| *** harlowja has quit IRC | 01:42 | |
| *** nicolasbock has quit IRC | 01:50 | |
| *** harlowja has joined #openstack-watcher | 03:37 | |
| *** harlowja has quit IRC | 05:22 | |
| openstackgerrit | Hidekazu Nakamura proposed openstack/watcher-tempest-plugin master: Add zone migration strategy scenario job https://review.openstack.org/559001 | 05:55 |
|---|---|---|
| *** hidekazu has joined #openstack-watcher | 05:55 | |
| *** hidekazu has quit IRC | 06:15 | |
| *** AlexeyAbashkin has joined #openstack-watcher | 06:38 | |
| *** danpawlik has joined #openstack-watcher | 06:55 | |
| *** alexchadin has joined #openstack-watcher | 07:42 | |
| *** alexchad_ has joined #openstack-watcher | 07:51 | |
| *** alexchadin has quit IRC | 07:54 | |
| *** alexchadin has joined #openstack-watcher | 08:09 | |
| *** logan- has quit IRC | 08:11 | |
| *** alexchad_ has quit IRC | 08:11 | |
| *** logan- has joined #openstack-watcher | 08:13 | |
| *** hidekazu has joined #openstack-watcher | 08:18 | |
| hidekazu | alexchadin: hi | 08:18 |
| alexchadin | hidekazu: hi | 08:19 |
| hidekazu | alexchadin: i wrote zone migration tempest test. could you teach me how to add it to job? | 08:20 |
| hidekazu | alexchdin: i tried , but failing. https://review.openstack.org/#/c/559001/ | 08:20 |
| alexchadin | hidekazu: sure | 08:20 |
| alexchadin | hidekazu: just add job definition like it's done here: https://github.com/openstack/watcher/blob/master/.zuul.yaml#L41 | 08:21 |
| hidekazu | alexchadin: before adding test to tempest-plugin, i should add it's definition to watcher repo,right? | 08:22 |
| alexchadin | hidekazu: you may add this definition to .zuul.yaml file in tempest-plugin repo | 08:23 |
| alexchadin | hidekazu: your job will find watcher-tempest-multinode job by itself | 08:23 |
| alexchadin | hidekazu: once you find out that everything is fine with your test, you may add your job to .zuul.yaml in watcher repo | 08:24 |
| alexchadin | your job definition* | 08:25 |
| hidekazu | alexchadin: zuul said Job watcher-tempest-zone_migration not defined in https://review.openstack.org/#/c/559001/ | 08:26 |
| hidekazu | alexchadin: that will be solved by adding the definition to wathcer repo? | 08:26 |
| alexchadin | hidekazu: you can add zone-migration job definition to watcher-tempest-plugin repo for a while. if your zone-migration scenario works fine, you may remove zone-migration job definition from watcher-tempest-plugin and add it to watcher repo | 08:29 |
| alexchadin | hidekazu: once you add it to watcher repo, we can merge your https://review.openstack.org/#/c/559001/ | 08:30 |
| alexchadin | hidekazu: does it make sense? | 08:33 |
| alexchadin | :) | 08:33 |
| hidekazu | alexchdin: so https://review.openstack.org/#/c/559001/ -1 means my test does not work.. | 08:34 |
| alexchadin | hidekazu: let me help you | 08:36 |
| openstackgerrit | Alexander Chadin proposed openstack/watcher-tempest-plugin master: Add zone migration strategy scenario job https://review.openstack.org/559001 | 08:37 |
| alexchadin | hidekazu: I've updated your commit | 08:37 |
| hidekazu | alexchadin: ah, i understand you said now. | 08:38 |
| alexchadin | hidekazu: have you tried your zone-migration test in your local env? | 08:39 |
| hidekazu | alexchadin: yes, 3 compute nodes. | 08:39 |
| hidekazu | alexchadin: test works fine in my env. | 08:39 |
| alexchadin | hidekazu: no worries, I've spent 24 days to fix our functional and scenario jobs and migrate them to zuulv3 | 08:40 |
| hidekazu | alexchadin: great work! | 08:40 |
| alexchadin | hidekazu: thank you! | 08:41 |
| alexchadin | hidekazu: there is a thing I'd like to discuss with you | 08:45 |
| hidekazu | alexchadin: lets discuss | 08:46 |
