09:01 |
| irenab | thanks! | 09:01 |
| *** yamamoto has joined #openstack-kuryr | 09:02 | |
| *** yamamoto has quit IRC | 09:08 | |
| *** yamamoto has joined #openstack-kuryr | 09:08 | |
| *** limao has quit IRC | 09:23 | |
| *** limao has joined #openstack-kuryr | 09:23 | |
| *** yamamoto has quit IRC | 09:25 | |
| *** limao has quit IRC | 09:28 | |
| *** yamamoto has joined #openstack-kuryr | 09:34 | |
| *** janki is now known as janki|meeting | 09:37 | |
| *** janki|meeting is now known as janki | 09:39 | |
| *** yamamoto has quit IRC | 09:40 | |
| *** yamamoto has joined #openstack-kuryr | 09:53 | |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: octavia: Make Octavia ready devstack https://review.openstack.org/489157 | 09:54 |
| apuimedo | irenab: ltomasbo: please, check ^^ | 09:54 |
| apuimedo | vikasc: ^^ | 09:54 |
| apuimedo | I'll make a follow-up patch that changes the default to use L3 | 09:56 |
| vikasc | apuimedo, will try out, thanks! | 09:56 |
| apuimedo | :-) | 09:56 |
| apuimedo | I'll make a demo soon out of this | 09:56 |
| apuimedo | but with openshift | 09:56 |
| ltomasbo | apuimedo, great! I'll take a look asap | 09:58 |
| apuimedo | ltomasbo: vikasc: irenab: note that this local.conf will only work in master and pike+ | 09:59 |
| apuimedo | since I dropped neutron-lbaas | 09:59 |
| apuimedo | that is not necessary anymore | 09:59 |
| vikasc | noted. | 09:59 |
| apuimedo | if using Ocata + Octavia, enable the neutron-lbaas repo plugin as well | 09:59 |
| *** kiennt has quit IRC | 10:01 | |
| openstackgerrit | Danil Golov proposed openstack/kuryr-kubernetes master: [WIP] Allow setting specific ports for SRIOV handler https://review.openstack.org/478494 | 10:15 |
| irenab | apuimedo, worth to document the required compotents with regard to Releases | 10:36 |
| kzaitsev_ws | apuimedo: do you plan to abandon neutron lbaas alltogether or keep it as a configurable option? | 10:46 |
| kzaitsev_ws | also really cool that it doesn't require any actual code change... | 10:49 |
| irenab | ivc, did really good job | 10:49 |
| apuimedo | irenab: can you ellaborate? | 10:49 |
| apuimedo | kzaitsev_ws: it needs one code change for working in L3 mode | 10:50 |
| apuimedo | I'm working on it now | 10:50 |
| irenab | apuimedo, about the required services for different releases | 10:50 |
| irenab | maybe in local.conf or some installation documentation | 10:50 |
| irenab | I wonder if we need installation doc revision per release | 10:50 |
| irenab | so for Ocata, there will be neutron-lbaas, but for pike only Octavia | 10:51 |
| kzaitsev_ws | we can have a single "USE_OCTAVIA" that defaults to true and a config function that set's up all the things that are in local.conf | 10:51 |
| apuimedo | irenab: I was thinking to make a developing section in the docs | 10:51 |
| kzaitsev_ws | that would keep local.conf simple enough | 10:51 |
| irenab | this is a bit more of installation than developing | 10:52 |
| irenab | kzaitsev_ws, use_octavia will end up with different services based on the openstack version | 10:52 |
| kzaitsev_ws | irenab: it will, but I don't see a problem there if you want octavia you need them anyway | 10:53 |
| irenab | neutron_lbass is not required in Pike, since Octavia added API server | 10:53 |
| irenab | apuimedo, correct? | 10:54 |
| apuimedo | correct | 10:54 |
| irenab | this maybe just another config in local.conf | 10:54 |
| kzaitsev_ws | we're making developer think/know about what services are required for octavia, which feels to me like leaking abstraction. | 10:55 |
| irenab | it is more Octavia thing than kuryr's | 10:55 |
| kzaitsev_ws | aha, wouldn't it configure all the services it needs btw? | 10:55 |
| apuimedo | kzaitsev_ws: I still want deployers and developers to think whether they prefer l2 or l3 | 10:56 |
