| *** lhcheng has quit IRC | 00:29 | |
| *** lhcheng has joined #openstack-keystone | 00:34 | |
| *** ChanServ sets mode: +v lhcheng | 00:34 | |
| *** EinstCrazy has joined #openstack-keystone | 00:34 | |
| *** EinstCrazy has quit IRC | 00:34 | |
| *** EinstCrazy has joined #openstack-keystone | 00:35 | |
| *** BjoernT has joined #openstack-keystone | 00:36 | |
| *** EinstCra_ has joined #openstack-keystone | 00:36 | |
| *** markvoelker has joined #openstack-keystone | 00:38 | |
| *** EinstCrazy has quit IRC | 00:40 | |
| *** markvoelker has quit IRC | 00:43 | |
| *** tqtran has quit IRC | 00:44 | |
| *** sheel has joined #openstack-keystone | 00:44 | |
| *** roxanagh_ has joined #openstack-keystone | 00:46 | |
| *** roxanaghe has quit IRC | 00:47 | |
| *** mylu has quit IRC | 00:48 | |
| *** mylu has joined #openstack-keystone | 00:52 | |
| *** lhcheng has quit IRC | 00:53 | |
| *** sdake_ has joined #openstack-keystone | 01:16 | |
| *** BjoernT has quit IRC | 01:17 | |
| *** sdake has quit IRC | 01:19 | |
| *** timonwong has joined #openstack-keystone | 01:23 | |
| *** roxanaghe has joined #openstack-keystone | 01:33 | |
| *** roxanagh_ has quit IRC | 01:34 | |
| *** sdake_ is now known as sdake | 01:34 | |
| *** EinstCra_ has quit IRC | 01:35 | |
| *** EinstCrazy has joined #openstack-keystone | 01:36 | |
| *** mylu has quit IRC | 01:37 | |
| *** EinstCra_ has joined #openstack-keystone | 01:38 | |
| *** tqtran has joined #openstack-keystone | 01:40 | |
| *** EinstCrazy has quit IRC | 01:40 | |
| *** roxanaghe has quit IRC | 01:41 | |
| *** Don_Nalezyty has quit IRC | 01:42 | |
| *** jdennis has quit IRC | 01:42 | |
| *** jdennis has joined #openstack-keystone | 01:42 | |
| *** tqtran has quit IRC | 01:44 | |
| *** timonwong has quit IRC | 01:45 | |
| *** timonwong has joined #openstack-keystone | 01:47 | |
| *** mylu has joined #openstack-keystone | 01:48 | |
| *** lhcheng has joined #openstack-keystone | 01:49 | |
| *** ChanServ sets mode: +v lhcheng | 01:49 | |
| *** mylu has quit IRC | 01:50 | |
| *** mylu has joined #openstack-keystone | 01:53 | |
| *** timonwong has quit IRC | 01:56 | |
| *** timonwong has joined #openstack-keystone | 01:58 | |
| *** timonwong has quit IRC | 01:59 | |
| *** timonwong has joined #openstack-keystone | 02:03 | |
| *** timonwon_ has joined #openstack-keystone | 02:06 | |
| *** timonwong has quit IRC | 02:06 | |
| *** amit213 has quit IRC | 02:09 | |
| *** amit213 has joined #openstack-keystone | 02:09 | |
| *** lhcheng has quit IRC | 02:21 | |
| *** amit213 has quit IRC | 02:25 | |
| *** mylu has quit IRC | 02:26 | |
| *** amit213 has joined #openstack-keystone | 02:26 | |
| *** mylu has joined #openstack-keystone | 02:26 | |
| *** knikolla has joined #openstack-keystone | 02:27 | |
| *** sdake has quit IRC | 02:27 | |
| *** baffle has quit IRC | 02:27 | |
| *** amit213 has quit IRC | 02:28 | |
| *** amit2131 has joined #openstack-keystone | 02:28 | |
| *** sdake has joined #openstack-keystone | 02:29 | |
| *** stingaci has joined #openstack-keystone | 02:30 | |
| *** mylu has quit IRC | 02:32 | |
| *** mylu has joined #openstack-keystone | 02:34 | |
| *** sdake_ has joined #openstack-keystone | 02:35 | |
| *** sdake has quit IRC | 02:35 | |
| *** markvoelker has joined #openstack-keystone | 02:39 | |
| *** mylu has quit IRC | 02:39 | |
| *** amit213 has joined #openstack-keystone | 02:39 | |
| *** baffle has joined #openstack-keystone | 02:43 | |
