| *** gouthamr has joined #openstack-kuryr | 00:10 | |
| *** salv-orlando has joined #openstack-kuryr | 00:15 | |
| *** gouthamr has quit IRC | 00:16 | |
| *** salv-orlando has quit IRC | 00:19 | |
| *** neiljerram has quit IRC | 00:27 | |
| *** feisky has joined #openstack-kuryr | 00:31 | |
| *** c00281451_ has quit IRC | 00:45 | |
| *** c00281451_ has joined #openstack-kuryr | 00:45 | |
| *** gouthamr has joined #openstack-kuryr | 01:12 | |
| *** yedongcan has joined #openstack-kuryr | 01:34 | |
| *** hongbin has quit IRC | 01:51 | |
| *** salv-orlando has joined #openstack-kuryr | 02:15 | |
| *** salv-orlando has quit IRC | 02:20 | |
| *** janki has joined #openstack-kuryr | 03:37 | |
| *** salv-orlando has joined #openstack-kuryr | 03:50 | |
| *** salv-orlando has quit IRC | 03:54 | |
| *** yamamoto has joined #openstack-kuryr | 04:13 | |
| *** aojea has joined #openstack-kuryr | 04:18 | |
| *** hongbin has joined #openstack-kuryr | 04:21 | |
| *** aojea has quit IRC | 04:22 | |
| *** salv-orlando has joined #openstack-kuryr | 04:39 | |
| *** gouthamr has quit IRC | 04:51 | |
| *** hongbin has quit IRC | 04:57 | |
| openstackgerrit | Frederick F. Kautz IV proposed openstack/kuryr-kubernetes master: Add support for USE_SCREEN and default to screen if USE_SYSTEMD is not set https://review.openstack.org/469280 | 04:58 |
|---|---|---|
| *** xiangxinyong is now known as edison | 06:08 | |
| *** edison is now known as edisonxiang | 06:08 | |
| *** yamamoto has quit IRC | 06:30 | |
| *** ltomasbo|away is now known as ltomasbo | 06:41 | |
| *** atoth has quit IRC | 06:41 | |
| *** atoth has joined #openstack-kuryr | 06:42 | |
| *** yedongcan has quit IRC | 07:00 | |
| *** yedongcan has joined #openstack-kuryr | 07:08 | |
| *** dimak_ has joined #openstack-kuryr | 07:09 | |
| *** yamamoto has joined #openstack-kuryr | 07:13 | |
| *** neiljerram has joined #openstack-kuryr | 07:22 | |
| *** pcaruana has joined #openstack-kuryr | 07:26 | |
| *** aojea has joined #openstack-kuryr | 07:26 | |
| *** salv-orl_ has joined #openstack-kuryr | 07:27 | |
| *** salv-orlando has quit IRC | 07:30 | |
| *** dimak_ has quit IRC | 07:40 | |
| *** egonzalez has joined #openstack-kuryr | 07:44 | |
| *** kzaitsev_ws has joined #openstack-kuryr | 08:02 | |
| *** garyloug_ has joined #openstack-kuryr | 08:49 | |
| *** yedongcan has left #openstack-kuryr | 08:50 | |
| *** dmellado_ has joined #openstack-kuryr | 09:00 | |
| *** dmellado has quit IRC | 09:01 | |
| *** limao has joined #openstack-kuryr | 09:03 | |
| apuimedo | kzaitsev_ws: A day late, sorry. But the "Unexpected command output Device "eth0" does not exist" is usually just a race between kubelet and docker for checking devices. It is not really an error and can happen without any adverse effects | 09:05 |
| dmellado_ | apuimedo: o/ | 09:08 |
| dmellado_ | do you folks know where does our devstack plugin puts the kube/config file? | 09:11 |
| kzaitsev_ws | apuimedo: yeah. I had to recheck every step of the flow. it seems the problem was in ovs-agent<->neutron communication in the end. | 09:14 |
| kzaitsev_ws | at least the problem went away after I restarted those & the rabbit | 09:14 |
| apuimedo | dmellado_: it doesn't put it | 09:17 |
| apuimedo | it's hardcoded from the kubectl client we use | 09:17 |
| apuimedo | (extracted from hyperkube) | 09:18 |
| dmellado_ | hmmm I see | 09:18 |
| apuimedo | dmellado_: but you can pass kubectl --config | 09:18 |
| apuimedo | iirc | 09:18 |
| dmellado_ | apuimedo: I was trying to have the python kubernetes client load them | 09:18 |
| apuimedo | kzaitsev_ws: so long story short. You can pretty much always ignore the stupid eth0 can't be found | 09:18 |
| dmellado_ | in order to setup the client base class | 09:18 |
| apuimedo | I think in 1.7 it's just a warning | 09:18 |
| apuimedo | dmellado_: nah, nah | 09:19 |
| apuimedo | we can just configure the client with the snippet I passed you | 09:19 |
| apuimedo | do you still have it? | 09:19 |
| dmellado_ | apuimedo: it could've been lost in time | 09:19 |
| apuimedo | dmellado_: either that, or we generate a config in /opt/stack/.kube/config | 09:20 |