| apuimedo | for lb -> member | 10:56 |
| kzaitsev_ws | apuimedo: yeah, that sounds fair. | 10:56 |
| apuimedo | kzaitsev_ws: will you spin anew your multi vif cni support? | 10:59 |
| apuimedo | I'd like to start getting these things in | 11:00 |
| *** aojea has quit IRC | 11:05 | |
| kzaitsev_ws | apuimedo: yeah. I need to deal with a couple of things locally and would get to them. | 11:06 |
| apuimedo | ok | 11:08 |
| kzaitsev_ws | apuimedo: while we're at it. irenab and I thought of a way to reduce amount of changes in my patches that go to generic vif handler. | 11:13 |
| irenab | kzaitsev_ws, seems we should move to modular_vif_driver | 11:13 |
| kzaitsev_ws | we can add a config variable, smth like 'enabled_drivers=pool,sriov,additional_subnets' | 11:14 |
| apuimedo | irenab: kzaitsev_ws: modular vif driver? | 11:14 |
| kzaitsev_ws | and vif_handler would pass the pod object to each one of them and would collect the vifs they request from neutron. | 11:15 |
| irenab | otherwise the generic vif driver gets 'polluted' with many specific if/else cases | 11:15 |
| apuimedo | can we work around that with multi net? | 11:15 |
| irenab | similar with what modular l2 plugin does in neutron | 11:15 |
| irenab | apuimedo, I wouldn't say its related to multiple networks | 11:16 |
| kzaitsev_ws | irenab: +1 it's more about different ways to request ports | 11:16 |
| apuimedo | irenab: kzaitsev_ws: my idea was that when you define a network object you can sepcify the type | 11:18 |
| apuimedo | this way we don't need configuration for more than the dfeault network | 11:18 |
| kzaitsev_ws | like in danil's patch he's adding a way to request a Port by id. that's another set of if/else's. sriov and pool drivers are also smth vif-handler shouldn't necessarily know about | 11:18 |
| kzaitsev_ws | apuimedo: that way (at least for sriov) we would still have to put some VIF-specific logic to generic vif | 11:19 |
| * apuimedo going for lunch. We'll read again when I'm back | 11:19 | |
| apuimedo | see if I catch it | 11:20 |
| apuimedo | :P | 11:20 |
| irenab | apuimedo, its just more for refactoring to provide extendability at vif plugging layer | 11:20 |
| kzaitsev_ws | (= | 11:20 |
| irenab | apuimedo, bon appetite | 11:20 |
| *** dougbtv__ is now known as dougbtv | 11:26 | |
| kzaitsev_ws | apuimedo: I think irenab's idea/concern can be summed up as "do not alter generic vif handler if you need to parse some specific annotation" | 11:29 |
| irenab | kzaitsev_ws, +1 | 11:30 |
| irenab | even out of tree options can be supported without modification of the kuryr code | 11:31 |
| *** yamamoto has quit IRC | 11:31 | |
| irenab | apuimedo, I vefiried BM deployment of the kuryr all in one with devstack patch, it works properly | 11:31 |
| kzaitsev_ws | irenab: it almost feels to me like we can later just switch the code from reading the config to reading the network objects to determine which plugins have to be enabled. but I'm not 100% sure on that. | 11:32 |
| kzaitsev_ws | mayb it's the other way and we can implement network-obj-reading driver, that would parse an annotation ) again withour altering main vif_handler's code =) | 11:33 |
| irenab | kzaitsev_ws, I think you are right. About the other way around, I just not sure that we always have exact mapping to network-obj driver | 11:34 |
| irenab | as with the port_id annotation | 11:34 |
| irenab | its just specific port to request | 11:34 |
| *** caowei has quit IRC | 11:38 | |
| apuimedo | irenab: which patch? | 11:41 |
| irenab | the one for multi node | 11:41 |
| apuimedo | so you used the local.conf in one | 11:42 |
| irenab | I just checked that the most common case is not broken :-) | 11:42 |
| apuimedo | and another VM with yours? | 11:42 |
| apuimedo | ah | 11:42 |
| irenab | I do not have the setup to verify multi node BM | 11:42 |
