| *** oanson has joined #openstack-dragonflow | 01:30 | |
| openstackgerrit | yuan wei proposed openstack/dragonflow: Reset related conntrack entries when the configuration about SG changed https://review.openstack.org/393710 | 01:42 |
|---|---|---|
| *** oanson has quit IRC | 01:53 | |
| openstackgerrit | hujie proposed openstack/dragonflow: Implement db consistency logic in df local controller https://review.openstack.org/336377 | 02:12 |
| *** gongysh has joined #openstack-dragonflow | 02:12 | |
| *** rajivk has joined #openstack-dragonflow | 02:55 | |
| *** gongysh has quit IRC | 03:36 | |
| *** rajivk has quit IRC | 03:36 | |
| *** rajivk has joined #openstack-dragonflow | 03:49 | |
| openstackgerrit | yuan wei proposed openstack/dragonflow: Reset related conntrack entries when the configuration about SG changed https://review.openstack.org/393710 | 05:23 |
| openstackgerrit | rajiv proposed openstack/dragonflow: Moved config options of metadata_service_app https://review.openstack.org/394282 | 06:23 |
| *** gampel has joined #openstack-dragonflow | 06:48 | |
| *** irenab has joined #openstack-dragonflow | 06:56 | |
| *** dimak has joined #openstack-dragonflow | 07:01 | |
| *** gongysh has joined #openstack-dragonflow | 07:12 | |
| *** yuval has joined #openstack-dragonflow | 07:32 | |
| openstackgerrit | Christian Schulze-Wiehenbrauk proposed openstack/dragonflow: Fixes sphinx warnings and errors. https://review.openstack.org/394304 | 07:48 |
| openstackgerrit | Christian Schulze-Wiehenbrauk proposed openstack/dragonflow: Fixes sphinx warnings and errors. https://review.openstack.org/394305 | 07:51 |
| ntr0py | I abandoned the first change, forgot to use other branch. | 07:53 |
| openstackgerrit | hujie proposed openstack/dragonflow: Implement db consistency logic in df local controller https://review.openstack.org/336377 | 08:33 |
| *** oanson has joined #openstack-dragonflow | 08:41 | |
| *** openstackgerrit has quit IRC | 08:48 | |
| *** openstackgerrit has joined #openstack-dragonflow | 08:49 | |
| *** xiaohhui has joined #openstack-dragonflow | 08:54 | |
| *** xiaohhui has left #openstack-dragonflow | 08:55 | |
| *** jingting has joined #openstack-dragonflow | 08:57 | |
| *** lihi has joined #openstack-dragonflow | 08:58 | |
| *** zenoway has joined #openstack-dragonflow | 08:59 | |
| oanson | Hi | 08:59 |
| oanson | Weekly meeting starts in a minute, in #openstack-meeting-4 | 09:00 |
| oanson | I'd be happy if everyone could join | 09:00 |
| *** saggi has joined #openstack-dragonflow | 09:01 | |
| oanson | nick-ma_, you in? | 09:01 |
| *** yuli_s has joined #openstack-dragonflow | 09:01 | |
| openstackgerrit | yuan wei proposed openstack/dragonflow: add active detection app for allowed address pairs https://review.openstack.org/339975 | 09:10 |
| *** xiaohhui has joined #openstack-dragonflow | 10:00 | |
| xiaohhui | I want to bring this review out https://review.openstack.org/#/c/339975/ | 10:01 |
| oanson | xiaohhui, one we have a Newton branch (I'll take care of is ASAP), we can back port important patches | 10:01 |
| oanson | I suggest we'll discuss it once the patch is merged into master | 10:01 |
| xiaohhui | This is actually the first patch of this functionality. I think we don't need backport, but have it merged, we can proceed further work | 10:02 |
| oanson | rajivk, nick-ma_, xiaohhui, can we continue the discussion of bug 1638151 | 10:02 |
| openstack | bug 1638151 in DragonFlow "Router schedule error in L3 router plugin as there are multi-external network" [High,New] https://launchpad.net/bugs/1638151 - Assigned to rajiv (rajiv-kumar) | 10:02 |
| xiaohhui | Currently, df has a configuration external_network_bridge | 10:03 |
| oanson | yes | 10:03 |
| xiaohhui | so only one external provider network is supported, | 10:03 |
| xiaohhui | That is my understanding. | 10:04 |
| oanson | I searched the code on github. I can see the bridge only used in dnat. | 10:05 |
| oanson | The bridge name, that is | 10:05 |
