| *** Yumeng has quit IRC | 00:09 | |
| *** hvprash has quit IRC | 00:13 | |
| *** thorst has joined #openstack-watcher | 01:04 | |
| *** thorst has quit IRC | 01:22 | |
| *** Kevin_Zheng has joined #openstack-watcher | 01:33 | |
| *** thorst has joined #openstack-watcher | 01:58 | |
| *** Guest50630 has quit IRC | 02:03 | |
| *** thorst has quit IRC | 02:03 | |
| *** zigo has joined #openstack-watcher | 02:05 | |
| *** zigo is now known as Guest59427 | 02:06 | |
| *** yuanying has quit IRC | 02:48 | |
| *** yuanying has joined #openstack-watcher | 02:58 | |
| *** yuanying has quit IRC | 03:02 | |
| *** odyssey4me has quit IRC | 03:14 | |
| *** odyssey4me has joined #openstack-watcher | 03:15 | |
| *** yuanying has joined #openstack-watcher | 03:30 | |
| *** yuanying has quit IRC | 03:33 | |
| *** licanwei has joined #openstack-watcher | 03:35 | |
| *** yuanying has joined #openstack-watcher | 03:36 | |
| *** yuanying has quit IRC | 03:45 | |
| *** yuanying has joined #openstack-watcher | 03:49 | |
| *** hvprash has joined #openstack-watcher | 04:02 | |
| *** gabor_antal_ has joined #openstack-watcher | 05:35 | |
| *** gabor_antal has quit IRC | 05:37 | |
| *** dtardivel has joined #openstack-watcher | 07:07 | |
| *** vincentfrancoise has joined #openstack-watcher | 07:23 | |
| *** Guest59427 is now known as zigo | 07:33 | |
| *** acabot has joined #openstack-watcher | 07:37 | |
| *** alexchadin has joined #openstack-watcher | 07:55 | |
| *** Yumeng has joined #openstack-watcher | 08:06 | |
| *** alexchad_ has joined #openstack-watcher | 08:15 | |
| *** alexchadin has quit IRC | 08:18 | |
| *** befreax has joined #openstack-watcher | 08:48 | |
| *** befreax is now known as tmetsch | 08:49 | |
| *** alexchad_ has quit IRC | 09:01 | |
| *** alexchadin has joined #openstack-watcher | 09:02 | |
| acabot | weekly meeting starts on #openstack-meeting-4 | 09:03 |
|---|---|---|
| *** alexchad_ has joined #openstack-watcher | 09:05 | |
| *** alexchad_ is now known as alexchadin_ | 09:05 | |
| vincentfrancoise | alexchadin: hi | 09:06 |
| *** alexchadin has quit IRC | 09:06 | |
| vincentfrancoise | alexchadin: just as an FYI, I'll try to merge your service BP before EOD since it already has a +2 | 09:07 |
| alexchadin_ | vincentfrancoise: nice to hear:) | 09:10 |
| alexchadin_ | jed56: ping | 09:37 |
| *** alexchadin_ is now known as alexchadin | 09:46 | |
| jed56 | alexchadin: hi | 10:05 |
| alexchadin | jed56: how are you?:) | 10:05 |
| jed56 | very well and you ? | 10:07 |
| alexchadin | I'm fine | 10:08 |
| alexchadin | I think that we need to save action weights, maybe rename it to the priorities | 10:09 |
| jed56 | I have to go sorry | 10:09 |
| jed56 | Where do you want to store the priorities ? | 10:10 |
| alexchadin | we can store it in the action table | 10:10 |
| *** alexchadin has quit IRC | 10:23 | |
| *** alexchadin has joined #openstack-watcher | 11:38 | |
| *** thorst has joined #openstack-watcher | 11:47 | |
| vincentfrancoise | hi alexchadin | 11:55 |
| alexchadin | hi vincentfrancoise | 11:55 |
| vincentfrancoise | jed56 is busy but kind of followed the debate | 11:56 |
| vincentfrancoise | alexchadin: so I'll try to answer your questions ;) | 11:56 |
| alexchadin | ok:) | 11:56 |
| alexchadin | I think, that we don't need next field (action table) anymore | 11:57 |
| vincentfrancoise | I was wondering: why you were willing to store the weights? | 11:57 |
| alexchadin | what if some strategy would want to set to one of migration action weight 2? | 11:58 |