| alexchadin | hidekazu: I'd like to implement filters in Watcher to sync them with Nova | 08:46 |
| alexchadin | hidekazu: like affinity/anti-affinity filters, aggregate, availability zones filters | 08:47 |
| alexchadin | hidekazu: to respect Nova Scheduler filters | 08:47 |
| alexchadin | hidekazu: since Nova doesn't open API for Scheduler, I don't see other ways to achieve it | 08:48 |
| hidekazu | alexchadin: you want strategy to migrate instances in according to the filter | 08:48 |
| alexchadin | hidekazu: right, and I want strategy to take into account filters to build good action plan with minimal risk of migration failures | 08:49 |
| alexchadin | I know, it isn't very good way, but our business level requirements asks about respecting Nova filters | 08:50 |
| hidekazu | alexchadin: cant we use nova-scheduler by mesagging? | 08:51 |
| alexchadin | hidekazu: by oslo.messaging? | 08:51 |
| hidekazu | alexchadin: yes, just an idea or question :) | 08:51 |
| alexchadin | hidekazu: I don't think that watcher can access Nova's internal service by messaging | 08:52 |
| alexchadin | hidekazu: I suppose it's private API to let Nova's service interact to each others | 08:53 |
| hidekazu | alexchadin: when booting instance, nova interacts with neutron by messaging. | 08:55 |
| alexchadin | hidekazu: could you please give me a link? | 08:55 |
| hidekazu | alexchadin: i saw code 4 years ago.. | 08:59 |
| alexchadin | hidekazu: I'll investigate it | 09:02 |
| hidekazu | alexchadin: hmm, it may be wrong. | 09:03 |
| alexchadin | hidekazu: they use API, not RPC | 09:10 |
| hidekazu | alexchadin: sorry, i was wrong. | 09:11 |
| alexchadin | hidekazu: so what do you think about this situation? | 09:11 |
| hidekazu | alexchadin: you will copy and paste allmost code from nova? | 09:13 |
| alexchadin | hidekazu: we can start with a little amount of filters and start to adapt them to Watcher | 09:14 |
| alexchadin | hidekazu: we can't use the same code since it works with internal Nova objects and they should be replaced with API-called objects | 09:15 |
| hidekazu | alexchadin: did you investigate placement api? | 09:15 |
| alexchadin | hidekazu: a little bit. what exactly do you mean? | 09:16 |
| hidekazu | alexchadin: nova scheduler calls placement api. but i do not know detail . | 09:18 |
| alexchadin | hidekazu: it may call it to gather information about provided compute resources by each node | 09:19 |
| hidekazu | i have stomachache. so go home and take a rest. | 10:14 |
| *** hidekazu has left #openstack-watcher | 10:15 | |
| *** nicolasbock has joined #openstack-watcher | 10:30 | |
| *** alexchadin has quit IRC | 11:32 | |
| *** alexchadin has joined #openstack-watcher | 11:42 | |
| *** alexchadin has quit IRC | 11:47 | |
| *** alexchadin has joined #openstack-watcher | 12:14 | |
| *** alexchadin has quit IRC | 12:19 | |
| *** alexchadin has joined #openstack-watcher | 12:20 | |
| *** alexchadin has quit IRC | 12:38 | |
| openstackgerrit | Alexander Chadin proposed openstack/watcher master: [WiP] docs update https://review.openstack.org/558559 | 12:43 |
| *** jdandrea has joined #openstack-watcher | 14:12 | |
| *** openstackgerrit has quit IRC | 15:34 | |
| *** AlexeyAbashkin has quit IRC | 16:10 | |
| *** harlowja has joined #openstack-watcher | 17:21 | |
| *** harlowja has quit IRC | 17:25 | |
| *** AlexeyAbashkin has joined #openstack-watcher | 17:34 | |
| *** AlexeyAbashkin has quit IRC | 17:38 | |
| *** harlowja has joined #openstack-watcher | 18:29 | |
| *** openstackgerrit has joined #openstack-watcher | 20:23 | |
| openstackgerrit | Egor Panfilov proposed openstack/watcher-specs master: Watcher Planner Selector https://review.openstack.org/558614 | 20:23 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!