| dmellado_ | just like the teenage mutant ninja turtles | 09:20 |
| apuimedo | and then generate it | 09:20 |
| dmellado_ | I was thinking about config.load_kube_config() | 09:20 |
| apuimedo | sorry, and then use it for both | 09:20 |
| apuimedo | dmellado_: you lazy man | 09:20 |
| apuimedo | xD | 09:20 |
| dmellado_ | yeah, I am xD | 09:20 |
| apuimedo | ok, so let's generate the config then | 09:20 |
| apuimedo | one sec | 09:20 |
| kzaitsev_ws | apuimedo: true =) | 09:21 |
| kzaitsev_ws | it was just a symptom of smth different | 09:22 |
| kzaitsev_ws | janonymous: what did you mean by "Reference should be [1]..etc currently #" I actually copied the approach from some other spec. Can convert it to more link-friendly, just gotta lookup how that's done | 09:28 |
| kzaitsev_ws | ah, you probably meant the list in the end | 09:28 |
| janonymous | yup | 09:28 |
| janonymous | kzaitsev_ws: it has hashes instead of numbers | 09:28 |
| kzaitsev_ws | well, they're translated into numbered list =) | 09:29 |
| kzaitsev_ws | the hashes I mean | 09:29 |
| *** limao has quit IRC | 09:29 | |
| janonymous | ohh cool, didn't knew that :D | 09:30 |
| kzaitsev_ws | I think there should be some syntax in rst to make it more friendly. currently [1]... are not even links. should be smth wiki-like | 09:30 |
| apuimedo | dmellado_: I'll send the patch now | 09:30 |
| dmellado_ | apuimedo: ack | 09:30 |
| apuimedo | I was just trying it out | 09:31 |
| kzaitsev_ws | ah, footnotes they're called | 09:31 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: devstack: Generate kubeconfig for kubectl and test https://review.openstack.org/469400 | 09:45 |
| apuimedo | dmellado_: ^^ | 09:45 |
| apuimedo | vikasc: ping | 09:45 |
| vikasc | apuimedo, pong | 09:46 |
| dmellado_ | apuimedo: got it | 09:46 |
| *** dmellado_ is now known as dmellado | 09:46 | |
| apuimedo | did you finally get to review those patches I asked you on monday? | 09:46 |
| apuimedo | I'd like to start merging stuff before they get stale | 09:46 |
| dmellado | apuimedo: now I need to struggle on how to mix the k8s client with the tempest plugin structure | 09:46 |
| apuimedo | https://review.openstack.org/#/c/468100/ was one of them | 09:46 |
| dmellado | I might go a bit off in this case to have at least a working test | 09:47 |
| apuimedo | dmellado: why? | 09:47 |
| apuimedo | Do you really need to wrap it? | 09:47 |
| dmellado | apuimedo: because the way a service client is defined in tempest | 09:47 |
| apuimedo | Can't you have the base class create the CoreV1Api object | 09:47 |
| dmellado | I might create a class with some methods already defined | 09:47 |
| apuimedo | put it in self.api | 09:47 |
| apuimedo | and then just use api.foo | 09:47 |
| dmellado | i.e KubernetesClient | 09:47 |
| apuimedo | from tests | 09:47 |
| vikasc | apuimedo, i am sorry , got caught up in another high priority internal work item. | 09:47 |
| dmellado | then def create_pod(foo) | 09:47 |
| dmellado | and so | 09:47 |
| vikasc | apuimedo, i will do it today for sure | 09:48 |
| dmellado | apuimedo: that'd be the fastest path | 09:48 |
| apuimedo | dmellado: which? | 09:48 |
| dmellado | so I might go for 'functionality first' | 09:48 |
| dmellado | then 'make it nice' second | 09:48 |
| dmellado | just like 'america first' but in the test case | 09:48 |
| dmellado | xD | 09:48 |
| dmellado | importing it directly | 09:48 |
| apuimedo | vikasc: can I help you somehow so we get it in the next hour or so? | 09:48 |
| apuimedo | (before lunch in my case :P ) | 09:48 |
| apuimedo | dmellado: test first | 09:48 |
| apuimedo | kzaitsev_ws: please, review https://review.openstack.org/#/c/468962/ | 09:49 |
| apuimedo | garyloug_: ping | 09:49 |
| kzaitsev_ws | apuimedo: it should work ) | 09:50 |
| kzaitsev_ws | they actually have -f shell, but I'm not sure if I like it more | 09:50 |
| vikasc | apuimedo, thanks , i will let you know if i need help. | 09:51 |
| apuimedo | ip -4 r add garyloug_ via mchiappero | 09:51 |
| apuimedo | mchiappero: ping | 09:51 |
| apuimedo | :-) | 09:51 |
| kzaitsev_ws | it gives smth-like "cidr=value" for -c cidr | 09:51 |