| vincentfrancoise | so your problem is more about being able to select a specific planner for a specific strategy right? | 11:59 |
| vincentfrancoise | and pass some parameters down to it | 11:59 |
| alexchadin | take a look | 12:00 |
| alexchadin | http://paste.openstack.org/show/585447/ | 12:00 |
| alexchadin | this is part of proposed changes | 12:00 |
| vincentfrancoise | do not forget that the weight-based planner is only one of the possible implementations for a planner | 12:02 |
| vincentfrancoise | so the problem I see with the priority is that it is too specific | 12:03 |
| alexchadin | too specific? can you describe what you mean? | 12:04 |
| vincentfrancoise | so first off, I would argue that planner parameters should be put into some parameters sub structure like we do between the audit and the strategy | 12:04 |
| vincentfrancoise | by specific I mean that you already assume that we will be using the default planner even though it might not be the case | 12:05 |
| *** tmetsch has quit IRC | 12:06 | |
| alexchadin | ok, we can leave weights to load them from conf | 12:07 |
| vincentfrancoise | I agree, it's easier for now | 12:07 |
| alexchadin | then we still need map between action types and weight | 12:07 |
| vincentfrancoise | and it's not really necessary for you to complete this blueprint | 12:08 |
| alexchadin | is it ok? | 12:08 |
| vincentfrancoise | that mapping already exist I think | 12:08 |
| alexchadin | yes, it is exist:) | 12:08 |
| vincentfrancoise | so yeah it should stay | 12:09 |
| alexchadin | I propose to name it priorities | 12:09 |
| alexchadin | it makes more sense | 12:09 |
| vincentfrancoise | the parameters thing is probably a different blueprint that we will probably want in the future | 12:09 |
| vincentfrancoise | it's already renamed as priority in the code so yeah | 12:10 |
| vincentfrancoise | see https://github.com/openstack/watcher/blob/master/watcher/decision_engine/planner/default.py#L76 | 12:10 |
| alexchadin | Have you seen changes in workflow? | 12:11 |
| alexchadin | I mean, the diagram | 12:11 |
| vincentfrancoise | yeah but let me re-read it so I make sure I understand it | 12:11 |
| openstackgerrit | Merged openstack/watcher: Stop adding ServiceAvailable group option https://review.openstack.org/381024 | 12:12 |
| openstackgerrit | Merged openstack/watcher: Updated from global requirements https://review.openstack.org/383149 | 12:13 |
| vincentfrancoise | just to make sure I understand the "node" field you want to introduce: what defines the node name you will put in it? | 12:14 |
| vincentfrancoise | I mean, what does it refer to? | 12:14 |
| alexchadin | It may comes from source_node for migration type or resource_id for change_node_state | 12:16 |
| alexchadin | This is the node the action is related to | 12:16 |
| vincentfrancoise | yeah but where do you create the node you refer to and how does it get its name? | 12:17 |
| alexchadin | We add value to the node column (action table) at the planner layer | 12:19 |
| alexchadin | this is just compute node | 12:20 |
| vincentfrancoise | aaaah | 12:20 |
| vincentfrancoise | so it's not a graph node | 12:20 |
| vincentfrancoise | it's a compute node | 12:20 |
| alexchadin | yeap:) | 12:21 |
| vincentfrancoise | ok so that was not clear at all for me | 12:21 |
| alexchadin | seems need to rename Node to ComputeNode | 12:21 |
| vincentfrancoise | yeah but my problem will be the following then? | 12:22 |
