| openstackgerrit | Merged openstack/dragonflow: Re-download flows when reconnect to ryu controller https://review.openstack.org/378193 | 04:26 |
|---|---|---|
| openstackgerrit | Merged openstack/dragonflow: uniform method for allowed address pair https://review.openstack.org/415705 | 05:43 |
| *** yuli_s has quit IRC | 07:33 | |
| *** yuval has joined #openstack-dragonflow | 08:02 | |
| *** yuli_s has joined #openstack-dragonflow | 08:54 | |
| yuli_s | hello | 08:54 |
| oanson | dimak, ping | 08:56 |
| dimak | oanson, hey | 08:56 |
| oanson | about the nb refactor, there is another issue we may need to address: locking | 08:56 |
| dimak | what parts | 08:56 |
| oanson | Technically can wait for later, but we should make sure we know it exists | 08:56 |
| oanson | Currently done in the Neutron mechanism driver. | 08:56 |
| oanson | It uses the SQL DB to lock the records being modified | 08:57 |
| dimak | I'm not that familiar | 08:57 |
| dimak | Oh the transaction - retry loop? | 08:57 |
| oanson | If you look in the mechanism driver, you will see the locking decorators | 08:57 |
| oanson | dimak, no, I think its completely dragonflow code. | 08:57 |
| oanson | Did you see hujie's modifications last week? | 08:58 |
| oanson | That's the location | 08:58 |
| dimak | Yes | 08:58 |
| oanson | We use a select-for-update to lock the record. Modify it. Then release. We have a retry mechanism there, yes | 08:58 |
| dimak | umm | 08:59 |
| oanson | I just wanted to mention it. It can be a work item: Move locking from mech_driver to model API | 08:59 |
| irenab | oanson, but its the usage model. Do yolu want to handle locking internally? | 09:00 |
| oanson | Everyone, Dragonflow weekly starting in a few seconds in #openstack-meeting-4 | 09:00 |
| dimak | We'll still need mysql | 09:00 |
| oanson | irenab, possibly, yes | 09:00 |
| dimak | no? | 09:00 |
| oanson | If we provide a REST api, or a CLI, the locking should be transparent. No? | 09:00 |
| irenab | I thin k the question if the locking is a separate layer or part of the model | 09:01 |
| oanson | irenab, let's continue it during/after the weekly | 09:01 |
| irenab | oanson, ok | 09:04 |
| *** yuli_s has quit IRC | 11:33 | |
| *** yuli_s has joined #openstack-dragonflow | 11:33 | |
| *** yamamoto has joined #openstack-dragonflow | 12:32 | |
| *** yamamoto has quit IRC | 13:18 | |
| openstackgerrit | WangJian proposed openstack/dragonflow: Support Firewall as a service https://review.openstack.org/415373 | 13:32 |
| *** saggi has quit IRC | 13:52 | |
| *** saggi has joined #openstack-dragonflow | 13:52 | |
| *** yamamoto has joined #openstack-dragonflow | 14:19 | |
| *** yamamoto has quit IRC | 14:29 | |
| openstackgerrit | hujie proposed openstack/dragonflow: separate control plane and data plane ip address https://review.openstack.org/415116 | 14:59 |
| *** yuval has quit IRC | 16:05 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/dragonflow: Updated from global requirements https://review.openstack.org/415958 | 16:08 |
| *** bzhao_ has joined #openstack-dragonflow | 22:12 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!