| apuimedo | vikasc: ok | 09:51 |
| garyloug_ | apuimedo, hi | 09:51 |
| apuimedo | garyloug_: :-) | 09:51 |
| apuimedo | garyloug_: can you give me some insight into the issues you were experiencing that caused you to w-1? | 09:52 |
| apuimedo | kzaitsev_ws: I don't like -f shell | 09:52 |
| apuimedo | :-) | 09:52 |
| dmellado | (all-plugin) [fedora@devstackrdo tempest(master)]$ tempest list-plugins | 09:52 |
| dmellado | +---------------------+--------------------------------------------------------------+ | 09:52 |
| dmellado | | Name | EntryPoint | | 09:52 |
| apuimedo | it's lazy programming | 09:52 |
| dmellado | +---------------------+--------------------------------------------------------------+ | 09:52 |
| dmellado | | keystone_tests | keystone_tempest_plugin.plugin:KeystoneTempestPlugin | | 09:52 |
| dmellado | | cinder_tests | cinder.tests.tempest.plugin:CinderTempestPlugin | | 09:52 |
| dmellado | | neutron_tests | neutron.tests.tempest.plugin:NeutronTempestPlugin | | 09:52 |
| dmellado | | neutron_lbaas | neutron_lbaas.tests.tempest.plugin:NeutronLbaasTempestPlugin | | 09:52 |
| dmellado | | kuryr_tempest_tests | kuryr_tempest_plugin.plugin:KuryrTempestPlugin | | 09:52 |
| dmellado | +---------------------+--------------------------------------------------------------+ | 09:52 |
| dmellado | \o/ | 09:52 |
| dmellado | now let's add the dreaded test | 09:52 |
| kzaitsev_ws | me to ) | 09:52 |
| apuimedo | dmellado: does that mean that to run our tests first it will run all the other tests or just that it loads the plugin in case we want to check stuff created on those? | 09:53 |
| apuimedo | btw, you can take cinder out of the list | 09:53 |
| apuimedo | we ignore it until fuxi-k8s is a thing | 09:53 |
| dmellado | apuimedo: that just means that it loads the entry point from our plugin | 09:54 |
| dmellado | alongside any another plugins that might be installed there | 09:54 |
| apuimedo | ok | 09:54 |
| apuimedo | dmellado: the good thing is that this approach and using the k8s client should work for both k8s and any downstream like openshift | 09:55 |
| apuimedo | (even for downstream only objects) | 09:55 |
| dmellado | apuimedo: yeah, that's my idea | 09:55 |
| openstackgerrit | Jaivish Kothari(janonymous) proposed openstack/kuryr-kubernetes master: [WIP]Kubernetes client usage from upstream Changed: - watch draft - configuration draft - unicode data/type change https://review.openstack.org/454555 | 09:56 |
| apuimedo | dmellado: can I get a nice +1 :-) https://review.openstack.org/#/c/469400/ | 09:56 |
| apuimedo | ? | 09:56 |
| dmellado | oh, do you deserve it? | 09:57 |
| dmellado | xD | 09:57 |
| dmellado | btw, when will you be heading off to IL? | 09:57 |
| openstackgerrit | Jaivish Kothari(janonymous) proposed openstack/kuryr-kubernetes master: [WIP]Kubernetes client usage from upstream https://review.openstack.org/454555 | 09:57 |
| apuimedo | Sunday 00:05 IIRC | 10:00 |
| dmellado | ouch | 10:02 |
| dmellado | that looks late | 10:02 |
| apuimedo | dmellado: indeed | 10:06 |
| apuimedo | red eye flight | 10:06 |
| janonymous | ivc_: i pushed a draft of change #454555 | 10:06 |
| apuimedo | it's the best I could get to go there, speak and go back home | 10:06 |
| apuimedo | thanks janonymous! | 10:06 |
| apuimedo | I'll review right away | 10:06 |
| janonymous | :D Heh.. | 10:06 |
| janonymous | not with red eyes :P | 10:06 |
| janonymous | apuimdeo: i had some questions with ivc_, so uploaded a very drft version to get comments | 10:07 |
| apuimedo | janonymous: yes. I read your exchange with ivc_ this morning | 10:08 |
| janonymous | apuimedo: would you have a link of that presentation too? | 10:10 |
| *** feisky has quit IRC | 10:11 | |
| apuimedo | janonymous: which presentation? | 10:11 |
| janonymous | apuimedo: IL one i guess | 10:11 |
| vikasc | apuimedo, changes look goog to me, https://review.openstack.org/#/c/468100/3 | 10:12 |
| vikasc | apuimedo, should we update sample localrc also? | 10:12 |
| ltomasbo | apuimedo, I tested again the devstack-heat patch, and it works | 10:12 |
| apuimedo | janonymous: yes. As soon as irenab and I deliver it I'll put it on slideshare | 10:12 |