| xiaohhui | That may not be the root cause to the bug, but I think we need to make sure if we support such case "multi-external network" | 10:05 |
| oanson | I guess it should be used in snat too, but that still uses Neutron code + Network node | 10:05 |
| oanson | xiaohhui, sounds like an enhancement for dnat application. | 10:05 |
| oanson | In general we need a better snat implementation. | 10:06 |
| oanson | Let me see if I can find someone to write it for us. I may have a contact | 10:06 |
| xiaohhui | I am thinking of distributed snat recently, | 10:06 |
| oanson | xiaohhui, yes. | 10:06 |
| dimak | oanson, I don't think we the external bridge code anymore | 10:06 |
| oanson | In the summit, we discussed the option of a double masquerade snat | 10:07 |
| oanson | dimak, very possibly. The code has to be reviewed | 10:07 |
| oanson | The first step is to NAT the packet with the compute node's | 10:07 |
| dimak | I drop some of it in my dnat app patch | 10:07 |
| dimak | but I can see if we can drop the rest of references | 10:08 |
| oanson | The second stage is the router NATing the packet with its external IP | 10:08 |
| xiaohhui | you mean one external IP for all, or one external IP for each compute node? | 10:08 |
| oanson | I don't think we can do without it at all. We need to know via which interface to egress | 10:08 |
| oanson | xiaohhui, technically both. From the compute node's point of view, one external IP for each compute node | 10:09 |
| oanson | If the gateway wants a single IP address for all, then it NATs a second time. | 10:09 |
| oanson | It's not the best of solutions, but it's very simple | 10:09 |
| oanson | As a reference implementation, it should do | 10:09 |
| xiaohhui | so, the nat will take lots of external ips, if there is many compute nodes | 10:10 |
| oanson | Not if the gateway NATs again | 10:11 |
| oanson | Or segments the NAT. Or deals with it outside of dragonflow :) | 10:11 |
| oanson | If the compute nodes have private/hidden IPs, then the external IP will be the gateway's | 10:11 |
| oanson | So that's not a problem | 10:11 |
| xiaohhui | Why not just centralized snat in one node, with scheduler to even the load. | 10:12 |
| oanson | xiaohhui, because then its not fully distributed :) | 10:12 |
| oanson | That could be another snat implementation. But I think it would be better to have at least one fully distributed version | 10:12 |
| xiaohhui | so, in your case, the external IP will be in one df controller node, or just a virtual IP? | 10:13 |
| oanson | depends on what you mean by virtual IP? | 10:13 |
| *** gongysh has quit IRC | 10:13 | |
| oanson | It will be one of the DF controller nodes (The one hosting the VM) or the cloud's gateway IP | 10:13 |
| oanson | I stress the concept of double masquerade / double NAT, since that's the essence of this solution. | 10:14 |
| xiaohhui | OK, I think we can have further discussion in spec. | 10:15 |
| oanson | Sounds good | 10:15 |
| oanson | I think it's lunch time (or dinner time) | 10:16 |
| xiaohhui | yeah, | 10:16 |
| oanson | Enjoy your meal. I plan to enjoy mine :) | 10:17 |
| xiaohhui | bye~ | 10:17 |
| oanson | Bye | 10:17 |
| yuli_s | lunch time ! :) | 10:17 |
| rajivk | Tea Time :) | 10:17 |
| openstackgerrit | yuan wei proposed openstack/dragonflow: Reset related conntrack entries when the configuration about SG changed https://review.openstack.org/393710 | 10:23 |
| *** hujie has joined #openstack-dragonflow | 10:39 | |
| *** hujie has left #openstack-dragonflow | 10:39 | |
| *** gampel has quit IRC | 10:40 | |
| *** irenab has quit IRC | 10:43 | |
| openstackgerrit | Hong Hui Xiao proposed openstack/dragonflow: Spec to support check if chassis is alive https://review.openstack.org/385719 | 11:08 |
| *** irenab has joined #openstack-dragonflow | 11:36 | |
| *** dimak has quit IRC | 11:53 | |
| *** gsagie has joined #openstack-dragonflow | 11:58 | |
| *** gampel has joined #openstack-dragonflow | 11:59 | |
| *** gampel has quit IRC | 12:15 | |
