| *** yanyanhu has joined #openstack-higgins | 01:50 | |
| *** yuanying has quit IRC | 02:55 | |
| *** coolsvap is now known as coolsvap_ | 03:43 | |
| *** yuanying has joined #openstack-higgins | 03:50 | |
| *** Wenzhi_ has joined #openstack-higgins | 04:19 | |
| *** Qiming has quit IRC | 04:22 | |
| *** Wenzhi has quit IRC | 04:23 | |
| *** Qiming has joined #openstack-higgins | 04:23 | |
| *** coolsvap_ has quit IRC | 04:51 | |
| *** coolsvap has joined #openstack-higgins | 05:21 | |
| *** sheel has quit IRC | 05:45 | |
| *** yanyanhu has quit IRC | 08:54 | |
| *** coolsvap has quit IRC | 12:52 | |
| *** hongbin has joined #openstack-higgins | 13:53 | |
| hongbin | dims: there? | 13:59 |
|---|---|---|
| dims | pong hongbin | 14:33 |
| hongbin | dims: I want to discuss with you about : https://etherpad.openstack.org/p/zun-architecture-decisions | 14:36 |
| hongbin | dims: I did a research on optioin #1 | 14:36 |
| hongbin | Frankly, I couldn't find anything we can do for that, besides a wrapper of the COE api | 14:37 |
| dims | hongbin right, that's the first job. then we layer on an "application" concept that uses the CRUD to the COE's | 14:38 |
| hongbin | OK | 14:38 |
| dims | back in a bit. we're getting a new sofa :) | 14:38 |
| hongbin | I see | 14:38 |
| hongbin | Let me think | 14:39 |
| hongbin | dims: But for the wrapper of COEs. It seems a library is more suitable than a service | 14:40 |
| hongbin | dims: Because COE already store all the states in their data store, The wrapper dont' need to store it twice | 14:40 |
| hongbin | Otherwise, it wont' scale at all | 14:40 |
| hongbin | Then, it looks a library is more suitable for wrapping the APIs | 14:41 |
| hongbin | Like libvirt | 14:41 |
| hongbin | dims: ^^ | 14:42 |
| dims | hongbin : right, that's where i was starting in that poc code | 14:48 |
| dims | hongbin : you don't need to convince me to go with another design option :) whatever works for whoever is working on this stuff is ok by me | 14:49 |
| hongbin | dims: Yes, but I want to communicate with you about what I think, and see if I am thinking the wrong way :) | 14:51 |
| hongbin | Frankly, the more I researched, the more I believed that container runtimes and COEs are totally two different entities | 14:55 |
| dims | if you think of them as layers....docker/rkt first then swarm/k8s/mesos...then comes helm(https://helm.sh/) and docker dab (https://github.com/docker/docker/blob/master/experimental/docker-stacks-and-bundles.md). We should be in that space | 14:55 |
| hongbin | Maybe it is better to treat them separatedly | 14:55 |
| dims | hongbin : yes, we should forget about the docker/rkt directly | 14:56 |
| hongbin | ack | 14:56 |
| hongbin | Will do more research on that | 14:57 |
| *** irenab has quit IRC | 15:47 | |
| *** irenab has joined #openstack-higgins | 15:49 | |
| *** openstackstatus has quit IRC | 18:40 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!