| apuimedo | the video I don't know when it will be posted | 10:12 |
| janonymous | apuimedo: Cool! :) | 10:13 |
| ltomasbo | apuimedo, I even use it to deploy your infra k8s API patch too, and it created the lbaas without any problem | 10:13 |
| apuimedo | vikasc: we should only update the undercloud/overcloud ones | 10:13 |
| apuimedo | I'll do it in a follow-up patch | 10:13 |
| apuimedo | so that when people use overcloud, they don't add the kubelet ifaces | 10:14 |
| apuimedo | ltomasbo: I really want to merge that too | 10:14 |
| apuimedo | but there's no +2s to that | 10:14 |
| vikasc | apuimedo, cool | 10:14 |
| apuimedo | I'm considering that since it is contrib. I'll just merge it | 10:14 |
| apuimedo | (referring to devstack-heat) | 10:14 |
| apuimedo | vikasc: https://review.openstack.org/#/c/467241/ would be very useful to get merged too? | 10:20 |
| apuimedo | s/?// | 10:20 |
| vikasc | apuimedo, reviewing it | 10:24 |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: devstack: Add configuration for kubelet probes https://review.openstack.org/468100 | 10:29 |
| openstackgerrit | Merged openstack/kuryr-kubernetes master: contrib: Add devstack-heat https://review.openstack.org/466291 | 10:29 |
| *** atoth has quit IRC | 10:38 | |
| dmellado | apuimedo: it feels good when dependant patches gets merged so soon xD | 10:46 |
| *** salv-orl_ has quit IRC | 10:59 | |
| *** yamamoto has quit IRC | 11:02 | |
| *** rwallner has joined #openstack-kuryr | 11:23 | |
| *** atoth has joined #openstack-kuryr | 11:39 | |
| *** rwallner has quit IRC | 11:40 | |
| *** rwallner has joined #openstack-kuryr | 11:41 | |
| *** yamamoto has joined #openstack-kuryr | 11:43 | |
| *** rwallner has quit IRC | 11:46 | |
| *** rwallner has joined #openstack-kuryr | 11:47 | |
| *** janki has quit IRC | 12:26 | |
| *** rwallner has quit IRC | 12:30 | |
| *** rwallner has joined #openstack-kuryr | 12:31 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Nested vlan vif pool driver extension to precreate reusable subports https://review.openstack.org/436894 | 12:36 |
| ltomasbo | apuimedo, I directly included the modifications to take into account irenab's patch about config options | 13:00 |
| ltomasbo | they are moved to the vif_pool driver | 13:00 |
| *** gouthamr has joined #openstack-kuryr | 13:18 | |
| apuimedo | ltomasbo: perfect thanks! | 13:19 |
| apuimedo | s/perfect /perfect, / | 13:19 |
| openstackgerrit | Kirill Zaitsev proposed openstack/kuryr-kubernetes master: Add kuryr-sriov spec proposal https://review.openstack.org/465661 | 13:23 |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Nested vlan vif pool driver extension to precreate reusable subports https://review.openstack.org/436894 | 13:24 |
| openstackgerrit | Kirill Zaitsev proposed openstack/kuryr-kubernetes master: Add kuryr-sriov spec proposal https://review.openstack.org/465661 | 13:26 |
| *** rwallner has quit IRC | 13:28 | |
| mchiappero | apuimedo: pong | 13:39 |
| mchiappero | apuimedo: sorry for the delay | 13:39 |
| mchiappero | the change in the config file on nested is a bit slowing us down | 13:40 |
| mchiappero | well, actually I'm a bit busy with other stuff at the moment | 13:40 |
| mchiappero | there is of course a workaround that is ugly. The other options would be to merge the two MACVLAN related patches, or invert their order | 13:41 |
| mchiappero | did you manage to test is correctly BTW? | 13:41 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr master: Updated from global requirements https://review.openstack.org/462106 | 13:46 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-kubernetes master: Updated from global requirements https://review.openstack.org/469489 | 13:47 |
| *** rwallner has joined #openstack-kuryr | 13:52 | |
| mchiappero | apuimedo, irenab & co.: what would you do at this stage? In theory it would be better to refactor the hierarchy of driver first and add macvlan later. Or merge into a single change. But it's another round of reviews | 14:00 |
| openstackgerrit | Kirill Zaitsev proposed openstack/kuryr-kubernetes master: Add kuryr-sriov spec proposal https://review.openstack.org/465661 | 14:01 |