| openstackgerrit | Hong Hui Xiao proposed openstack/dragonflow: Spec to support check if chassis is alive https://review.openstack.org/385719 | 12:16 |
| *** dimak has joined #openstack-dragonflow | 12:18 | |
| *** gampel has joined #openstack-dragonflow | 12:22 | |
| openstackgerrit | Hong Hui Xiao proposed openstack/dragonflow: Store local network id to nb db https://review.openstack.org/394017 | 12:30 |
| *** yuli_s has quit IRC | 12:48 | |
| irenab | oanson, hi | 13:38 |
| oanson | irenab, hi | 13:39 |
| irenab | started to put notes regarding DF api at this page https://etherpad.openstack.org/p/dragonflow_api | 13:39 |
| oanson | I see | 13:40 |
| oanson | Very cool | 13:40 |
| oanson | This is what we discussed this morning, right? | 13:42 |
| irenab | lets use it to collaborate on this topic | 13:42 |
| irenab | yes, just a summary of what we already discussed | 13:43 |
| oanson | Sounds good. | 13:43 |
| irenab | cool, thanks | 13:43 |
| oanson | Thank you! | 13:44 |
| *** oanson has quit IRC | 14:06 | |
| *** gsagie has quit IRC | 14:23 | |
| *** gampel has quit IRC | 14:45 | |
| *** gampel1 has joined #openstack-dragonflow | 14:45 | |
| openstackgerrit | Omer Anson proposed openstack/dragonflow: Support passing credentials via env in fullstack https://review.openstack.org/394483 | 15:10 |
| openstackgerrit | Omer Anson proposed openstack/dragonflow: Support passing credentials via env in fullstack https://review.openstack.org/394483 | 15:11 |
| *** irenab has quit IRC | 15:16 | |
| *** oanson has joined #openstack-dragonflow | 15:19 | |
| *** gampel1 has quit IRC | 15:19 | |
| *** gampel has joined #openstack-dragonflow | 15:29 | |
| saggi | gampel: How are you | 15:30 |
| *** serverascode has quit IRC | 15:41 | |
| openstackgerrit | Merged openstack/dragonflow: Nitpicks: Remove redundant pass https://review.openstack.org/393392 | 15:42 |
| openstackgerrit | Merged openstack/dragonflow: Nitpicks: Fix typo in RedisPortStatusNotifier https://review.openstack.org/393393 | 15:42 |
| openstackgerrit | Merged openstack/dragonflow: Nitpicks: Fix invalid attribute access https://review.openstack.org/393394 | 15:42 |
| openstackgerrit | Merged openstack/dragonflow: Fix unable to update subnet when subnet disables dhcp https://review.openstack.org/393328 | 15:43 |
| openstackgerrit | Dima Kuznetsov proposed openstack/dragonflow: [WIP] Service Function Chaining specification https://review.openstack.org/394498 | 15:46 |
| *** serverascode has joined #openstack-dragonflow | 15:48 | |
| *** zenoway has quit IRC | 16:02 | |
| *** zenoway has joined #openstack-dragonflow | 16:02 | |
| *** zenoway has quit IRC | 16:06 | |
| *** yuval has quit IRC | 16:08 | |
| *** dimak has quit IRC | 16:13 | |
| gampel | hey saggi are you still there | 16:18 |
| *** gampel has left #openstack-dragonflow | 16:20 | |
| openstackgerrit | Merged openstack/dragonflow: make mod flow uniform and delete duplicated args https://review.openstack.org/391731 | 16:28 |
| openstackgerrit | Merged openstack/dragonflow: Connect nb db in l3 plugin after neutron server fork https://review.openstack.org/393809 | 16:35 |
| openstackgerrit | Merged openstack/dragonflow: Add portstats table to df-db https://review.openstack.org/392671 | 16:36 |
| openstackgerrit | Merged openstack/dragonflow: Remove duplicating log setup https://review.openstack.org/393802 | 16:38 |
| openstackgerrit | Merged openstack/dragonflow: Updated from global requirements https://review.openstack.org/393637 | 16:38 |
| *** lihi has quit IRC | 16:48 | |
| openstackgerrit | Omer Anson proposed openstack/dragonflow: Support passing credentials via env in fullstack https://review.openstack.org/394483 | 16:59 |
| openstackgerrit | Steve Kipp proposed openstack/dragonflow: Add unit tests to df_local_controller https://review.openstack.org/390083 | 18:03 |
| *** oanson has quit IRC | 20:41 | |
| openstackgerrit | Dima Kuznetsov proposed openstack/dragonflow: [WIP] Service Function Chaining specification https://review.openstack.org/394498 | 20:53 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!