| *** gampel has joined #openstack-smaug | 00:34 | |
| *** gampel has quit IRC | 00:54 | |
| *** c00281451 has quit IRC | 01:27 | |
| *** wangfeng_yellow has quit IRC | 02:52 | |
| yinwei | saggi, atom is base class of Retry and BaseTask, not the base class of flow. | 03:58 |
|---|---|---|
| yinwei | actually, I just want to confirm the interface of Provider to build task flow, so that ProtectionManager could get that task flow, compose with other tasks to get a top level flow, and start to execute | 04:00 |
| yinwei | first such an interface like provider.build_graph_flow(backup) is missing in current PluggableProvider design | 04:01 |
| yinwei | second, I'm thinking the return of that interface could be a graph flow. The components of this graph flow could be tasks/sub graph-flows built from resource graph walking. | 04:03 |
| yinwei | Those components are added to the top graph flow, without links, and thus would be executed on parallel. | 04:04 |
| yinwei | Each component (task/sub graph-flow) corresponds to one connected sub-graph in the whole resource graph. Waling through one connected sub-graph, we will get a task or a sub-graphflow. The links inside sub-grapflow are built on leaving node, pointing to its dependent tasks. | 04:10 |
| yinwei | Provider just return the top level graph flow is ok, ProtectionManager who will executed task flow just take the flow returned by Provider as a whole and compose any additional tasks before or after it. | 04:12 |
| openstackgerrit | ting wang proposed openstack/smaug: Clean flake8 ignore https://review.openstack.org/267460 | 10:50 |
| *** yinweimac has joined #openstack-smaug | 12:48 | |
| *** yinweimac has quit IRC | 12:53 | |
| *** WANG_Feng has quit IRC | 14:02 | |
| *** wanghao has quit IRC | 14:17 | |
| *** gampel has joined #openstack-smaug | 14:26 | |
| *** WANG_Feng has joined #openstack-smaug | 15:57 | |
| *** gampel has quit IRC | 20:14 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!