| *** salv-orlando has joined #openstack-kuryr | 14:10 | |
| *** dell has joined #openstack-kuryr | 14:13 | |
| dmellado | mchiappero: I'd go for the refactor first | 14:13 |
| dmellado | otherwise you may risk getting yourself frustrated by a ton of rebases :\ | 14:13 |
| apuimedo | mchiappero: I'd rather merge as is if it works | 14:14 |
| apuimedo | we're a bit short handed this week on core reviewers | 14:15 |
| dmellado | apuimedo: we've a creature to be ramping up | 14:15 |
| dmellado | xD | 14:15 |
| dmellado | mchiappero: in any case I'll review the patches | 14:15 |
| apuimedo | ? | 14:15 |
| * dmellado isolated himself from anything but reviews tomorrow | 14:15 | |
| apuimedo | dmellado: no more refactors | 14:15 |
| apuimedo | I want the big patch in finally | 14:16 |
| dmellado | hmmm, let's wait for the reviews xD | 14:16 |
| apuimedo | and then we can clean up | 14:16 |
| dmellado | mchiappero: pls add me to whichever approach you follow in the end and I'll review it ;) | 14:18 |
| *** hongbin has joined #openstack-kuryr | 14:25 | |
| dell | apuimedo:hi, I am zengchen. Several days ago, I committed a path which wanted to add a new sub-project of Fuxi-golang to Kuryr. In the maillist, Spyros said Rexray had integrated Cinder to supply volume for Docker and it is implemented in go language. So Fuxi-golang conflicts with it. What's your opinion about adding Fuxi-golang. Thanks. | 14:26 |
| *** dell is now known as zengchen | 14:27 | |
| apuimedo | Hi zengchen! Nice to see you here | 14:27 |
| zengchen | apuimedo:sorry, foget to change my nick name. | 14:27 |
| zengchen | apuimedo:nice to talk to you | 14:27 |
| apuimedo | zengchen: can you get me a link to this conflicting cinder golang project? | 14:27 |
| zengchen | apuimedo:ok, wait a moment. | 14:28 |
| apuimedo | brb | 14:28 |
| zengchen | apuimedo:https://github.com/codedellemc/rexray/releases/tag/v0.9.0 | 14:28 |
| apuimedo | zengchen: this is not part of openstack | 14:29 |
| zengchen | apuimedo: yes, it is. but it had finished part of work of Fuxi-golang. | 14:30 |
| apuimedo | zengchen: there's also https://github.com/j-griffith/cinder-docker-driver | 14:30 |
| apuimedo | I mean, this rexray is not on big tent or anything | 14:30 |
| zengchen | apuimedo: yes, it is. So you mean Fuxi-golang should be worked on? | 14:32 |
| zengchen | apuimedo: john had implemented the work of converting Cinder to supply volume for Docker. But imo, he will not say no to Fuxi-golang. | 14:35 |
| zengchen | apuimedo:you can see the discussing on the maillist with title of '[openstack-dev] [fuxi][kuryr] Where to commit codes for Fuxi-golang' | 14:36 |
| hongbin | apuimedo: toni, from what i understood, rexray assume you run docker on a nova instance, which is very different from what fuxi is trying to do for the baremetal use cases | 14:46 |
| hongbin | apuimedo: and yes, it is not a openstack project | 14:46 |
| zengchen | hongbin:to my understanding rexray can work on baremetal. you can see the doc on https://github.com/codedellemc/rexray/blob/master/.docs/user-guide/docker-plugins.md | 14:51 |
| kzaitsev_ws | ivc_: apuimedo: irenab: I think I'll implement just the tip of multi-vif handler, i.e. being able to work with multiple vifs, but not the exact mechanism for requesting them | 14:54 |
| kzaitsev_ws | except for the sriov, where I've defined the mechanism =) | 14:55 |
| apuimedo | sorry. I was afk | 15:02 |
| apuimedo | zengchen: ^^ | 15:02 |
| zengchen | apuimedo::) | 15:02 |
| apuimedo | hongbin: zengchen: You both know cinder much better than me | 15:03 |
| apuimedo | from my limited context, rexray looks like flocker | 15:04 |
| zengchen | apuimedo: yes, flocker is another volume plugin for docker. | 15:04 |
| apuimedo | that also supports cinder | 15:04 |
| zengchen | apuimedo:sorry, i don't know much about flocker. let me check. | 15:05 |
| apuimedo | I am fine if fuxi people decide to just contribute to rexray or flocker instead. As long as fuxi people can implement their use cases, we'll all be happy | 15:06 |
| apuimedo | jgriffith: could you weigh in on the above? | 15:07 |
