| *** yamamoto_ has joined #openstack-dragonflow | 02:09 | |
| *** _pino has joined #openstack-dragonflow | 03:25 | |
| _pino | oanson: dimak: Is there any OpenFlow/OVS register holding the tenant/project ID? If not, I'd like to set it for an experiment I'm doing. I'm thinking of just modifying | 03:31 |
|---|---|---|
| _pino | https://github.com/openstack/dragonflow/blob/master/dragonflow/controller/apps/l2.py#L328 | 03:32 |
| _pino | ... just adding another action to set a free register with the value of the tenant_id. | 03:33 |
| _pino | But I don't know how to get the tenant/project id from the lport - is it just the device_owner field? | 03:33 |
| _pino | Or is the tenant_id saved in lport.topic? | 03:35 |
| oanson | _pino, this isn't done. | 03:37 |
| oanson | I suggest you use the classifier app, which also sets reg6 | 03:37 |
| oanson | The tenant ID is usually in topic. | 03:37 |
| oanson | For VM/container lports, I think it's always in 'topic' | 03:37 |
| oanson | _pino, see e.g. https://github.com/openstack/dragonflow/blob/master/dragonflow/controller/apps/classifier.py?utf8=%E2%9C%93#L85 | 03:38 |
| oanson | But note that topic is a UUID (I think that's 128 bits). Registers only have 32 bits. So I'd recommend a mapping. | 03:41 |
| _pino | oanson: thanks! | 03:48 |
| oanson | _pino, no worries. May I ask what you need this for? | 03:48 |
| _pino | One more thing... reg6 and reg7 are set to lport.unique_id | 03:48 |
| _pino | Who assigns those unique IDs? | 03:48 |
| oanson | Unique IDs are assigned in the Neutron Server by our mechanism driver. We use the distributed database to generate the unique keys in an atomic manner so there'll be no collisions | 03:49 |
| _pino | Are those OF/OVS port numbers (unique, but start at 1 and are on the order of number of VMs on the host). | 03:49 |
| _pino | Ah, ok, so they're unique across the entire deployment. I was just thinking that reg6/reg7 might have enough bits to encode both the port and the tenant ID. | 03:50 |
| oanson | Yes. In essence, they already do :) | 03:50 |
| oanson | But the encoding isn't very useful. | 03:50 |
| _pino | Anyway, that's over-optimizing for my current purposes - I'll ask you about it again later. | 03:51 |
| oanson | Sure. | 03:51 |
| *** _pino has quit IRC | 03:53 | |
| openstackgerrit | Omer Anson proposed openstack/dragonflow master: use qos DriverBase from neutron-lib https://review.openstack.org/501562 | 04:05 |
| oanson | dimak, lihi ^^^^^ This might be a gate fix due to https://review.openstack.org/#/c/487194 | 04:06 |
| *** yamamot__ has joined #openstack-dragonflow | 04:20 | |
| *** yamamoto_ has quit IRC | 04:21 | |
| *** dims has quit IRC | 04:40 | |
| *** lihi has quit IRC | 05:06 | |
| *** yamamot__ has quit IRC | 05:07 | |
| *** lihi has joined #openstack-dragonflow | 05:07 | |
| *** yamamoto_ has joined #openstack-dragonflow | 05:07 | |
| *** leyal has quit IRC | 05:08 | |
| *** leyal has joined #openstack-dragonflow | 05:08 | |
| *** oanson has quit IRC | 05:58 | |
| *** oanson has joined #openstack-dragonflow | 05:59 | |
| openstackgerrit | Omer Anson proposed openstack/dragonflow master: use neutron notifier correctly https://review.openstack.org/499166 | 06:11 |
| oanson | lihi, dimak, I'd be happy if you could also look at https://review.openstack.org/#/c/501240 ? | 06:57 |
| oanson | Thanks! | 06:57 |
| *** shachar has joined #openstack-dragonflow | 07:16 | |
| *** zenoway has joined #openstack-dragonflow | 07:22 | |
| openstackgerrit | Eyal Leshem proposed openstack/dragonflow master: Compere IP to IP https://review.openstack.org/501233 | 07:27 |
| openstackgerrit | Eyal Leshem proposed openstack/dragonflow master: Compere IP to IP https://review.openstack.org/501233 | 07:31 |
| openstackgerrit | Eyal Leshem proposed openstack/dragonflow master: use neutron notifier correctly https://review.openstack.org/499166 | 07:43 |
| openstackgerrit | yanghuichan proposed openstack/dragonflow master: Fix wrong links in dragonflow https://review.openstack.org/501619 | 08:09 |
