| openstackgerrit | Prudhvi Rao Shedimbi proposed openstack/watcher: Implemented wacther decision engine config module https://review.openstack.org/401006 | 00:53 |
|---|---|---|
| *** diga has quit IRC | 02:21 | |
| *** thorst has joined #openstack-watcher | 02:32 | |
| *** thorst has joined #openstack-watcher | 02:33 | |
| openstackgerrit | Merged openstack/puppet-watcher: Remove rabbitmq connection parameters https://review.openstack.org/395244 | 02:33 |
| *** thorst has quit IRC | 02:40 | |
| *** thorst has joined #openstack-watcher | 02:41 | |
| *** thorst has quit IRC | 02:45 | |
| *** thorst has joined #openstack-watcher | 03:09 | |
| openstackgerrit | Rong Han proposed openstack/watcher: Remove redundan lines. https://review.openstack.org/401050 | 03:14 |
| openstackgerrit | Rong Han proposed openstack/watcher: Solve some spelling mistakes. https://review.openstack.org/399402 | 03:23 |
| *** jwcroppe has joined #openstack-watcher | 03:23 | |
| *** jwcroppe has quit IRC | 03:24 | |
| *** jwcroppe has joined #openstack-watcher | 03:24 | |
| *** jwcroppe has quit IRC | 03:24 | |
| *** jwcroppe has joined #openstack-watcher | 03:25 | |
| *** thorst has joined #openstack-watcher | 03:36 | |
| *** thorst has quit IRC | 03:37 | |
| openstackgerrit | bailin.zhang proposed openstack/watcher-specs: Fix a spelling error https://review.openstack.org/401072 | 06:09 |
| *** michaelgugino has quit IRC | 06:23 | |
| *** Yumeng has joined #openstack-watcher | 06:31 | |
| *** Yumeng has quit IRC | 07:04 | |
| openstackgerrit | Rong Han proposed openstack/watcher-specs: Support Description For Dynamic Action https://review.openstack.org/401111 | 07:50 |
| *** diga has joined #openstack-watcher | 08:17 | |
| *** dtardivel has joined #openstack-watcher | 08:41 | |
| *** alexchadin has joined #openstack-watcher | 08:51 | |
| openstackgerrit | Alexander Chadin proposed openstack/watcher: Add periods input parameter https://review.openstack.org/399539 | 08:55 |
| openstackgerrit | Merged openstack/watcher: Remove redundan lines. https://review.openstack.org/401050 | 08:58 |
| openstackgerrit | Merged openstack/watcher: Solve some spelling mistakes. https://review.openstack.org/399402 | 08:58 |
| acabot | hi guys, weekly meeting will start in 2 min on #openstack-meeting-4 | 08:59 |
| openstackgerrit | Merged openstack/watcher: Fix inconsistent descriptions in docstring in action_plan.py https://review.openstack.org/399846 | 08:59 |
| openstackgerrit | Rong Han proposed openstack/watcher-specs: Support Description For Dynamic Action https://review.openstack.org/401111 | 09:00 |
| *** diga has quit IRC | 09:02 | |
| alexchadin | dtardivel: hi! | 09:04 |
| dtardivel | hi alexchadin | 09:05 |
| dtardivel | we are on meeting :) | 09:05 |
| dtardivel | can you join us ? | 09:05 |
| alexchadin | dtardivel: already | 09:06 |
| dtardivel | yep I see it :) | 09:07 |
| *** diga has joined #openstack-watcher | 09:12 | |
| *** befreax has joined #openstack-watcher | 09:14 | |
| *** befreax is now known as tmetsch | 09:14 | |
| openstackgerrit | Merged openstack/watcher: Fix the wrong ref for 'Compute node' https://review.openstack.org/398673 | 09:41 |
| *** diga_ has joined #openstack-watcher | 09:41 | |
| openstackgerrit | Merged openstack/watcher: Fix some typos in action.py & action_plan.py & audit.py https://review.openstack.org/398938 | 09:41 |
| *** diga has quit IRC | 09:42 | |
| openstackgerrit | Merged openstack/python-watcherclient: Updated from global requirements https://review.openstack.org/398620 | 09:46 |
| alexchadin | dtardivel: can we discuss planner improvements? | 09:47 |
| dtardivel | yep | 09:50 |
| *** tmetsch is now known as befreax | 09:50 | |
| alexchadin | As you said here, disabling compute node is not linked action to the migration since we still can migrate instances after node is disabled https://review.openstack.org/#/c/385871/2/specs/ocata/approved/planner-storage-action-plan.rst@40 | 09:52 |
| alexchadin | but I agree, that we need mechanism to support 2 or more parents | 09:53 |
| alexchadin | in the future, since we don't have such type of action that will have some requirements | 09:53 |
| alexchadin | we can store parents as list in database (json or jsonb in PostgreSQL, for example) | 09:54 |
| dtardivel | yes I think so | 09:56 |
| *** alexchadin has quit IRC | 09:57 | |
| *** alexchad_ has joined #openstack-watcher | 09:57 | |
| alexchad_ | dtardivel: sorry for disconnecting | 09:57 |
| alexchad_ | dtardivel: "yes I think so" - is it your last message? | 09:58 |
| dtardivel | yep | 09:58 |
