| *** yamamoto has joined #openstack-kuryr | 00:02 | |
| fkautz | apuimedo: the credential changes with each run? is this for a multi-tenant use case? | 00:02 |
|---|---|---|
| openstackgerrit | Akash Gangil proposed openstack/kuryr: Enabled neutronclient debug messages https://review.openstack.org/311893 | 00:05 |
| *** salv-orlando has joined #openstack-kuryr | 00:07 | |
| *** yamamoto has quit IRC | 00:09 | |
| *** abhiraut has joined #openstack-kuryr | 00:13 | |
| *** salv-orlando has quit IRC | 00:14 | |
| *** salv-orlando has joined #openstack-kuryr | 00:14 | |
| apuimedo | fkautz: you meant the opt for docker engine? | 00:14 |
| *** salv-orlando has quit IRC | 00:15 | |
| *** salv-orlando has joined #openstack-kuryr | 00:16 | |
| fkautz | yea | 00:17 |
| fkautz | i'm going to reach out to docker about it, want to make sure i have the story straight | 00:17 |
| apuimedo | fkautz: thanks | 00:18 |
| apuimedo | well, I'm not sure docker opts reach all the way to libnetwork at all | 00:18 |
| apuimedo | I thought you meant docker network opt labels | 00:19 |
| fkautz | there are a set of network opts that i think should reach, e.g. docker network create --opt | 00:19 |
| *** salv-orlando has quit IRC | 00:22 | |
| apuimedo | fkautz: maybe I misunderstand, but IIUC, you refer to --opt/-o in docker-network-create | 00:24 |
| apuimedo | is that right? | 00:24 |
| fkautz | That's correct | 00:24 |
| fkautz | There is no net opt in docker run | 00:25 |
| fkautz | But there is in docker network create | 00:25 |
| apuimedo | this was my initial idea, but the fact that we don't get it for `run` is a bit problematic :P | 00:25 |
| *** salv-orlando has joined #openstack-kuryr | 00:25 | |
| fkautz | I think they are trying to maintain a uniform interface for running containers | 00:26 |
| fkautz | There is also no options for storage drivers | 00:26 |
| fkautz | But there is for logging | 00:26 |
| apuimedo | fkautz: that's why the "docker login-like" proposal is good. It would not modify the interface at all | 00:27 |
| apuimedo | it would set up options in .docker/config.json | 00:27 |
| apuimedo | and if it matches the name of a driver (storage or network), it would be used for operations | 00:28 |
| apuimedo | that involve that driver | 00:28 |
| *** abhiraut has quit IRC | 00:29 | |
| fkautz | So you are looking for a way to add metadata that is operational data on run | 00:30 |
| fkautz | And that is user, not system specific | 00:31 |
| *** abhiraut has joined #openstack-kuryr | 00:31 | |
| fkautz | And only if the driver matches | 00:31 |
| apuimedo | exactly | 00:32 |
| apuimedo | fkautz: that's the best idea I had so far | 00:32 |
| fkautz | The security model is still weak, but that's due to docker enabling privileged if you have access to the socket | 00:34 |
| apuimedo | yeah | 00:34 |
| fkautz | Can't do much about that in this context | 00:34 |
| fkautz | I'll pitch the idea to them | 00:34 |
| fkautz | They might still reject since network create opt is available | 00:35 |
| apuimedo | fkautz: great | 00:35 |
| fkautz | And it is technically workable | 00:36 |
| fkautz | It's also an interface change so it'll be closely scrutinized | 00:36 |
| apuimedo | fkautz: the counter argument to that is that network and storage drivers perform operations on run and others that could use some security | 00:36 |
| apuimedo | the good thing is, it is backwards compatible | 00:37 |
| *** gangil has quit IRC | 00:37 | |
| *** abhiraut has quit IRC | 00:37 | |
| fkautz | Agreed, going to argue future use cases and backwards compatibility | 00:37 |
| apuimedo | :-) | 00:37 |
| apuimedo | let me know if I can help. But that will have to be tomorrow. Or I'll never make it to the office before noon :P | 00:38 |
| apuimedo | I thought maybe the return jet lag would have me up until the kuryr meeting time | 00:38 |
| apuimedo | but it seems not :P | 00:38 |
| fkautz | No worries | 00:39 |
| fkautz | Thanks for answering sooner than I expected :p | 00:39 |
