| *** thorst has joined #openstack-watcher | 00:02 | |
| *** thorst has quit IRC | 00:04 | |
| *** vtech has quit IRC | 00:04 | |
| *** thorst has joined #openstack-watcher | 00:04 | |
| *** tpeoples has quit IRC | 00:08 | |
| *** tpeoples has joined #openstack-watcher | 00:11 | |
| *** thorst has quit IRC | 00:13 | |
| *** thorst has joined #openstack-watcher | 01:11 | |
| *** thorst has quit IRC | 01:18 | |
| *** thorst has joined #openstack-watcher | 02:16 | |
| *** thorst has quit IRC | 02:23 | |
| *** thorst has joined #openstack-watcher | 03:21 | |
| *** thorst has quit IRC | 03:27 | |
| *** thorst has joined #openstack-watcher | 04:26 | |
| *** thorst has quit IRC | 04:33 | |
| *** Kevin_Zheng has quit IRC | 05:02 | |
| *** jwcroppe has joined #openstack-watcher | 05:23 | |
| *** thorst has joined #openstack-watcher | 05:31 | |
| *** thorst has quit IRC | 05:37 | |
| *** jwcroppe has quit IRC | 06:30 | |
| *** thorst has joined #openstack-watcher | 06:36 | |
| *** thorst has quit IRC | 06:42 | |
| *** thorst has joined #openstack-watcher | 07:40 | |
| *** thorst has quit IRC | 07:47 | |
| *** vtech has joined #openstack-watcher | 07:54 | |
| *** vincentfrancoise has joined #openstack-watcher | 08:29 | |
| *** thorst has joined #openstack-watcher | 08:46 | |
| *** thorst has quit IRC | 08:52 | |
| *** thorst has joined #openstack-watcher | 09:50 | |
| *** thorst has quit IRC | 09:57 | |
| *** openstackgerrit has quit IRC | 10:02 | |
| *** openstackgerrit_ has joined #openstack-watcher | 10:02 | |
| *** openstackgerrit_ has quit IRC | 10:03 | |
| *** dtardivel has joined #openstack-watcher | 10:10 | |
| *** thorst has joined #openstack-watcher | 10:56 | |
| *** thorst has quit IRC | 11:03 | |
| *** vtech has quit IRC | 11:16 | |
| *** jwcroppe has joined #openstack-watcher | 11:26 | |
| *** openstackgerrit has joined #openstack-watcher | 11:53 | |
| *** openstackgerrit has quit IRC | 11:54 | |
| *** openstackgerrit_ has joined #openstack-watcher | 11:54 | |
| *** openstackgerrit_ is now known as openstackgerrit | 11:55 | |
| *** vtech has joined #openstack-watcher | 11:56 | |
| *** openstackgerrit has quit IRC | 11:59 | |
| *** thorst has joined #openstack-watcher | 12:01 | |
| *** openstackgerrit has joined #openstack-watcher | 12:07 | |
| *** thorst has quit IRC | 12:08 | |
| *** thorst has joined #openstack-watcher | 12:48 | |
| *** cjvolzka has joined #openstack-watcher | 14:10 | |
| *** cjvolzka has quit IRC | 14:13 | |
| *** cjvolzka has joined #openstack-watcher | 14:14 | |
| openstackgerrit | Edwin Zhai proposed openstack/watcher-specs: optimization threshold spec https://review.openstack.org/274688 | 14:36 |
|---|---|---|
| dtardivel | Hi tpeoples | 14:46 |
| *** jwcroppe has quit IRC | 15:07 | |
| *** vtech has quit IRC | 15:20 | |
| *** vtech has joined #openstack-watcher | 15:21 | |
| tpeoples | Hey dtardivel. I'm trying to do something like: http://paste.openstack.org/show/485624/ in order to handle the 4.2.0 change, but I can't get it working due to DuplicateOptErrors in the UTs, but it seems to work fine in devstack. The thing here, though, is that the project_domain_id and user_domain_id need to be defined in the conf file, which is what | 15:40 |
| tpeoples | keystoneauth1.loading expects | 15:40 |
| dtardivel | Yes I saw your message this week-end | 15:40 |
| tpeoples | It seems like keystoneauth1.loading is the way to go here, though | 15:41 |
| dtardivel | vincentfrancoise and I have found a way to get them. Can I push another patchset above yours ? | 15:41 |
| tpeoples | sure | 15:41 |
| tpeoples | i was banging my head on Friday trying to figure out what was needed :$ | 15:41 |
| vincentfrancoise | hi tpeoples | 15:42 |
| tpeoples | hey vincentfrancoise | 15:42 |
| vincentfrancoise | with dtardivel we were also struggling quite a bit figuring all this out... | 15:42 |
| dtardivel | https://www.irccloud.com/pastebin/5SPioRkS/ | 15:42 |
| tpeoples | ok so you're still manually creating the admin creds dictionary. did you guys try playing with the keystoneauth1.loading library at all? | 15:44 |
| dtardivel | It works on my testing env (containers distributed watcher). Not tested again on devstack | 15:44 |
| tpeoples | dtardivel: i can test it on devstack. are you going to push it up? + my comment above^ | 15:46 |
| *** jwcroppe has joined #openstack-watcher | 15:47 | |
| dtardivel | We have errors on unit tests related to xxx_helpers() with this patchset. We working on them : ) | 15:47 |
| vincentfrancoise | side note: the thing didn't quite understand with all this is that I couldn't find any example on other projects using "load_auth_from_conf_options" or "register_conf_options" | 15:47 |
| *** jwcroppe_ has joined #openstack-watcher | 15:48 | |
| *** jwcroppe has quit IRC | 15:51 | |
| tpeoples | lol i know vincentfrancoise, the docs on this are awful. every project seems to do it differently | 15:53 |