| alexchad_ | dtardivel: great:) | 09:59 |
| dtardivel | can you add in your spec the new fields of actions ? | 10:02 |
| dtardivel | based from https://github.com/openstack/watcher/blob/master/watcher/objects/action.py#L44 | 10:03 |
| dtardivel | I did not understand why you said both BP should be merged together | 10:04 |
| alexchad_ | dtardivel: you mean to remove first_action_id? | 10:04 |
| alexchad_ | dtardivel: how will applier work without changes if we remove next fields also? | 10:05 |
| alexchad_ | field* | 10:05 |
| dtardivel | If in the first one, you update the Action data model, by introducing a new 'parents' field and you rewrite the planner algorithm to take into account this new way to store an action in Db, we should have no impact in the applier (the action plan built by the planner should be similar as before) | 10:08 |
| dtardivel | in other words, in db we will have new model, by but from Applier point of view, we will have same data structure with 'next' pointer. This is the proposal from vincentfrancoise to be able to split this BP in two parts. | 10:10 |
| dtardivel | so the new planner job will be to translate 'parents' link to 'next' links | 10:12 |
| dtardivel | i'm not familiar with this part of code. | 10:12 |
| dtardivel | Does it make sense from your point of view ? | 10:13 |
| alexchad_ | dtardivel: I'm exploring this part of code | 10:13 |
| alexchad_ | dtardivel: give me a sec:) | 10:13 |
| dtardivel | you have time for reflexion | 10:14 |
| dtardivel | It's just a input | 10:14 |
| dtardivel | The first part of the work will concern only the data model and the planner. The second part will concern the applier and the graph modelisation (and you can see from my previous comments that there will be lot of use cases to cover) | 10:17 |
| dtardivel | take time to think about vincentfrancoise proposal :) | 10:17 |
| alexchad_ | dtardivel: well, we have almost done our work with the first part | 10:17 |
| alexchad_ | dtardivel: ok, I agree | 10:25 |
| dtardivel | If you agree with our proposal, could you rewrite your pseudo code and the list of work items as well | 10:26 |
| alexchad_ | https://github.com/openstack/watcher/blob/master/watcher/applier/default.py#L60 | 10:26 |
| alexchad_ | here we get list of actions | 10:26 |
| dtardivel | and also define exactly the scope of this spec :) | 10:26 |
| alexchad_ | and we don't use 'next' field in applier | 10:26 |
| dtardivel | oh oh I see | 10:28 |
| dtardivel | in fact 'next' is used only by planner to built of ordered list of actions, named action plan. | 10:30 |
| dtardivel | So if we remove 'next' and add 'parents' fielf, the action plan new algo should take into account the new field 'parents' to built the ordered action list. in our first implementation, 'parents' will have only a unique action uuid | 10:32 |
| dtardivel | are you agree ? | 10:33 |
| dtardivel | s/to built of/to build an | 10:33 |
| *** pkoniszewski has quit IRC | 10:35 | |
| *** alexchad_ has quit IRC | 10:37 | |
| *** alexchadin has joined #openstack-watcher | 10:37 | |
| dtardivel | hummm I read the code is 'next' field realy useful for now ? | 10:39 |
| *** alexchadin has quit IRC | 10:42 | |
| *** pkoniszewski has joined #openstack-watcher | 10:44 | |
| *** danpawlik has joined #openstack-watcher | 10:54 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/watcher: Updated from global requirements https://review.openstack.org/401152 | 11:05 |
| *** diga_ has quit IRC | 11:48 | |
| *** exploreshaifali has joined #openstack-watcher | 11:50 | |
| *** openstackgerrit has quit IRC | 12:03 | |
| *** openstackgerrit has joined #openstack-watcher | 12:03 | |
| *** shaifali has joined #openstack-watcher | 12:12 | |
| *** exploreshaifali has quit IRC | 12:12 | |
| *** alexchadin has joined #openstack-watcher | 12:41 | |
| alexchadin | dtardivel: ping | 12:42 |
| dtardivel | alexchadin: pong | 13:14 |
| alexchadin | seems that 'next' field is used only in API | 13:14 |
| alexchadin | right? | 13:14 |
| dtardivel | yes. but I will need to clarify with jed56 | 13:14 |
| dtardivel | I think the applier is not well implemented. It is confused for me now. For me the goal of the planner will be to build a graph of actions (the simple graph will be a linked chain of actions for now). | 13:17 |
| dtardivel | The goal of the applier should be to to convert the action graph to something readable by the workflow engine used, i.e. taskflow for our default implementation . | 13:18 |
| dtardivel | right ? | 13:18 |
| alexchadin | right | 13:25 |
| alexchadin | so planner only creates linked list now | 13:25 |