| *** omnipresent has joined #openstack-dragonflow | 08:09 | |
| *** zenoway has quit IRC | 08:12 | |
| openstackgerrit | Merged openstack/dragonflow master: use qos DriverBase from neutron-lib https://review.openstack.org/501562 | 08:40 |
| *** omnipresent has quit IRC | 08:43 | |
| *** zenoway has joined #openstack-dragonflow | 08:46 | |
| openstackgerrit | Merged openstack/dragonflow master: Update log info https://review.openstack.org/501048 | 09:05 |
| openstackgerrit | Eyal Leshem proposed openstack/dragonflow master: Compere IP to IP https://review.openstack.org/501233 | 09:10 |
| openstackgerrit | Eyal Leshem proposed openstack/dragonflow master: use neutron notifier correctly https://review.openstack.org/499166 | 09:14 |
| *** yamamoto_ has quit IRC | 09:28 | |
| *** yamamoto has joined #openstack-dragonflow | 09:46 | |
| openstackgerrit | Eyal Leshem proposed openstack/dragonflow master: DHCP port per lswitch spec https://review.openstack.org/498464 | 10:03 |
| openstackgerrit | Eyal Leshem proposed openstack/dragonflow master: Fix wrong links in dragonflow https://review.openstack.org/501619 | 10:04 |
| openstackgerrit | Eyal Leshem proposed openstack/dragonflow master: DHCP port per lswitch spec https://review.openstack.org/498464 | 10:08 |
| *** zenoway has quit IRC | 10:08 | |
| *** zenoway has joined #openstack-dragonflow | 10:09 | |
| *** zenoway has quit IRC | 10:13 | |
| *** yamamoto has quit IRC | 10:14 | |
| *** oanson has quit IRC | 10:16 | |
| *** oanson has joined #openstack-dragonflow | 10:18 | |
| *** zenoway has joined #openstack-dragonflow | 10:18 | |
| *** yamamoto has joined #openstack-dragonflow | 10:19 | |
| *** zenoway has quit IRC | 10:23 | |
| *** zenoway has joined #openstack-dragonflow | 10:23 | |
| *** oanson_ has joined #openstack-dragonflow | 10:32 | |
| *** oanson_ has quit IRC | 10:38 | |
| openstackgerrit | Eyal Leshem proposed openstack/dragonflow master: DHCP port per lswitch spec https://review.openstack.org/498464 | 11:48 |
| openstackgerrit | Merged openstack/dragonflow master: Clear db_store by calling clear method https://review.openstack.org/501240 | 11:57 |
| *** dims has joined #openstack-dragonflow | 12:07 | |
| openstackgerrit | Merged openstack/dragonflow master: Add an in-memory NB database driver https://review.openstack.org/494248 | 12:24 |
| *** mlavalle has joined #openstack-dragonflow | 14:05 | |
| *** irenab has quit IRC | 15:12 | |
| *** irenab has joined #openstack-dragonflow | 15:12 | |
| oanson | mlavalle, hi, are you in? | 15:35 |
| mlavalle | yeah | 15:35 |
| oanson | Can I trouble you with an L3 flavour question? | 15:35 |
| mlavalle | sup? | 15:35 |
| mlavalle | I'll try to help | 15:35 |
| oanson | No complaints here :). What's with you? | 15:35 |
| oanson | I read the docs. I'm not sure I understand how the L3 implementation connects with the rest of the backend. | 15:36 |
| oanson | How does the L3 flavour implementation get the packets? How does it send them out? | 15:36 |
| mlavalle | yeah the document is not very good | 15:37 |
| oanson | Once Dragonflow L3 flavour is implemented, we can use it as a running example and a tutorial | 15:38 |
| mlavalle | let's move this conversation to the Neutron channel. running a meeting right now. But kevinbento can helpo you there | 15:38 |
| oanson | Sure | 15:38 |
| *** zenoway has quit IRC | 15:53 | |
| mlavalle | oanson: if kevin doesn't answer in a reasonable time for you (it is the middle of the night, isn't it) I'll follow up with you on Monday during th IRC meeting | 16:12 |
| oanson | mlavalle, sure. Thanks. It's past midnight here, and my plane will be leaving soon :) | 16:12 |
| oanson | (Mostly the plane time is the issue) | 16:13 |
| mlavalle | I'll follow up with you | 16:13 |
| oanson | Thanks! | 16:16 |
| *** dims has quit IRC | 18:34 | |
| *** dims has joined #openstack-dragonflow | 18:36 | |
| *** dims has quit IRC | 18:41 | |
| *** dims has joined #openstack-dragonflow | 18:45 | |
| *** mlavalle has quit IRC | 23:49 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!