| vincentfrancoise | +1000 | 15:54 |
| tpeoples | vincentfrancoise: https://github.com/openstack/neutron/blob/master/neutron/notifiers/nova.py#L49 | 15:55 |
| tpeoples | https://github.com/openstack/ceilometer/blob/master/ceilometer/keystone_client.py#L105-L124 | 15:56 |
| tpeoples | those were the only two examples of things using load_*_from_conf_options i could find | 15:57 |
| vincentfrancoise | good catch! I was looking at magnum, ironic, nova, glance and swift (and zaqar because they made a ticket on launchpad)... | 15:58 |
| tpeoples | vincentfrancoise: what i posted in the paste above works, but seems to cause issues with the UTs. I'm not sure if something in our keystone middleware code is also implicitly registering those options | 16:01 |
| vincentfrancoise | Overall I think we got to the same issue, the only thing I don't like much about using keystoneauth1 in your approach is that it looks like we are kind of breaking the "abstraction" keystoneclient provides | 16:04 |
| vincentfrancoise | But TBH, I think that if you get it working this way, I won't be picky... | 16:05 |
| tpeoples | vincentfrancoise: just asked in #openstack-keystone but i doubt i'll get a reply :( lol | 16:12 |
| vincentfrancoise | wait and see I guess... | 16:28 |
| tpeoples | vincentfrancoise: what do you think about having a CONF file that included this: http://paste.openstack.org/show/485633/ ugly cause it's duplicated, but i think this might get around the duplicate opt errors | 16:31 |
| *** gabor_antal has joined #openstack-watcher | 16:53 | |
| vincentfrancoise | hahaha, with dtardivel we were thinking about doing it kind of like in https://github.com/openstack/ceilometer/blob/master/ceilometer/keystone_client.py#L106 were they have put in a 'service_credentials' group | 16:55 |
| vincentfrancoise | so that would pretty much generate some dupes as well | 16:56 |
| *** efoley has joined #openstack-watcher | 16:58 | |
| *** jwcroppe_ has quit IRC | 17:14 | |
| *** vtech has quit IRC | 17:16 | |
| tpeoples | i think that may be the easiest solution vincentfrancoise | 17:22 |
| tpeoples | i am heading to lunch with jwcroppe in a few minutes but i will see if i can get something with that working | 17:23 |
| vincentfrancoise | I'll be leaving soon too. Have a good day :) | 17:24 |
| *** efoley_ has joined #openstack-watcher | 17:29 | |
| *** efoley has quit IRC | 17:31 | |
| *** efoley_ has quit IRC | 17:39 | |
| vincentfrancoise | So just before leaving, here are our conclusions | 17:41 |
| vincentfrancoise | 1st option: we do it the DevStack way where we would have a single section with the old params (admin_tenant_name, ...) | 17:42 |
| vincentfrancoise | 2nd option: We do it with 2 sections like in ceilometer but that would imply more code changes and further documentation, but we wouldn't have the deprecation warnings anymore | 17:43 |
| vincentfrancoise | As dtardivel points out, option 1 would only mean having to explain in the doc we need to add the auth_plugin parameters within the [keystone_authtoken] section | 17:44 |
| vincentfrancoise | Either way, there seems to be no good solution ... Will leave you deciding or if you have some clues from the keystone guys | 17:45 |
| vincentfrancoise | But because Watcher does not work outside a DevStack anymore, I think we should find a solution fairly quickly event if it's not perfect | 17:46 |
| *** vincentfrancoise has quit IRC | 17:51 | |
| *** vtech has joined #openstack-watcher | 18:14 | |
| *** vtech has quit IRC | 18:24 | |
| openstackgerrit | Gábor Antal proposed openstack/watcher: Removed unused parameter in dt_deserializer() https://review.openstack.org/274838 | 18:29 |
| *** jwcroppe has joined #openstack-watcher | 18:29 | |
| *** vtech has joined #openstack-watcher | 18:49 | |
| *** dtardivel has quit IRC | 20:48 | |
| *** thorst has quit IRC | 21:12 | |
| *** vfrancoise has joined #openstack-watcher | 21:22 | |
| *** dtardivel has joined #openstack-watcher | 21:29 | |
| *** thorst has joined #openstack-watcher | 21:31 | |
| sballe | jwcroppe: ping | 22:11 |
| sballe | acabot: hi | 22:11 |
| jwcroppe | sballe: hey - otp | 22:11 |
| sballe | Can you attend the cross projdct meeting tomorrow ? I am at a conference and cannot be there | 22:12 |
| sballe | I might see if I can attend on IRC while the session is on but no guarantee that I can | 22:12 |
| *** vtech has quit IRC | 22:15 | |
| jwcroppe | What time is it? We have mid-cycle | 22:23 |
| jwcroppe | doubtful I can make tomorrow unless it's late at night | 22:23 |
| jwcroppe | sballe: ^^ | 22:37 |
| *** cjvolzka has quit IRC | 23:05 | |
| tpeoples | jwcroppe: i think it's at 3 CST (21 UTC) | 23:07 |
| *** jwcroppe has quit IRC | 23:11 | |
| *** vfrancoise has quit IRC | 23:14 | |
| *** dtardivel has quit IRC | 23:38 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!