| openstackgerrit | Hidekazu Nakamura proposed openstack/watcher: Add additional depencencies of CentOS 7 https://review.openstack.org/415038 | 00:36 |
|---|---|---|
| *** stack_ has joined #openstack-watcher | 02:10 | |
| *** stack_ has left #openstack-watcher | 02:11 | |
| *** licanwei has joined #openstack-watcher | 02:12 | |
| openstackgerrit | Hidekazu Nakamura proposed openstack/watcher: Update configuration document https://review.openstack.org/415088 | 04:02 |
| *** hidekazu has joined #openstack-watcher | 04:09 | |
| *** atuly has quit IRC | 04:22 | |
| openstackgerrit | Hidekazu Nakamura proposed openstack/watcher: Fix TypeError if no input_parameters added https://review.openstack.org/415096 | 04:52 |
| *** vtech has joined #openstack-watcher | 05:37 | |
| *** vtech has quit IRC | 05:46 | |
| *** atuly has joined #openstack-watcher | 06:27 | |
| *** licanwei has left #openstack-watcher | 06:43 | |
| openstackgerrit | Alexander Chadin proposed openstack/watcher: New default planner https://review.openstack.org/406991 | 08:33 |
| *** alexchadin has joined #openstack-watcher | 08:34 | |
| *** vincentfrancoise has joined #openstack-watcher | 08:39 | |
| openstackgerrit | Merged openstack/watcher: Fix reference http https://review.openstack.org/414865 | 09:26 |
| openstackgerrit | Merged openstack/watcher: Add additional depencencies of CentOS 7 https://review.openstack.org/415038 | 09:27 |
| vincentfrancoise | alexchadin: hi | 09:31 |
| alexchadin | vincentfrancoise: morning | 09:31 |
| vincentfrancoise | alexchadin: I was quickly going through your latest PS for the planner | 09:32 |
| vincentfrancoise | and I was wondering why you added a new resize action | 09:32 |
| vincentfrancoise | it seems to me it is not really related to this BP | 09:33 |
| alexchadin | vincentfrancoise: Resize may be need in some new strategies | 09:33 |
| alexchadin | vincentfrancoise: Yes, it isn't related to the BP directly | 09:34 |
| vincentfrancoise | can you instead push this in a separate PS the meake the new planner code easier to read? | 09:34 |
| alexchadin | vincentfrancoise: but it shows how `parents` field is filled | 09:34 |
| alexchadin | vincentfrancoise: It will take some more time to separate it, but I think yes, I can do it | 09:35 |
| alexchadin | if it is necessary | 09:35 |
| vincentfrancoise | oh let me see the parents | 09:35 |
| alexchadin | David said that my previous patch made actions centralized, not pluggable | 09:36 |
| vincentfrancoise | it's just that I wasn't understanding the reason why you pushed it along with the planner | 09:36 |
| alexchadin | that's why I made new `validate_parents` method and actions is now pluggable again | 09:37 |
| alexchadin | vincentfrancoise: you mean resize? | 09:37 |
| vincentfrancoise | alexchadin: yeah but now that you explain it I get why | 09:38 |
| alexchadin | vincentfrancoise: I didn't test applying of resize action, but planner build actions well | 09:40 |
| vincentfrancoise | this PS is hard to review because I didn't follow all the talks that happened when I was on holidays :( | 09:40 |
| alexchadin | vincentfrancoise: the main goal is to make applying of actions in parallel | 09:41 |
| alexchadin | vincentfrancoise: we have separate this goal to two BP: modifying of planner and changes to applier | 09:42 |
| vincentfrancoise | alexchadin: yeah I remember that | 09:42 |
| alexchadin | vincentfrancoise: I've uploaded first BP that adds `parents` field to action table that will have list of action uuids | 09:43 |
| alexchadin | vincentfrancoise: but, some actions shouldn't be linked | 09:43 |
| alexchadin | vincentfrancoise: for example, migration, nop, sleep, change_nova_node_status - all of them can be done without any dependencies | 09:44 |
| alexchadin | vincentfrancoise: but there are resize and turn_node_to_acpi_s3 (it isn't defined as action type yet) actions that should have parents | 09:45 |
| alexchadin | resize of specified action must go after migration of this action | 09:45 |
| alexchadin | and turn_node_to_acpi_s3 must go only when all migration and resize actions are done to specified host | 09:46 |
| vincentfrancoise | alexchadin: so validate_parents() will be the acting according the parents types? | 09:47 |
| alexchadin | vincentfrancoise: yeap | 09:47 |
| alexchadin | vincentfrancoise: I've added dummy_with_resize strategy, take a look at it | 09:47 |
| vincentfrancoise | alexchadin: does it have a tempest test to run the resize? | 09:50 |
| alexchadin | vincentfrancoise: not yet | 09:50 |
| vincentfrancoise | alexchadin: ok | 09:50 |
| *** alexchadin has quit IRC | 10:16 | |
| *** alexchadin has joined #openstack-watcher | 10:16 | |
| alexchadin | vincentfrancoise: does review.openstack.org work good for you? | 10:22 |
| openstackgerrit | Alexander Chadin proposed openstack/watcher: New default planner https://review.openstack.org/406991 | 10:22 |
| vincentfrancoise | alexchadin: yeap | 10:23 |