| vincentfrancoise | what do we do in the future we we deal with, say, a storage model or a network model? | 12:22 |
| vincentfrancoise | we cannot hardcode a compute_node field in the action table | 12:22 |
| alexchadin | I meant to change Node in the Graph the ComputeNode | 12:23 |
| alexchadin | to the* | 12:24 |
| alexchadin | BTW, any actions we do related to the resources, that are placed on one of the nodes, compute or network or storage | 12:25 |
| vincentfrancoise | ok, let me try to explain how I understand it to see if that's what you meant | 12:26 |
| vincentfrancoise | The action table will not contain a "next" field anymore | 12:26 |
| vincentfrancoise | instead we add a "node" field | 12:26 |
| vincentfrancoise | which will reference the parent action (via its resource_id or its UUID/ID ?) | 12:27 |
| alexchadin | It will just contain Node Name | 12:27 |
| alexchadin | for example, compute1 | 12:28 |
| vincentfrancoise | IMHO, it's better to reference it via a nullable foreign key on another action. | 12:28 |
| alexchadin | hm | 12:30 |
| vincentfrancoise | so the Main Flow's top-level action will have no "node" set | 12:30 |
| vincentfrancoise | but all the subsequent actions will have something in it | 12:30 |
| vincentfrancoise | what do you think? | 12:31 |
| alexchadin | you want to reference actions to each other? | 12:32 |
| vincentfrancoise | yes | 12:33 |
| alexchadin | then where we will store Node Name?:) | 12:33 |
| alexchadin | hm | 12:36 |
| alexchadin | let me think | 12:36 |
| vincentfrancoise | let me think about it too | 12:37 |
| dtardivel | alexchadin: While talking with vincentfrancoise, could you have a quick look on https://review.openstack.org/#/c/382803/ and set +2/+1 if it's ok for you. Thx ;) | 12:38 |
| alexchadin | dtardivel: oh, great commit! | 12:38 |
| dtardivel | alexchadin: validated on my devstack !!! | 12:39 |
| alexchadin | dtardivel: Need to create sticker: Trust me, it works on devstack | 12:39 |
| vincentfrancoise | alexchadin: hahaha | 12:39 |
| *** alexchadin has quit IRC | 12:42 | |
| *** alexchadin has joined #openstack-watcher | 12:42 | |
| openstackgerrit | Merged openstack/watcher: Added composite unique name constraints https://review.openstack.org/382803 | 12:45 |
| openstackgerrit | Merged openstack/python-watcherclient: Add service support https://review.openstack.org/373533 | 12:48 |
| alexchadin | hahaha | 12:48 |
| alexchadin | Need to rebase:) | 12:49 |
| openstackgerrit | Alexander Chadin proposed openstack/watcher: Add service supervisor https://review.openstack.org/371632 | 12:52 |
| openstackgerrit | Alexander Chadin proposed openstack/watcher: Add service object to the watcher_db_schema https://review.openstack.org/377634 | 12:53 |
| alexchadin | vincentfrancoise: We can get actions by node name and order them respecting the priorities. | 13:01 |
| alexchadin | vincentfrancoise: It will be done in workflow engine | 13:02 |
| alexchadin | vincentfrancoise: then we don't need to reference them to each other | 13:02 |
| *** alexchadin has quit IRC | 13:12 | |
| openstackgerrit | Merged openstack/watcher: Add service supervisor https://review.openstack.org/371632 | 13:41 |
| *** michaelgugino has joined #openstack-watcher | 14:17 | |
| *** brunograz has joined #openstack-watcher | 14:43 | |
| *** jimbaker has joined #openstack-watcher | 15:20 | |
| acabot | brunograz : hello Bruno | 15:25 |
| brunograz | Hi acabot | 15:25 |