| zengchen | apuimedo: ok. for me i hope fuxi-golang can be implemented. | 15:09 |
| zengchen | apuimedo: thanks very much for your suggestion. | 15:10 |
| jgriffith | apuimedo reading scroll back | 15:11 |
| apuimedo | thanks jgriffith | 15:11 |
| jgriffith | Oh, the RexRay debate huh? | 15:11 |
| jgriffith | So that's fine if people go that route, but the reality is as I pointed out on the ML, the fact that it's owned/maintained by Dell/EMC most storage folks aren't going to be contributing to that | 15:12 |
| jgriffith | so that's your trade-off | 15:12 |
| *** limao has joined #openstack-kuryr | 15:12 | |
| jgriffith | also, I'm frankly not clear why the extra layer of RexRay/Libstorage would be desirable but that's just me | 15:13 |
| apuimedo | jgriffith: I lack all the context | 15:13 |
| jgriffith | apuimedo no worries | 15:13 |
| apuimedo | I'm not up to date on storage | 15:13 |
| apuimedo | jgriffith: so this project is mostly EMC? | 15:14 |
| jgriffith | apuimedo There's RexRay, Cinder-Docker-Driver, Flocker (RIP), Rancher, and a whole host of other options out there | 15:14 |
| jgriffith | apuimedo not mostly "is" | 15:14 |
| jgriffith | which is fine, and they do a fine job | 15:14 |
| apuimedo | jgriffith: I was saying mostly in case they got outside contributions | 15:14 |
| jgriffith | but their goal is also a bit different IMO | 15:14 |
| jgriffith | They aim to be a Cinder for Containers | 15:15 |
| jgriffith | So their abstraction is a number of EMC/Dell products (and maybe others), as well as GCE, AWS and OpenStack/Cinder | 15:15 |
| zengchen | jgriffith: yes. totaly agree. | 15:15 |
| apuimedo | how do they differ from the now deceased Flocker? | 15:15 |
| jgriffith | They don't :) | 15:15 |
| apuimedo | because it sounds the same to me | 15:16 |
| jgriffith | I mean, in a nut shell | 15:16 |
| apuimedo | it's just a middle layer for any cloud storage | 15:16 |
| jgriffith | there's a lot of technical differences and implementation differences, but in terms of the goal it's pretty similar | 15:16 |
| jgriffith | apuimedo cloud or bare-metal | 15:16 |
| *** limao_ has joined #openstack-kuryr | 15:16 | |
| apuimedo | right | 15:16 |
| jgriffith | apuimedo but it also doesn't have support for things like stand-alone cinder etc | 15:17 |
| *** limao has quit IRC | 15:17 | |
| jgriffith | it's cloud-provider or not cloud-provider | 15:17 |
| apuimedo | I jgriffith: I don't think products with so many backends can be expected to give the most of cinder | 15:17 |
| jgriffith | anyway, the biggest detractors IMO are: | 15:17 |
| jgriffith | 1. The limited nature of building a community around it | 15:18 |
| jgriffith | 2. The architecture of abstractions on top of abstractions | 15:18 |
| jgriffith | apuimedo yeah, I think they do a fine job, but I think it's much more than what we'd need | 15:18 |
| jgriffith | apuimedo honestly the Docker interface impl is quite simple | 15:19 |
| jgriffith | and these days not so interesting | 15:19 |
| jgriffith | what's more interesting to folks is orchestrators like K8's or even Habitat at some point | 15:19 |
| jgriffith | and of course you can use the docker volume api's in Swarm and even DCOS contexts | 15:19 |
| jgriffith | anyway... I'm rambling :) | 15:20 |
| apuimedo | from my perspective, which lacks any enterprise ambition, what I'd like is to follow the same path we hvae with kuryr | 15:20 |
| apuimedo | one simple but powerful docker driver | 15:20 |
| apuimedo | then a good k8s integration | 15:20 |
| jgriffith | Yeah, I would really like ot be vendor agnositc | 15:20 |
| apuimedo | (both efforts unrelated) | 15:20 |
| jgriffith | and honestly the K8's component can be obtained upstream anyway | 15:20 |
| jgriffith | already a lot of work going on there | 15:20 |
| apuimedo | obviously vendor agnostic | 15:21 |
| apuimedo | zengchen: jgriffith: hongbin: I'd go forward with fuxi-golang | 15:21 |
| apuimedo | if there are things that the rexray impl has better, it's apache license 2 as well | 15:21 |
| zengchen | apuimedo: great! | 15:21 |
| apuimedo | just like j-griffith's | 15:21 |
| apuimedo | let's start anew in community | 15:22 |