| *** diogogmt has quit IRC | 00:40 | |
| apuimedo | fkautz: I almost missed the message though. I should finally do something about the irc bouncer :P | 00:41 |
| apuimedo | but before heading to bed I remembered I heard some beep while watching games of thrones earlier :P | 00:41 |
| *** abhiraut has joined #openstack-kuryr | 00:41 | |
| *** diogogmt has joined #openstack-kuryr | 00:42 | |
| apuimedo | fkautz: you told me a hundred times, so sorry about that... The one you use is irc cloud, right? | 00:42 |
| *** abhiraut has quit IRC | 00:43 | |
| *** gangil has joined #openstack-kuryr | 00:43 | |
| *** gangil has joined #openstack-kuryr | 00:43 | |
| fkautz | That's right | 00:44 |
| fkautz | apuimedo: irccloud which I think is UK based | 00:45 |
| apuimedo | thanks | 00:46 |
| *** apuimedo has quit IRC | 00:46 | |
| *** fawadkhaliq has quit IRC | 00:50 | |
| *** gangil has quit IRC | 00:56 | |
| *** gangil has joined #openstack-kuryr | 00:58 | |
| *** gangil has joined #openstack-kuryr | 00:58 | |
| *** salv-orlando has quit IRC | 01:04 | |
| gangil | fkautz: hi, I pushed a patch https://review.openstack.org/#/c/311893/ for the bug in the commit message. It was a pretty simple fix and I don't see taku online. just fyi. | 01:10 |
| *** abhiraut has joined #openstack-kuryr | 01:13 | |
| *** abhiraut has quit IRC | 01:17 | |
| *** banix has quit IRC | 01:28 | |
| fkautz | gangil: thanks, I'll take a look when I'm back home | 01:29 |
| gangil | Also, I am just wondering is there a reason we use flask instead of oslo.service? Or more like it was the first prototype version and its been working alright? | 01:32 |
| fkautz | Best to ask apuimedo that question | 01:34 |
| fkautz | gangil: I'm not too familiar with what oslo provides in this case. I assume it provides a rest api? Flask might be used to accurately produce the libnetwork rest api. | 01:38 |
| fkautz | I assume oslo is still useful for systemd support and other utilities | 01:38 |
| *** gangil has quit IRC | 01:39 | |
| *** itsuugo has quit IRC | 01:53 | |
| *** diogogmt has quit IRC | 02:20 | |
| *** yamamoto has joined #openstack-kuryr | 02:24 | |
| *** yamamoto has quit IRC | 02:37 | |
| *** yamamoto has joined #openstack-kuryr | 02:50 | |
| *** mspreitz has joined #openstack-kuryr | 03:00 | |
| mspreitz | time for the weekly IRC now, right? | 03:03 |
| mspreitz | Is anybody alive in here? | 03:10 |
| *** mspreitz has quit IRC | 03:17 | |
| *** wanghua has joined #openstack-kuryr | 03:21 | |
| *** diogogmt has joined #openstack-kuryr | 03:25 | |
| *** yamamoto has quit IRC | 03:26 | |
| *** oanson has joined #openstack-kuryr | 04:03 | |
| *** gangil has joined #openstack-kuryr | 04:14 | |
| *** gangil has joined #openstack-kuryr | 04:14 | |
| *** yamamoto has joined #openstack-kuryr | 04:23 | |
| *** yamamoto has quit IRC | 04:23 | |
| *** oanson has quit IRC | 04:47 | |
| *** irenab has joined #openstack-kuryr | 05:06 | |
| *** yamamoto has joined #openstack-kuryr | 05:24 | |
| *** gangil has quit IRC | 05:34 | |
| *** yamamoto has quit IRC | 05:40 | |
| *** yamamoto has joined #openstack-kuryr | 06:08 | |
| *** salv-orlando has joined #openstack-kuryr | 06:11 | |
| *** gangil has joined #openstack-kuryr | 06:11 | |
| *** gangil has joined #openstack-kuryr | 06:11 | |
| *** yamamoto has quit IRC | 06:23 | |
| *** gangil has quit IRC | 06:35 | |
| *** salv-orlando has quit IRC | 07:04 | |
| *** wangfeng has joined #openstack-kuryr | 07:46 | |
| *** yuanying has joined #openstack-kuryr | 07:55 | |
| *** fkautz has quit IRC | 07:58 | |
| *** flaviof has quit IRC | 07:58 | |
| *** huats__ has joined #openstack-kuryr | 07:58 | |
| *** yuanying_ has quit IRC | 07:58 | |
| *** huats has quit IRC | 07:58 | |
| *** huats__ is now known as huats | 07:59 | |
| *** fkautz has joined #openstack-kuryr | 08:00 | |
| *** flaviof has joined #openstack-kuryr | 08:02 | |
| *** wangfeng has quit IRC | 08:27 | |
| *** itsuugo has joined #openstack-kuryr | 08:34 | |