| vincentfrancoise | alexchadin: there is something I want to double check before telling you | 10:23 |
| alexchadin | vincentfrancoise: yes? | 10:23 |
| vincentfrancoise | alexchadin: about using the action factory within the planner | 10:23 |
| vincentfrancoise | alexchadin: if you use the action factory to load the action plugins within the planner | 10:24 |
| vincentfrancoise | alexchadin: that means you need to install an action both where you have the decision engine and where the applier are running | 10:24 |
| alexchadin | but we usually install all the code despite of services we plan to use | 10:25 |
| vincentfrancoise | alexchadin: so the separation of concern is not full here | 10:25 |
| alexchadin | we have full nova code set on compute node, along with conductor, api, etc | 10:26 |
| vincentfrancoise | alexchadin: in practice yes but that's the idea behind using action dicts within the strategies in the first place | 10:27 |
| vincentfrancoise | alexchadin: like https://github.com/openstack/watcher/blob/master/watcher/decision_engine/strategy/strategies/basic_consolidation.py#L295-L297 does not reference the action object on purpose | 10:28 |
| alexchadin | I see | 10:28 |
| alexchadin | hm | 10:29 |
| vincentfrancoise | alexchadin: the idea is that api, decision_engine and applier should be considered as separate projects | 10:29 |
| vincentfrancoise | even if under the same repo | 10:29 |
| alexchadin | we can move actions to the common code set:) | 10:29 |
| alexchadin | like it is done with objects | 10:29 |
| vincentfrancoise | alexchadin: action is already in 'objects' | 10:30 |
| alexchadin | I mean, to move watcher/applier/actions folder somewhere | 10:31 |
| vincentfrancoise | alexchadin: that's why I was telling you that I wanted to double check | 10:33 |
| vincentfrancoise | alexchadin: it's because I don't have a good solution to propose you yet :p | 10:33 |
| alexchadin | oh | 10:33 |
| alexchadin | btw | 10:33 |
| alexchadin | merry christmas:) | 10:34 |
| vincentfrancoise | alexchadin: merry christmas to you too :) | 10:34 |
| alexchadin | vincentfrancoise: we can move watcher/watcher/applier/actions to the watcher/watcher/, on the same level where are notifications, objects, tests.. | 10:46 |
| vincentfrancoise | alexchadin: we could but I do not like the idea of making actions a first-order package | 10:47 |
| alexchadin | vincentfrancoise: me too | 10:47 |
| alexchadin | :( | 10:47 |
| vincentfrancoise | alexchadin: as Time Peters says: Special cases aren't special enough to break the rules. :p | 10:48 |
| alexchadin | soooo we won't make any changes:D | 10:48 |
| vincentfrancoise | alexchadin: that's why I'm thinking on how to make this all work | 10:49 |
| vincentfrancoise | will come back to you if I find anything | 10:49 |
| alexchadin | I will be back in one hour estimate | 10:49 |
| *** alexchadin has quit IRC | 10:49 | |
| vincentfrancoise | I'll be on lunch break soon so if I come back to you that would be in ~3h I guess | 10:50 |
| *** alexchadin has joined #openstack-watcher | 13:11 | |
| *** alexchadin1 has joined #openstack-watcher | 13:14 | |
| *** alexchadin has quit IRC | 13:18 | |
| *** alexchadin1 has quit IRC | 13:18 | |
| *** alexchadin has joined #openstack-watcher | 13:19 | |
| vincentfrancoise | alexchadin: ping | 13:21 |
| alexchadin | vincentfrancoise: pong | 13:26 |
| vincentfrancoise | alexchadin: where is turn_host_to_acpi_s3_state turn_host_to_acpi_s3_state? | 13:27 |
| vincentfrancoise | alexchadin: I can't see any new action with this name | 13:27 |
| alexchadin | vincentfrancoise: so, there is no action yet | 13:28 |
| alexchadin | vincentfrancoise: cause there is no strategy that would use it soon | 13:29 |
| alexchadin | vincentfrancoise: I've leaved it just to show how parents field should work | 13:29 |
| vincentfrancoise | alexchadin: IMHO we should have a strategy that takes as parameter an explicit action to execute | 13:30 |
| vincentfrancoise | alexchadin: that way we could easily test all the actiona | 13:30 |
| vincentfrancoise | actions* | 13:30 |
| vincentfrancoise | instead of doing a new dummy_with* everytime | 13:31 |
| alexchadin | vincentfrancoise: maybe | 13:31 |
| vincentfrancoise | but that's just a nice to have | 13:32 |
| alexchadin | vincentfrancoise: something like "fullset" strategy:) | 13:32 |
| vincentfrancoise | yep | 13:32 |
| *** alexchadin has quit IRC | 14:04 | |
| *** lamt has joined #openstack-watcher | 14:23 | |
| *** openstackgerrit_ has joined #openstack-watcher | 15:55 | |
| *** openstackgerrit_ has quit IRC | 15:57 | |
| openstackgerrit | Merged openstack/watcher: Fix TypeError if no input_parameters added https://review.openstack.org/415096 | 16:21 |
| *** vincentfrancoise has quit IRC | 17:57 | |
| *** vtech has joined #openstack-watcher | 22:57 | |
| *** vtech has quit IRC | 23:21 | |
| *** hidekazu has quit IRC | 23:34 | |
| *** lamt has quit IRC | 23:36 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!