| jgriffith | apuimedo I think so, honestly it's pretty simple... the missing piece is the connection stuff for stand-alone and that's missing in the alternatives as well | 15:22 |
| apuimedo | jgriffith: the brick? | 15:22 |
| apuimedo | os-brick | 15:22 |
| jgriffith | apuimedo yes.. connections outside of a nova context | 15:22 |
| apuimedo | right | 15:23 |
| apuimedo | zengchen: hongbin: jgriffith: One thing we didn't talk through for fuxi-golang is the core team | 15:23 |
| zengchen | apuimedo: jgriffith: hongbin said he will work on os-brick to reimplement it in go language. | 15:23 |
| apuimedo | I see that in the submission it has a different core team than fuxi | 15:23 |
| apuimedo | that is fine for me | 15:23 |
| apuimedo | zengchen: but I wouldn't like the core team to be single vendor | 15:24 |
| zengchen | apuimedo: we can change it to fuxi core team. | 15:24 |
| apuimedo | jgriffith should be in both fuxi-golang and the os-brick golang subprojects | 15:24 |
| apuimedo | zengchen: I don't mind about that | 15:24 |
| apuimedo | but make sure that all the parties have somebody from the start | 15:25 |
| apuimedo | otherwise this is just going to be a Huawei thing | 15:25 |
| jgriffith | apuimedo agreed | 15:25 |
| hongbin | apuimedo: i have no problem of that | 15:25 |
| jgriffith | apuimedo I plan to participate, I think others will as well | 15:25 |
| zengchen | apuimedo: agree | 15:25 |
| jgriffith | Cinder folks that is | 15:25 |
| apuimedo | jgriffith: if you have some other candidate for core, I think starting with 3-4 cores would be ideal | 15:26 |
| apuimedo | I assume hongbin will be a core too | 15:26 |
| jgriffith | apuimedo let me bring it up at cinder meeting and see if there some folks interested | 15:26 |
| apuimedo | I could be if there is really lack of reviewers, but I'm quite busy as it is | 15:26 |
| *** kzaitsev_ws has quit IRC | 15:27 | |
| apuimedo | zengchen: hongbin: are you both going to be driving this? | 15:27 |
| apuimedo | jgriffith: that sounds good | 15:27 |
| apuimedo | otherwise starting projects with lack of cores gets painful in merge times | 15:27 |
| zengchen | apuimedo: i am | 15:27 |
| hongbin | apuimedo: i will drive the core team of gos-brick, i believe zengchen will found the core team of fuxi-go | 15:27 |
| apuimedo | hongbin: you should probably get other cores from cinder/manila in gos-brick | 15:28 |
| apuimedo | too | 15:28 |
| *** kzaitsev_ws has joined #openstack-kuryr | 15:28 | |
| hongbin | apuimedo: sure, i would love to add them if some of them are interested | 15:28 |
| apuimedo | jgriffith: when is the next cinder meeting? | 15:29 |
| apuimedo | could you add that to the agenda? | 15:29 |
| jgriffith | today | 15:29 |
| jgriffith | yes | 15:29 |
| apuimedo | jgriffith: in half an hour, right? | 15:31 |
| jgriffith | Yes, agenda item added | 15:32 |
| apuimedo | thanks jgriffith | 15:32 |
| hongbin | jgriffith: i think we can add the whole cinder core team as a sub-team of gos-brick if they would love to be part of the project | 15:33 |
| jgriffith | hongbin agree | 15:36 |
| apuimedo | agreed | 15:36 |
| jgriffith | hongbin and honestly I think it should have some coupling to the Cinder team regardless | 15:36 |
| zengchen | apuimedo: Fuxi also integrate Manila. so it is better to ask some cores of Manila to join. | 15:37 |
| hongbin | jgriffith: yes, it should, for me, the goal is to hand it to cinder team after it is matured (if they like to take it). | 15:37 |
| jgriffith | hongbin I certainly hope the rest of the team is an interested as I am | 15:38 |
| hongbin | yes :) | 15:38 |
| apuimedo | :-) | 15:39 |
| hongbin | for fuxi-go, it might belong to a different team, likely to be a kuryr subteam as it is right now | 15:40 |
| hongbin | but it doesn't matter too much for me | 15:40 |
| apuimedo | it doesn't matter much to me either | 15:40 |
| apuimedo | as long as there's a healthy project in openstack opening the way for cinder and manila to containers | 15:41 |
| zengchen | apuimedo: agree. | 15:41 |
| zengchen | apuimedo: hongbin: jgriffith: i have to leave for sleeping. Thanks very much for your sugestions on Fuxi-golang. | 15:42 |