| *** oanson has joined #openstack-kuryr | 09:20 | |
| *** oanson has quit IRC | 09:27 | |
| *** apuimedo has joined #openstack-kuryr | 09:44 | |
| apuimedo | huats: ping | 10:41 |
| huats | apuimedo: hello | 10:41 |
| apuimedo | huats: do you have that ovs kuryr environment up and running? | 10:42 |
| huats | actually I have :) | 10:42 |
| apuimedo | nice! | 10:42 |
| apuimedo | Could you check one thing for the Kubernetes design? | 10:42 |
| apuimedo | Could you try creating a network with a subnet | 10:42 |
| apuimedo | with neutron | 10:43 |
| apuimedo | then create a load balancer | 10:43 |
| apuimedo | with a VIP in that subnet | 10:43 |
| huats | but I have deployed it many times yesterday since I was working on a bug (the possibility to define the port for etc and docker-engine) | 10:43 |
| huats | I can do that | 10:43 |
| apuimedo | and finally start a couple of containers in another neutron net | 10:43 |
| apuimedo | and use neutron to add them as members of the load balancer | 10:43 |
| huats | How is it related to kubernetes (on the thing you have detailled I don't see the relationship) | 10:44 |
| huats | ? | 10:44 |
| huats | (just to know) | 10:44 |
| apuimedo | huats: :-) | 10:44 |
| apuimedo | mspreitzer is asking in the mailing list if the load balancer usage that I described on the presentation and work session works with ovs | 10:45 |
| huats | ok | 10:45 |
| huats | let me a few minutes :) | 10:45 |
| apuimedo | you can use https://hub.docker.com/r/celebdor/webinar/tags/ | 10:45 |
| apuimedo | as the test container if you want | 10:45 |
| apuimedo | http://lists.openstack.org/pipermail/openstack-dev/2016-May/093666.html | 10:47 |
| apuimedo | this was the question | 10:47 |
| apuimedo | I'll send an answer now. Maybe you can reply | 10:47 |
| apuimedo | when you test it | 10:48 |
| huats | I will | 10:48 |
| apuimedo | thanks! | 10:58 |
| *** banix has joined #openstack-kuryr | 11:34 | |
| *** banix has quit IRC | 11:36 | |
| huats | apuimedo: tell me once you are back | 12:01 |
| *** asti has joined #openstack-kuryr | 12:17 | |
| *** apuimedo has quit IRC | 12:28 | |
| *** apuimedo has joined #openstack-kuryr | 12:28 | |
| apuimedo | huats: I'm back | 12:29 |
| huats | ok | 12:29 |
| huats | apuimedo: just to let you know that I am rebuilding my env since the previous one didn't have lbaas installed (it is not the case on the local.conf that is given in exemple in kuryr) | 12:29 |
| huats | it won't be long before it is up and running | 12:30 |
| apuimedo | cool ;-) | 12:30 |
| *** wangfeng has joined #openstack-kuryr | 12:36 | |
| openstackgerrit | Christophe Sauthier proposed openstack/kuryr: Enable the change of docker-engine and etcd listen TCP port in devstack https://review.openstack.org/312043 | 12:44 |
| *** apuimedo has quit IRC | 12:51 | |
| *** apuimedo has joined #openstack-kuryr | 12:51 | |
| apuimedo | thanks for the patch huats | 12:53 |
| *** gsagie has joined #openstack-kuryr | 13:01 | |
| *** wanghua has quit IRC | 13:04 | |
| *** oanson has joined #openstack-kuryr | 13:07 | |
| *** salv-orlando has joined #openstack-kuryr | 13:25 | |
| *** salv-orlando has quit IRC | 13:32 | |
| irenab | gsagie: apuimedo is there any etherpad notes from the OS Thursday afternoon kuryr sessions? | 13:49 |
| apuimedo | there is at least one | 13:49 |
| apuimedo | but I didn't open the laptop, you'll have to ask fawad | 13:49 |
| *** oanson has quit IRC | 13:58 | |
| *** diogogmt has quit IRC | 13:58 | |
| irenab | it is not linked to the summit etherpad page | 14:00 |
| *** apuimedo_ has joined #openstack-kuryr | 14:03 | |
| *** banix has joined #openstack-kuryr | 14:03 | |
| *** apuimedo has quit IRC | 14:07 | |
| *** apuimedo_ is now known as apuimedo | 14:07 | |
| *** apuimedo has quit IRC | 14:10 | |
| *** apuimedo has joined #openstack-kuryr | 14:24 | |
| *** diogogmt has joined #openstack-kuryr | 14:30 | |
| *** diogogmt has quit IRC | 14:32 | |
| *** irenab_ has joined #openstack-kuryr | 14:37 | |
