| *** harlowja has quit IRC | 00:09 | |
| *** sanfern has quit IRC | 00:10 | |
| *** sanfern has joined #openstack-watcher | 00:10 | |
| *** pkoniszewski has quit IRC | 00:13 | |
| *** pkoniszewski has joined #openstack-watcher | 00:16 | |
| *** thorst has joined #openstack-watcher | 00:25 | |
| *** harlowja has joined #openstack-watcher | 00:25 | |
| *** jwcroppe has quit IRC | 00:32 | |
| *** thorst has quit IRC | 00:42 | |
| *** thorst has joined #openstack-watcher | 00:42 | |
| *** thorst has quit IRC | 00:46 | |
| *** thorst has joined #openstack-watcher | 01:11 | |
| *** Yumeng__ has joined #openstack-watcher | 01:12 | |
| *** thorst has quit IRC | 01:16 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/watcher: Should use glanceclient to get images https://review.openstack.org/414445 | 01:33 |
|---|---|---|
| openstackgerrit | Zhenyu Zheng proposed openstack/watcher: Enable coverage report in console output https://review.openstack.org/418640 | 01:48 |
| *** thorst has joined #openstack-watcher | 02:02 | |
| *** thorst has quit IRC | 02:02 | |
| *** sanfern has quit IRC | 02:31 | |
| *** absubram has quit IRC | 03:18 | |
| *** thorst has joined #openstack-watcher | 03:20 | |
| *** thorst has quit IRC | 03:20 | |
| *** absubram has joined #openstack-watcher | 03:25 | |
| *** diga has joined #openstack-watcher | 03:27 | |
| *** absubram has quit IRC | 03:31 | |
| *** Yumeng__ has quit IRC | 04:01 | |
| *** sferna1_ has joined #openstack-watcher | 04:07 | |
| *** absubram has joined #openstack-watcher | 04:55 | |
| *** absubram_ has joined #openstack-watcher | 04:59 | |
| *** absubram has quit IRC | 04:59 | |
| *** absubram_ is now known as absubram | 04:59 | |
| *** thorst has joined #openstack-watcher | 05:09 | |
| *** adisky_ has joined #openstack-watcher | 05:11 | |
| *** thorst has quit IRC | 05:14 | |
| *** sferna1__ has joined #openstack-watcher | 06:05 | |
| *** sferna1_ has quit IRC | 06:08 | |
| *** sferna1__ has quit IRC | 06:31 | |
| *** sanfern has joined #openstack-watcher | 06:31 | |
| openstackgerrit | avnish proposed openstack/watcher-dashboard: Remove useless pytz requirements https://review.openstack.org/418730 | 06:40 |
| *** hidekazu has joined #openstack-watcher | 08:07 | |
| hidekazu | hello, I have a question about watcher high availability configuration. | 08:12 |
| hidekazu | Can watcher be configured as active-active? | 08:12 |
| openstackgerrit | Hidekazu Nakamura proposed openstack/watcher: Add period input parameter to basic strategy https://review.openstack.org/416432 | 08:27 |
| vincentfrancoise | hidekazu: hi | 08:29 |
| hidekazu | vincentfrancoise: hi | 08:30 |
| vincentfrancoise | hidekazu: HA is not possible for all the watcher services | 08:31 |
| vincentfrancoise | hidekazu: the API is stateless so you should be able to deploy many of them | 08:31 |
| hidekazu | vincentfrancoise: oh, why? data model in memory? | 08:32 |
| vincentfrancoise | hidekazu: exactly | 08:32 |
| hidekazu | vincentfrancoise: api is as i expected. | 08:32 |
| vincentfrancoise | hidekazu: because the data model is currently in-memory, that may cause some issues | 08:33 |
| vincentfrancoise | hidekazu: I personally didn't dig into this much yet | 08:33 |
| vincentfrancoise | but someone already asked a similar question | 08:33 |
| hidekazu | vincentfrancoise: i have never heard this topic. | 08:34 |
| vincentfrancoise | yes let me find it | 08:34 |
| hidekazu | vincentfrancoise: if we release for production purpose, ha feature is needed. | 08:35 |
| vincentfrancoise | https://bugs.launchpad.net/watcher/+bug/1598946 | 08:36 |
| openstack | Launchpad bug 1598946 in watcher "Initial execution of parallel engines can be racey " [Undecided,Won't fix] - Assigned to Zhai, Edwin (edwin-zhai) | 08:36 |