| *** rwallner has quit IRC | 15:42 | |
| apuimedo | you're welcome zengchen | 15:43 |
| apuimedo | have a good night | 15:43 |
| zengchen | apuimedo: thanks. | 15:43 |
| *** rwallner has joined #openstack-kuryr | 15:43 | |
| *** zengchen has quit IRC | 15:44 | |
| jgriffith | thank you zengchen!! | 15:51 |
| *** aojea has quit IRC | 15:59 | |
| *** limao_ has quit IRC | 16:07 | |
| openstackgerrit | Luis Tomas Bolivar proposed openstack/kuryr-kubernetes master: Nested vlan vif pool driver extension to precreate reusable subports https://review.openstack.org/436894 | 16:12 |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr-kubernetes master: devstack: Use config file generation https://review.openstack.org/469557 | 16:21 |
| apuimedo | dmellado: ^^ | 16:22 |
| *** ltomasbo is now known as ltomasbo|away | 16:32 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr master: Updated from global requirements https://review.openstack.org/462106 | 16:37 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr-libnetwork master: Updated from global requirements https://review.openstack.org/461376 | 16:37 |
| *** rwallner has quit IRC | 16:38 | |
| *** pcaruana has quit IRC | 16:50 | |
| *** rwallner has joined #openstack-kuryr | 16:55 | |
| *** salv-orlando has quit IRC | 16:55 | |
| *** rwallner has quit IRC | 16:58 | |
| *** rwallner has joined #openstack-kuryr | 17:02 | |
| *** egonzalez has quit IRC | 17:09 | |
| *** garyloug_ has quit IRC | 17:24 | |
| *** aojea has joined #openstack-kuryr | 17:29 | |
| *** tonanhngo has joined #openstack-kuryr | 17:31 | |
| *** aojea has quit IRC | 17:50 | |
| *** aojea has joined #openstack-kuryr | 17:51 | |
| *** atoth has quit IRC | 17:53 | |
| *** aojea has quit IRC | 17:55 | |
| *** aojea has joined #openstack-kuryr | 17:57 | |
| *** aojea has quit IRC | 18:03 | |
| *** aojea has joined #openstack-kuryr | 18:18 | |
| *** aojea has quit IRC | 18:22 | |
| *** dimak_ has joined #openstack-kuryr | 19:01 | |
| *** rwallner has quit IRC | 19:12 | |
| *** rwallner has joined #openstack-kuryr | 19:16 | |
| *** rwallner_ has joined #openstack-kuryr | 19:17 | |
| *** rwallner has quit IRC | 19:21 | |
| *** aojea has joined #openstack-kuryr | 19:28 | |
| *** yamamoto has quit IRC | 19:33 | |
| *** dimak_ has quit IRC | 19:37 | |
| *** egonzalez has joined #openstack-kuryr | 19:45 | |
| *** pcaruana has joined #openstack-kuryr | 19:45 | |
| *** rwallner_ has quit IRC | 19:53 | |
| *** salv-orlando has joined #openstack-kuryr | 20:00 | |
| *** pcaruana has quit IRC | 20:02 | |
| *** salv-orlando has quit IRC | 20:09 | |
| *** salv-orlando has joined #openstack-kuryr | 20:11 | |
| *** yamamoto has joined #openstack-kuryr | 20:33 | |
| *** yamamoto has quit IRC | 20:43 | |
| *** pcaruana has joined #openstack-kuryr | 21:12 | |
| *** aojea has quit IRC | 21:13 | |
| *** rwallner has joined #openstack-kuryr | 21:14 | |
| *** salv-orlando has quit IRC | 21:16 | |
| *** rwallner has quit IRC | 21:18 | |
| openstackgerrit | Frederick F. Kautz IV proposed openstack/kuryr-kubernetes master: Add support for USE_SCREEN and default to screen if USE_SYSTEMD is not set https://review.openstack.org/469280 | 21:28 |
| *** pcaruana has quit IRC | 21:29 | |
| hongbin | jgriffith: apuimedo : john and cinder ptl were added to this group, please feel free to add someone who also interest in this project | 21:47 |
| hongbin | jgriffith: apuimedo https://review.openstack.org/#/admin/groups/1781,members | 21:47 |
| openstackgerrit | Frederick F. Kautz IV proposed openstack/kuryr-kubernetes master: Add support for USE_SCREEN. https://review.openstack.org/469280 | 22:06 |
| *** neiljerram has quit IRC | 22:09 | |
| *** gouthamr has quit IRC | 22:20 | |
| *** egonzalez has quit IRC | 22:21 | |
| *** limao has joined #openstack-kuryr | 22:39 | |
| *** limao_ has joined #openstack-kuryr | 22:42 | |
| *** limao has quit IRC | 22:44 | |
| *** gouthamr has joined #openstack-kuryr | 23:11 | |
| *** limao_ has quit IRC | 23:40 | |
| *** limao has joined #openstack-kuryr | 23:44 | |
| *** limao has quit IRC | 23:50 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!