| *** irenab has quit IRC | 14:39 | |
| *** irenab_ is now known as irenab | 14:39 | |
| *** fawadkhaliq has joined #openstack-kuryr | 14:51 | |
| apuimedo | fkautz: ok. I just finished setting up my bouncer | 15:00 |
| apuimedo | let's see how that goes | 15:00 |
| *** diogogmt has joined #openstack-kuryr | 15:06 | |
| huats | apuimedo: I am currently working on the ovs lbaas thing | 15:10 |
| apuimedo | huats: great! | 15:10 |
| huats | but I have a hard time to find time to do it :) | 15:10 |
| apuimedo | no worries ;-) | 15:11 |
| apuimedo | we're all busy | 15:11 |
| huats | In the worst case I'll have some time tonight | 15:11 |
| apuimedo | +1 | 15:16 |
| *** salv-orlando has joined #openstack-kuryr | 15:36 | |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr: Move libnetwork specific options to its own group https://review.openstack.org/308738 | 15:43 |
| apuimedo | gsagie: banix: check ^^ out | 15:44 |
| apuimedo | the sooner I get this one out of the way, the sooner I can start the rebase and push of the k8s patches | 15:45 |
| banix | apuimedo: will do | 15:45 |
| banix | apuimedo: have a minute? | 15:45 |
| banix | i was thinking about the router option you were mentioning | 15:45 |
| apuimedo | banix: on the phone and walking, but yes | 15:45 |
| banix | chasing your son? | 15:45 |
| banix | :) | 15:45 |
| banix | i’ll write you rreply when you get a chance; no rush | 15:46 |
| banix | so i was wondering iof the cleanest approach would be to simply expect the existing network have been connected to a router if required | 15:46 |
| banix | otherwise we have to also cover a way to connect the router to external network etc. | 15:47 |
| banix | it seems to me we need to do some of these out of band anyway and doing it for connecting to a router though docker network create option may not be the best thing to do | 15:47 |
| apuimedo | Walking to the bus | 15:48 |
| banix | let me know what you think when you get a chance | 15:48 |
| banix | do not walk and text :) | 15:48 |
| apuimedo | I think in case we specify a network, we should fail if a router is specified. The router opt should only be for creation | 15:49 |
| banix | when you say “router option should only be for creation” you mean creation of network. right? | 15:50 |
| apuimedo | Yes, for implicit neutron network creation | 15:50 |
| banix | so I am questioning that. let’s say we do that. how to we set the router gateway? do we specify that in options too? | 15:51 |
| apuimedo | Nah, the router should already have a gateway set. We are not creating it | 15:53 |
| apuimedo | Or that's how I imagined it | 15:53 |
| apuimedo | You want to optionally create them? | 15:54 |
| apuimedo | Mmm... That sounds useful too | 15:54 |
| apuimedo | In that case, couldn't we just say kuryr.router.gateway | 15:56 |
| apuimedo | And that would mean implicit gateway creation and attachment? | 15:56 |
| banix | apuimedo: well i am arguing for the opposite. i am actually saying that for such cases it may be better to have an existing Neutron network connected to router and then we just specify the network and use it for a docker network | 15:56 |
| apuimedo | That's currently possible already | 15:57 |
| banix | I am thinking if we want to specify a router, then we may need to specify the gateway, and more and more | 15:57 |
| banix | yes that’s correct | 15:57 |
| banix | currently we can do that | 15:57 |
| banix | so i am saying if this will be a better way of dealing with what you have been describing | 15:57 |
| banix | just a different way of doing it | 15:58 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kuryr: Updated from global requirements https://review.openstack.org/304896 | 15:58 |
| apuimedo | I don't know. From where I look at it. Allowing plugging to existing routers (that already have gateway) or getting New routers with a specified gateway seems pretty useful to minimize neutron command usage by tenants | 15:59 |
| banix | ok i hear you | 16:00 |