| irenab | but master+worker BM works properly | 11:42 |
| apuimedo | oh, you meant real baremetal? | 11:42 |
| irenab | and nested master with another nested worker too | 11:43 |
| irenab | yes | 11:43 |
| apuimedo | cool | 11:44 |
| *** yamamoto has joined #openstack-kuryr | 12:01 | |
| *** yamamoto has quit IRC | 12:04 | |
| *** yamamoto has joined #openstack-kuryr | 12:20 | |
| *** yamamoto has quit IRC | 12:32 | |
| *** yamamoto has joined #openstack-kuryr | 12:33 | |
| *** yamamoto has quit IRC | 12:33 | |
| *** yamamoto has joined #openstack-kuryr | 12:33 | |
| apuimedo | kzaitsev_ws: I can't find the "setup" phase | 12:33 |
| kzaitsev_ws | apuimedo: there should be some =) for example we could use override_defaults | 12:37 |
| *** zengchen has quit IRC | 12:37 | |
| apuimedo | kzaitsev_ws: I was trying to pick between 'source' and 'override_defaults' | 12:37 |
| *** zengchen has joined #openstack-kuryr | 12:37 | |
| apuimedo | also considering just enabling them in our devstack/settings | 12:37 |
| kzaitsev_ws | that's an option too | 12:38 |
| apuimedo | kzaitsev_ws: irenab: which do you like more? | 12:38 |
| kzaitsev_ws | i don't think just enabling in settings is a good idea though. feels confusing. I would expect settings to be settings without much logic | 12:39 |
| apuimedo | okey dokey | 12:39 |
| apuimedo | so I'll try with 'source' phase | 12:39 |
| apuimedo | no, that's too late, I think | 12:40 |
| kzaitsev_ws | just checked murano's plugin. murano requires heat and it's not enabled anywhere at the moment. =/ yet I was sure it was somewhere. | 12:41 |
| kzaitsev_ws | apuimedo: maybe even that would not work | 12:41 |
| apuimedo | kzaitsev_ws: because the official position is that you are supposed to put it in local.conf | 12:42 |
| kzaitsev_ws | https://github.com/openstack-dev/devstack/blob/master/stack.sh#L584-L588 | 12:42 |
| apuimedo | just like I did in my patch and like we do with etcd and so on | 12:42 |
| kzaitsev_ws | fetch_plugins before source probably means that plugins can't implicitly download other plugins =( | 12:43 |
| apuimedo | right | 12:43 |
| kzaitsev_ws | so my nice single variable wouldn't work | 12:43 |
| apuimedo | I think we have to leave it in local.conf or devstackgaterc | 12:43 |
| apuimedo | s/or/and/ | 12:43 |
| apuimedo | kzaitsev_ws: tbh, there should be a way for plugins to have /devstack/deps | 12:44 |
| apuimedo | and put there the plugins they depend on | 12:44 |
| kzaitsev_ws | apuimedo: sounds scary (= | 12:47 |
| *** rwallner_ has quit IRC | 12:47 | |
| *** rwallner has joined #openstack-kuryr | 12:47 | |
| kzaitsev_ws | ok other idea. if you have octavia enabled — run octavia preps. elif you have lpaasv2 run lbaasv2 prep. else complain that services would not work? | 12:48 |
| apuimedo | kzaitsev_ws: what's prep? | 12:49 |
| kzaitsev_ws | we do that for horizon (configure dashboard only if you have enabled horizon) | 12:49 |
| apuimedo | oh, right | 12:50 |
| apuimedo | of course | 12:50 |
| kzaitsev_ws | apuimedo: enable additional services and do that thing security-groups | 12:50 |
| kzaitsev_ws | idk. maybe you're right and I'm wrong about enabled_services | 12:53 |
| kzaitsev_ws | and they *should* just go to local.conf | 12:53 |
| kzaitsev_ws | the security-groups code can be safely wrapped under 'if service_enabled octavia' though =) | 12:55 |
| *** rwallner has quit IRC | 12:55 | |
| *** rwallner has joined #openstack-kuryr | 12:56 | |
| apuimedo | yeah | 12:59 |
| apuimedo | at least that | 12:59 |
| apuimedo | I hate devstack | 13:09 |
| apuimedo | necessary evil | 13:09 |
| irenab | apuimedo, +1000 | 13:09 |
| *** yedongcan1 has joined #openstack-kuryr | 13:12 | |
| *** yedongcan has quit IRC | 13:14 | |
| apuimedo | irenab: amphorae is the plural of amphora ;-) | 13:15 |