| acabot | brunograz : we will soon merge a template to write doc strategy in Watcher https://review.openstack.org/#/c/382494/ | 15:26 |
| brunograz | acabot: we have been disconnected to IRC for a while :) - Apols for that | 15:26 |
| acabot | brunograz : no pb | 15:26 |
| acabot | brunograz : would it be possible that you update the doc for vm_workload_consolidation according to this template ? | 15:27 |
| acabot | https://github.com/openstack/watcher/blob/master/watcher/decision_engine/strategy/strategies/vm_workload_consolidation.py | 15:27 |
| brunograz | Ok, I see now. we will have a look on this and revert back to you | 15:27 |
| brunograz | is there any deadline for this? | 15:27 |
| acabot | I'd really like to have a clear doc for each strategy available in Watcher before the summit in 2 weeks | 15:28 |
| brunograz | acabot: we will try to make it by then ;) | 15:28 |
| acabot | before end of next week if possible, I think its pretty straightforward if you know the strategy | 15:28 |
| acabot | thanks a lot | 15:28 |
| acabot | brunograz : will you attend the summit ? | 15:30 |
| brunograz | nps - and sorry again for being a bit disconnected | 15:30 |
| brunograz | acabot: yes - myself and Sean will attend | 15:30 |
| acabot | ok great | 15:30 |
| brunograz | but we are not sure if we can make it for the fishbowl session | 15:31 |
| acabot | brunograz : if you plan to attend watcher sessions, here is the agenda https://etherpad.openstack.org/p/watcher-ocata-design-session | 15:31 |
| brunograz | oh great | 15:31 |
| brunograz | thanks :) | 15:32 |
| *** thorst is now known as thorst_afk | 16:04 | |
| *** thorst_afk is now known as thorst | 17:14 | |
| *** michaelgugino has quit IRC | 17:14 | |
| *** michaelgugino has joined #openstack-watcher | 17:28 | |
| *** dtardivel has quit IRC | 18:34 | |
| *** Kevin_Zheng has quit IRC | 18:59 | |
| *** Kevin_Zheng has joined #openstack-watcher | 19:02 | |
| openstackgerrit | Prashanth Hari proposed openstack/watcher-specs: Define grammar for workload characterization https://review.openstack.org/377100 | 19:34 |
| *** hvprash_ has joined #openstack-watcher | 19:45 | |
| *** hvprash_ has quit IRC | 19:50 | |
| *** thorst has quit IRC | 21:05 | |
| *** thorst has joined #openstack-watcher | 21:11 | |
| *** brunograz_ has joined #openstack-watcher | 21:12 | |
| *** thorst has quit IRC | 21:16 | |
| *** zigo_ has joined #openstack-watcher | 21:17 | |
| *** odyssey4me_ has joined #openstack-watcher | 21:17 | |
| *** odyssey4me has quit IRC | 21:18 | |
| *** odyssey4me_ is now known as odyssey4me | 21:18 | |
| *** brunograz has quit IRC | 21:19 | |
| *** licanwei has quit IRC | 21:19 | |
| *** zigo has quit IRC | 21:19 | |
| *** mestery has quit IRC | 21:23 | |
| *** kamtamtun has quit IRC | 21:23 | |
| *** tkaczynski has quit IRC | 21:23 | |
| *** jimbaker has quit IRC | 21:24 | |
| *** mestery has joined #openstack-watcher | 21:28 | |
| *** kamtamtun has joined #openstack-watcher | 21:28 | |
| *** tkaczynski has joined #openstack-watcher | 21:28 | |
| *** zigo_ is now known as zigo | 21:28 | |
| *** michaelgugino has quit IRC | 21:31 | |
| *** thorst has joined #openstack-watcher | 21:45 | |
| *** thorst has quit IRC | 21:49 | |
| *** thorst has joined #openstack-watcher | 21:55 | |
| *** thorst has quit IRC | 22:00 | |
| *** jimbaker has joined #openstack-watcher | 22:49 | |
| *** jimbaker has quit IRC | 23:00 | |
| *** jimbaker has joined #openstack-watcher | 23:15 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!