| apuimedo | It all depends the kind of ux we want to give in swarm | 16:00 |
| banix | yeah | 16:00 |
| apuimedo | And I can't really substitute operator input for this :P | 16:01 |
| apuimedo | I'm getting good at irc while avoiding lamp posts | 16:02 |
| *** salv-orlando has quit IRC | 16:03 | |
| banix | apuimedo: get home safe man! :) | 16:04 |
| fawadkhaliq | apuimedo: +1 on the gateway specification, I added this in plumgrid libnetwork stuff and users loved it. | 16:06 |
| apuimedo | Lamp posts are at disadvantage. Anyway I'm in the bus stop now | 16:06 |
| apuimedo | Let's hope my bouncer behaves | 16:06 |
| banix | fawadkhaliq: what did people love? having the option to specify the router stuff including gateway when creating Docker networks? | 16:07 |
| fawadkhaliq | banix: that's correct. specify the target router | 16:08 |
| fawadkhaliq | --router | 16:08 |
| banix | fawadkhaliq: ok thx for the input | 16:08 |
| fawadkhaliq | banix: yw | 16:08 |
| *** gangil has joined #openstack-kuryr | 16:09 | |
| *** gangil has joined #openstack-kuryr | 16:09 | |
| *** wangfeng has quit IRC | 16:17 | |
| *** fawadkhaliq has quit IRC | 16:22 | |
| gangil | apuimedo: could you throw some light on why we use flask instead of oslo services? | 16:39 |
| apuimedo | I'll try to when I get home | 16:49 |
| *** gangil has quit IRC | 17:17 | |
| *** gangil has joined #openstack-kuryr | 17:25 | |
| *** gangil has quit IRC | 17:25 | |
| *** gangil has joined #openstack-kuryr | 17:25 | |
| *** salv-orlando has joined #openstack-kuryr | 17:27 | |
| *** salv-orlando has quit IRC | 17:53 | |
| *** gangil has quit IRC | 18:02 | |
| *** gangil has joined #openstack-kuryr | 18:12 | |
| *** gangil has quit IRC | 18:12 | |
| *** gangil has joined #openstack-kuryr | 18:12 | |
| *** gangil has quit IRC | 18:31 | |
| *** salv-orlando has joined #openstack-kuryr | 18:40 | |
| *** salv-orlando has quit IRC | 18:45 | |
| *** salv-orlando has joined #openstack-kuryr | 18:48 | |
| *** salv-orlando has quit IRC | 18:48 | |
| *** salv-orlando has joined #openstack-kuryr | 18:51 | |
| *** salv-orlando has quit IRC | 18:54 | |
| *** banix has quit IRC | 18:59 | |
| *** banix has joined #openstack-kuryr | 19:20 | |
| *** trey has quit IRC | 19:42 | |
| *** banix has quit IRC | 19:42 | |
| *** gsagie has quit IRC | 19:42 | |
| *** banix has joined #openstack-kuryr | 19:51 | |
| *** gsagie has joined #openstack-kuryr | 19:51 | |
| *** trey has joined #openstack-kuryr | 19:56 | |
| *** banix has quit IRC | 20:24 | |
| *** salv-orlando has joined #openstack-kuryr | 20:44 | |
| *** salv-orlando has quit IRC | 20:50 | |
| *** salv-orlando has joined #openstack-kuryr | 20:54 | |
| *** gangil has joined #openstack-kuryr | 21:05 | |
| *** salv-orlando has quit IRC | 21:37 | |
| huats | apuimedo: I am trying to do what you asked me earlier | 21:41 |
| huats | can you explain me what is on your container ? so that I can try to reproduce things ? | 21:41 |
| *** asti has quit IRC | 21:44 | |
| *** salv-orlando has joined #openstack-kuryr | 21:51 | |
| *** huikang has joined #openstack-kuryr | 21:51 | |
| *** gangil has quit IRC | 22:04 | |
| *** salv-orlando has quit IRC | 22:06 | |
| *** salv-orlando has joined #openstack-kuryr | 22:15 | |
| *** huikang has quit IRC | 22:20 | |
| *** salv-orlando has quit IRC | 22:28 | |
| *** salv-orlando has joined #openstack-kuryr | 22:28 | |
| openstackgerrit | Antoni Segura Puimedon proposed openstack/kuryr: Move libnetwork specific options to its own group https://review.openstack.org/308738 | 22:34 |
| *** diogogmt has quit IRC | 22:35 | |
| *** salv-orlando has quit IRC | 22:42 | |
| *** apuimedo has quit IRC | 22:54 | |
| *** apuimedo has joined #openstack-kuryr | 22:54 | |
| *** diogogmt has joined #openstack-kuryr | 23:28 | |
| *** apuimedo has quit IRC | 23:30 | |
| *** apuimedo has joined #openstack-kuryr | 23:37 | |
| *** yamamoto has joined #openstack-kuryr | 23:47 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!