| *** atoth has joined #openstack-kuryr | 13:27 | |
| *** gouthamr has joined #openstack-kuryr | 13:31 | |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: octavia: Make Octavia ready devstack https://review.openstack.org/489157 | 13:39 |
| apuimedo | kzaitsev_ws: irenab: I think I may have addressed all your comments ^^ | 13:39 |
| *** yedongcan1 has left #openstack-kuryr | 13:45 | |
| kzaitsev_ws | ohmygosh =) does octavia require barbican? | 13:57 |
| apuimedo | kzaitsev_ws: I think so | 14:00 |
| apuimedo | they enable it in their local.conf | 14:01 |
| *** zengchen1 has joined #openstack-kuryr | 14:01 | |
| *** limao has joined #openstack-kuryr | 14:03 | |
| kzaitsev_ws | I'm always afraid of new cross-project 'big-tent'ish' dependencies (I know that big tent is gone, still there's no better work I guess) | 14:03 |
| kzaitsev_ws | yet | 14:03 |
| *** kural has joined #openstack-kuryr | 14:03 | |
| apuimedo | kzaitsev_ws: it is scary indeed | 14:08 |
| apuimedo | kzaitsev_ws: although in this case I'd expect barbican not to be necessary for L4 most usage | 14:08 |
| irenab | apuimedo: jenkins is not happy on Octavia patch | 14:09 |
| apuimedo | irenab: yes, yes. I made a mistake | 14:10 |
| apuimedo | I'm pushing a new version | 14:10 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: octavia: Make Octavia ready devstack https://review.openstack.org/489157 | 14:11 |
| apuimedo | irenab: here you go | 14:11 |
| *** janki has quit IRC | 14:17 | |
| openstackgerrit | Merged openstack/kuryr-libnetwork master: Allow new network ID pattern https://review.openstack.org/487802 | 14:26 |
| *** openstackgerrit has quit IRC | 14:33 | |
| *** openstackgerrit has joined #openstack-kuryr | 14:34 | |
| openstackgerrit | Merged openstack/kuryr-libnetwork master: Return network options when AllocateNetwork called https://review.openstack.org/487258 | 14:34 |
| *** aojea has joined #openstack-kuryr | 14:44 | |
| *** yamamoto has quit IRC | 14:45 | |
| *** limao has quit IRC | 14:52 | |
| *** hongbin has joined #openstack-kuryr | 14:52 | |
| *** limao has joined #openstack-kuryr | 14:54 | |
| janonymous | apuimedo: zengchen1 : so i wanted to know on drawbacks/limitations of python one client model | 14:59 |
| apuimedo | I didn't look at it in some time, but apparently duplicated events, reconnection and things like that | 15:00 |
| *** alraddarla has joined #openstack-kuryr | 15:02 | |
| janonymous | agree that go-client is good and robust, but was curious on things which need improvement in python- client | 15:02 |
| janonymous | this week i will test client patch e2e | 15:03 |
| apuimedo | good | 15:03 |
| apuimedo | try to take notes of the findings, to report | 15:03 |
| apuimedo | kzaitsev_ws: irenab: could you try https://review.openstack.org/#/c/489157/ ? | 15:04 |
| janonymous | apuimedo: right! i will note down | 15:04 |
| *** pcaruana has quit IRC | 15:04 | |
| zengchen1 | janonymous:do you see my comments on your patch | 15:04 |
| irenab | apuimedo: asap | 15:04 |
| *** alraddarla has left #openstack-kuryr | 15:05 | |
| janonymous | zengchen1: yes, i had a reason for that but while loop still dont make sense | 15:05 |
| janonymous | zengchen1: i wanted to avoid iterating 2 times so passed object directly.. | 15:05 |
| zengchen1 | janonymous:no, it is not right. | 15:06 |
| zengchen1 | janonymous:if you return a generator directly, the code of 'while True' will be useless. | 15:06 |
| janonymous | zengchen1: right! i wrote that above.. | 15:07 |
| irenab | apuimedo: posted some question | 15:08 |
| janonymous | zengchen1: i will test it out this week and correct errors | 15:08 |
| zengchen1 | janonymous:when you finish iterating the generator, how to re-watch the resource if break out the 'while True'. | 15:08 |
| zengchen1 | janonymous:ok. | 15:09 |