| alexchadin | and save it to DB | 13:26 |
| dtardivel | I'm discussing with jed56. He implemented this part a long time ago. | 13:32 |
| *** lamt has joined #openstack-watcher | 13:39 | |
| dtardivel | alexchadin: jed56 confirm to me that this is a bug on applier side. it shoudl take into account 'next' field to find the next action. | 13:43 |
| alexchadin | wow | 13:43 |
| dtardivel | So you're right, applier should be impected as well in the first bp | 13:43 |
| alexchadin | well | 13:44 |
| dtardivel | so forget what I wrote later | 13:44 |
| alexchadin | So we have to implement storage improvements and changes to applier | 13:45 |
| dtardivel | planner will build a list of action (the action plan) chained by the 'parents' field. And applier should be able to handle 'parents' field to create task for engine workflow (taskflow). And we will limit the first bp to simple chained list used for now. right ? | 13:46 |
| alexchadin | hm | 13:47 |
| alexchadin | ok | 13:48 |
| *** alexchadin has quit IRC | 13:50 | |
| *** shaifali_ has joined #openstack-watcher | 13:53 | |
| *** shaifali has quit IRC | 13:57 | |
| *** thorst has joined #openstack-watcher | 14:08 | |
| *** thorst has quit IRC | 14:19 | |
| *** shaifali has joined #openstack-watcher | 14:22 | |
| *** thorst has joined #openstack-watcher | 14:22 | |
| *** shaifali_ has quit IRC | 14:25 | |
| *** thorst has quit IRC | 14:31 | |
| *** shaifali_ has joined #openstack-watcher | 14:35 | |
| *** shaifali has quit IRC | 14:37 | |
| *** shaifali_ has quit IRC | 15:00 | |
| *** thorst has joined #openstack-watcher | 15:05 | |
| sballe_ | can I get some +2 on https://review.openstack.org/#/c/397411/10? dtardivel | 15:11 |
| sballe_ | acabot: ^^ | 15:11 |
| sballe_ | thx tp everybody who reviewed | 15:12 |
| *** thorst has quit IRC | 15:12 | |
| *** thorst has joined #openstack-watcher | 15:16 | |
| *** thorst has quit IRC | 15:20 | |
| acabot | sballe_ : I plan to review it tomorrow | 15:29 |
| sballe_ | thx | 15:30 |
| sballe_ | 4am is just too early for me to attend the IRC meeting. | 15:30 |
| sballe_ | I will read the minutes | 15:30 |
| sballe_ | acabot: | 15:30 |
| acabot | no pb | 15:31 |
| sballe_ | acabot: I should have the remaing two docs done really quickly now. I just didn;t want to have 3 docs in flight which would end up haveing similar comments across them | 15:40 |
| acabot | sballe_ : ok great | 15:44 |
| sballe_ | All: pshedimb is looking for for reviews on https://review.openstack.org/#/c/400367/3 | 15:45 |
| acabot | dtardivel will review https://review.openstack.org/#/q/topic:bp/centralise-config-opts | 15:47 |
| sballe_ | thx. pshedimb decided to commit smaller peices instead of all the code in one commit. Does that works for you or do you want it all in one commit? | 15:49 |
| acabot | no no please split the code | 15:50 |
| sballe_ | ok good. that what he did | 15:50 |
| *** thorst has joined #openstack-watcher | 16:38 | |
| *** thorst has quit IRC | 16:39 | |
| *** thorst has joined #openstack-watcher | 17:06 | |
| *** thorst_ has joined #openstack-watcher | 17:07 | |
| *** thorst has quit IRC | 17:11 | |
| *** thorst_ has quit IRC | 17:11 | |
| openstackgerrit | Prudhvi Rao Shedimbi proposed openstack/watcher: Implemented api config module https://review.openstack.org/400367 | 17:16 |
| openstackgerrit | Prudhvi Rao Shedimbi proposed openstack/watcher: Implemented applier config module https://review.openstack.org/400969 | 17:17 |
| openstackgerrit | Prudhvi Rao Shedimbi proposed openstack/watcher: Implemented wacther decision engine config module https://review.openstack.org/401006 | 17:17 |
| pshedimb | thanks sballe_ acabot | 17:17 |
| pshedimb | pushed in a few config options yest and will push in a few more today. | 17:18 |
| openstackgerrit | Prudhvi Rao Shedimbi proposed openstack/watcher: Implemented api config module https://review.openstack.org/400367 | 17:24 |
| openstackgerrit | Prudhvi Rao Shedimbi proposed openstack/watcher: Implemented applier config module https://review.openstack.org/400969 | 17:25 |
| openstackgerrit | Prudhvi Rao Shedimbi proposed openstack/watcher: Implemented wacther decision engine config module https://review.openstack.org/401006 | 17:26 |
| *** befreax has quit IRC | 18:53 | |
| *** dtardivel has quit IRC | 19:24 | |
| *** wootehfoot has joined #openstack-watcher | 21:15 | |
| *** thorst_ has joined #openstack-watcher | 21:47 | |
| *** thorst_ has quit IRC | 22:20 | |
| *** wootehfoot has quit IRC | 23:26 | |
| *** lamt has quit IRC | 23:43 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!