| vincentfrancoise | hidekazu: the rationale around not pushing too much towards HA is that Watcher is not a critical service | 08:37 |
| vincentfrancoise | hidekazu: its purpose is to make things better in the cluster so if it fails, nothing bad should actually come out of it | 08:38 |
| hidekazu | vincentfrancoise: yes | 08:38 |
| vincentfrancoise | hidekazu: but I totally agree that we should eventually find a way to make it HA-ready | 08:38 |
| hidekazu | vincentfrancoise: i agree better. | 08:39 |
| vincentfrancoise | hidekazu: one of the issues is that we need to introduce a persistent model to make the decision engine stateless | 08:39 |
| hidekazu | vincentfrancoise: so you introduce graph model instead of data model? | 08:41 |
| vincentfrancoise | hidekazu: the other issue is the question of action plan validaity | 08:41 |
| *** dtardivel has joined #openstack-watcher | 08:41 | |
| vincentfrancoise | hidekazu: it's not the main reason but we could eventually use a graph DB like neo4j to solve this yes | 08:42 |
| openstackgerrit | Merged openstack/watcher-dashboard: Remove useless pytz requirements https://review.openstack.org/418730 | 08:43 |
| vincentfrancoise | hidekazu: this model persistence needs more thought around it because models are pluggable, so what about the 3rd party ones? Should we impose that all model should be graph-based? Should we propose a single unified model that is graph-based? There are many ways to solve this | 08:44 |
| hidekazu | vincentfrancoise: decision-engine service can have worker process by configuration. but should not to be. right? | 08:44 |
| hidekazu | vincentfrancoise: applier is also. | 08:45 |
| vincentfrancoise | hidekazu: decision-engine already uses a worker pool to execute audits | 08:45 |
| vincentfrancoise | vfrancoise: but it's not multi-process because we need to share the memory space for the models | 08:46 |
| hidekazu | worker pool means green thread? | 08:46 |
| vincentfrancoise | hidekazu: the applier can be HA-ready without too much effort if we use taskflow I think | 08:46 |
| vincentfrancoise | hidekazu: because taskflow can somehow persist the workflow it executes if I'm not mistaken | 08:47 |
| vincentfrancoise | hidekazu: that's a standard thread pool (see https://github.com/openstack/watcher/blob/master/watcher/decision_engine/messaging/audit_endpoint.py#L36-L37) | 08:47 |
| hidekazu | vincentfrancoise: thanks, i misunderstood. | 08:49 |
| hidekazu | vincentfrancoise: i confirmed conf file. max_workers in watcher_decision_engine section is the number of threads. | 08:50 |
| vincentfrancoise | hidekazu: you got it :) | 08:50 |
| vincentfrancoise | hidekazu: so it's only a starting point for HA | 08:51 |
| vincentfrancoise | hidekazu: but I try to keep this in mind ever since so we doin't should ouselves in the foot once we decide to make HA a priority | 08:51 |
| vincentfrancoise | shoot ourselves* | 08:51 |
| hidekazu | vincentfrancoise: thanks for all :) | 08:52 |
| vincentfrancoise | but if you wish to tackle this problem, feel free :) | 08:52 |
| openstackgerrit | Deepak Jon proposed openstack/python-watcherclient: Removes unnecessary utf-8 encoding https://review.openstack.org/418779 | 08:54 |
| openstackgerrit | Merged openstack/watcher: Enable coverage report in console output https://review.openstack.org/415824 | 09:00 |
| openstackgerrit | avnish proposed openstack/watcher-dashboard: Remove useless ddt requirements https://review.openstack.org/418784 | 09:03 |
| *** Yumeng has joined #openstack-watcher | 09:05 | |
| *** hanrong has quit IRC | 09:26 | |
| *** hanrong has joined #openstack-watcher | 09:27 | |
| *** hidekazu has left #openstack-watcher | 09:44 | |
| *** Yumeng__ has joined #openstack-watcher | 09:46 | |