| zengchen1 | janonymous:i say the library of https://github.com/kubernetes/client-go/blob/master/tools/cache may be more correctly. because it use the two interface of list and watch to generate the real events | 15:10 |
| zengchen1 | janonymous: you can see this function to get the details. https://github.com/kubernetes/client-go/blob/master/tools/cache/controller.go#L279 | 15:12 |
| janonymous | zengchen1: more efficiently written i guess | 15:15 |
| *** yamamoto has joined #openstack-kuryr | 15:16 | |
| *** limao has quit IRC | 15:17 | |
| kzaitsev_ws | apuimedo: left devstack deploying with your commit. would probably check on it tomorrow, though (= | 15:19 |
| zengchen1 | janonymous:i encountered a scenario that i got no result when i invoke the watch interface directly to get an event of 'Add' for PVC which i created after i invoked the watch. | 15:19 |
| *** yamamoto has quit IRC | 15:21 | |
| janonymous | zengchen1: so watch was looking for which resource type? and was watch working correctly? | 15:21 |
| zengchen1 | janonymous: so, i am worry that it may be not right to get all events of resource only invoking the watch interface. | 15:21 |
| janonymous | zengchen1: i see channels usage in go for that.. but have to recheck what is used in python | 15:22 |
| janonymous | zengchen1: i guess there is some issue, watch should work ideally... | 15:22 |
| zengchen1 | janonymous: watch for PVC. I hope the mechanism of watch for python is same good as go. | 15:23 |
| zengchen1 | janonymous: i will try to find the reason. and will review your patch. | 15:24 |
| janonymous | zengchen1: great! thanks | 15:24 |
| janonymous | zengchen1: ping me if i can be of any assistance with client-working | 15:25 |
| zengchen1 | janonymous:sure. i will tell you what i have learned on that library. | 15:27 |
| janonymous | :) | 15:27 |
| *** yboaron_ has quit IRC | 15:32 | |
| *** yamamoto has joined #openstack-kuryr | 15:48 | |
| *** yamamoto has quit IRC | 15:58 | |
| *** zengchen1 has quit IRC | 16:00 | |
| *** egonzalez has quit IRC | 16:03 | |
| *** dougbtv_ has joined #openstack-kuryr | 16:32 | |
| *** garyloug has quit IRC | 16:58 | |
| *** kural has quit IRC | 17:27 | |
| *** kural has joined #openstack-kuryr | 18:20 | |
| *** tonanhngo has joined #openstack-kuryr | 18:26 | |
| *** aojea has quit IRC | 18:42 | |
| *** aojea has joined #openstack-kuryr | 18:42 | |
| *** aojea has quit IRC | 18:47 | |
| *** rwallner has quit IRC | 18:48 | |
| *** atoth has quit IRC | 19:41 | |
| *** aojea has joined #openstack-kuryr | 20:07 | |
| *** kural_ has joined #openstack-kuryr | 20:24 | |
| *** kural has quit IRC | 20:24 | |
| *** yamamoto_ has joined #openstack-kuryr | 21:06 | |
| *** yamamoto_ has quit IRC | 21:15 | |
| *** kural_ has quit IRC | 22:05 | |
| *** yamamoto has joined #openstack-kuryr | 22:16 | |
| *** aojea has quit IRC | 22:18 | |
| *** aojea has joined #openstack-kuryr | 22:19 | |
| *** lihi has quit IRC | 22:20 | |
| *** lihi has joined #openstack-kuryr | 22:20 | |
| *** yamamoto has quit IRC | 22:23 | |
| *** aojea has quit IRC | 22:24 | |
| *** yamamoto has joined #openstack-kuryr | 22:28 | |
| *** gouthamr has quit IRC | 22:46 | |
| *** tonanhngo has quit IRC | 22:53 | |
| *** premsankar has joined #openstack-kuryr | 22:54 | |
| *** tonanhngo has joined #openstack-kuryr | 23:05 | |
| *** tonanhngo has quit IRC | 23:08 | |
| *** tonanhngo has joined #openstack-kuryr | 23:09 | |
| *** tonanhngo has quit IRC | 23:09 | |
| *** hongbin has quit IRC | 23:11 | |
| *** tonanhngo has joined #openstack-kuryr | 23:13 | |
| *** gouthamr has joined #openstack-kuryr | 23:16 | |
| *** tonanhngo has quit IRC | 23:18 | |
| *** aojea has joined #openstack-kuryr | 23:20 | |
| *** aojea has quit IRC | 23:27 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!