| *** Yumeng has quit IRC | 10:03 | |
| *** hanrong has quit IRC | 10:04 | |
| *** hanrong has joined #openstack-watcher | 10:05 | |
| *** hanrong has quit IRC | 10:05 | |
| *** hanrong has joined #openstack-watcher | 10:06 | |
| *** Yumeng has joined #openstack-watcher | 10:10 | |
| *** Yumeng has quit IRC | 10:14 | |
| *** zhurong has joined #openstack-watcher | 10:46 | |
| openstackgerrit | Merged openstack/python-watcherclient: Removes unnecessary utf-8 encoding https://review.openstack.org/418779 | 10:51 |
| openstackgerrit | Merged openstack/watcher-dashboard: Remove useless ddt requirements https://review.openstack.org/418784 | 10:54 |
| openstackgerrit | Merged openstack/watcher: Removed unnecessary utf-8 encoding https://review.openstack.org/418805 | 10:55 |
| *** zhurong has quit IRC | 11:35 | |
| *** Yumeng__ has quit IRC | 12:11 | |
| *** hanrong1 has joined #openstack-watcher | 12:25 | |
| *** sanfern has quit IRC | 12:29 | |
| *** thorst has joined #openstack-watcher | 12:41 | |
| *** hanrong1 has quit IRC | 12:51 | |
| *** sanfern has joined #openstack-watcher | 13:51 | |
| sballe_ | morning | 13:56 |
| vincentfrancoise | sballe_: hi | 13:56 |
| acabot | morning | 13:59 |
| sballe_ | acabot: thx for setting up: https://etherpad.openstack.org/p/Watcher_abstracts_boston2017 I was rushing ot do it before todays meeting and you had already done it :) | 13:59 |
| acabot | weekly meeting starts in #openstack-meeting-4 | 13:59 |
| acabot | sballe_ : no pb ;-) | 13:59 |
| *** lamt has joined #openstack-watcher | 14:01 | |
| *** Yumeng__ has joined #openstack-watcher | 14:03 | |
| *** diga has quit IRC | 14:06 | |
| openstackgerrit | Vincent Françoise proposed openstack/watcher: Remove obsolete Resource element https://review.openstack.org/418974 | 14:33 |
| openstackgerrit | Vincent Françoise proposed openstack/watcher: Remove obsolete Resource element https://review.openstack.org/418974 | 14:36 |
| openstackgerrit | Vincent Françoise proposed openstack/watcher: [PoC] Graph cluster model instead of mapping one https://review.openstack.org/418434 | 14:36 |
| *** lamt has quit IRC | 14:40 | |
| *** jwcroppe has joined #openstack-watcher | 14:46 | |
| *** michaelgugino has joined #openstack-watcher | 14:53 | |
| *** lamt has joined #openstack-watcher | 16:03 | |
| *** ss4 has joined #openstack-watcher | 16:29 | |
| *** ss4 has quit IRC | 17:02 | |
| *** ss4 has joined #openstack-watcher | 17:14 | |
| openstackgerrit | David TARDIVEL proposed openstack/watcher: Add first alembic version for db migration https://review.openstack.org/419031 | 17:28 |
| *** sanfern has quit IRC | 17:30 | |
| *** vincentfrancoise has quit IRC | 17:38 | |
| *** absubram has quit IRC | 18:36 | |
| *** absubram has joined #openstack-watcher | 18:39 | |
| *** adisky_ has quit IRC | 18:39 | |
| *** Yumeng__ has quit IRC | 19:11 | |
| *** pshedimb has joined #openstack-watcher | 19:27 | |
| *** tmetsch has joined #openstack-watcher | 20:04 | |
| *** tmetsch has quit IRC | 20:16 | |
| *** ss4 has quit IRC | 20:50 | |
| *** jwcroppe has quit IRC | 21:14 | |
| *** thorst has quit IRC | 21:54 | |
| *** thorst has joined #openstack-watcher | 21:57 | |
| *** thorst has quit IRC | 22:02 | |
| *** michaelgugino has quit IRC | 22:27 | |
| *** dtardivel has quit IRC | 22:34 | |
| *** hanrong has quit IRC | 22:37 | |
| *** hanrong has joined #openstack-watcher | 22:41 | |
| *** lamt has quit IRC | 23:11 | |
| *** thorst_ has joined #openstack-watcher | 23:48 | |
| openstackgerrit | Matthew J Black proposed openstack/puppet-watcher: Allow python-memcache install from authtoken class https://review.openstack.org/419201 | 23:53 |
| *** thorst_ has quit IRC | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!