| *** markvoelker has quit IRC | 02:43 | |
| *** amit213 has quit IRC | 02:44 | |
| *** amit213 has joined #openstack-keystone | 02:45 | |
| *** sdake_ is now known as sdake | 02:45 | |
| *** mylu has joined #openstack-keystone | 02:47 | |
| *** amit213 has quit IRC | 02:51 | |
| *** amit213 has joined #openstack-keystone | 02:52 | |
| *** amit213 has quit IRC | 03:00 | |
| *** amit213 has joined #openstack-keystone | 03:00 | |
| *** sheel has quit IRC | 03:05 | |
| *** markvoelker has joined #openstack-keystone | 03:17 | |
| openstackgerrit | ayoung proposed openstack/keystone: Make all fixture project_ids into uuids https://review.openstack.org/306681 | 03:21 |
|---|---|---|
| *** EinstCra_ has quit IRC | 03:24 | |
| *** EinstCrazy has joined #openstack-keystone | 03:25 | |
| *** EinstCrazy has quit IRC | 03:29 | |
| *** EinstCrazy has joined #openstack-keystone | 03:29 | |
| *** lhcheng has joined #openstack-keystone | 03:31 | |
| *** ChanServ sets mode: +v lhcheng | 03:31 | |
| *** lhcheng_ has joined #openstack-keystone | 03:32 | |
| *** EinstCrazy has quit IRC | 03:33 | |
| *** lhcheng has quit IRC | 03:36 | |
| *** mylu has quit IRC | 03:40 | |
| *** timonwon_ has quit IRC | 03:42 | |
| *** lamt has quit IRC | 03:42 | |
| *** richm has quit IRC | 03:43 | |
| openstackgerrit | ayoung proposed openstack/keystone: Make fernet support trust auth against v2.0 https://review.openstack.org/278693 | 03:49 |
| openstackgerrit | ayoung proposed openstack/keystone: [WIP]Make fernet default token provider https://review.openstack.org/258650 | 03:49 |
| *** dave-mcc_ has quit IRC | 03:55 | |
| *** mylu has joined #openstack-keystone | 04:09 | |
| *** EinstCrazy has joined #openstack-keystone | 04:12 | |
| *** EinstCrazy has quit IRC | 04:12 | |
| *** EinstCrazy has joined #openstack-keystone | 04:12 | |
| *** stingaci has quit IRC | 04:21 | |
| *** tqtran has joined #openstack-keystone | 04:37 | |
| *** timonwong has joined #openstack-keystone | 05:22 | |
| *** sheel has joined #openstack-keystone | 05:35 | |
| *** mylu has quit IRC | 05:49 | |
| *** EinstCrazy has quit IRC | 06:15 | |
| *** lhcheng_ has quit IRC | 06:17 | |
| *** EinstCrazy has joined #openstack-keystone | 06:32 | |
| *** EinstCrazy has quit IRC | 06:37 | |
| *** EinstCrazy has joined #openstack-keystone | 06:41 | |
| *** zqfan has joined #openstack-keystone | 06:41 | |
| *** markvoelker has quit IRC | 06:44 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/oslo.policy: Imported Translations from Zanata https://review.openstack.org/306699 | 06:46 |
| *** EinstCrazy has quit IRC | 06:51 | |
| *** EinstCrazy has joined #openstack-keystone | 06:51 | |
| *** EinstCrazy has quit IRC | 06:55 | |
| *** sdake_ has joined #openstack-keystone | 06:56 | |
| *** EinstCrazy has joined #openstack-keystone | 06:56 | |
| *** EinstCrazy has quit IRC | 06:57 | |
| *** sdake has quit IRC | 06:58 | |
| *** timonwong has quit IRC | 07:01 | |
| *** EinstCrazy has joined #openstack-keystone | 07:03 | |
| *** EinstCrazy has quit IRC | 07:03 | |
| *** EinstCrazy has joined #openstack-keystone | 07:04 | |
| *** EinstCrazy has quit IRC | 07:09 | |
| *** EinstCrazy has joined #openstack-keystone | 07:10 | |
| *** EinstCrazy has quit IRC | 07:10 | |
| *** EinstCrazy has joined #openstack-keystone | 07:11 | |
| *** mvk has quit IRC | 07:31 | |
| *** markvoelker has joined #openstack-keystone | 07:45 | |
| *** markvoelker has quit IRC | 07:50 | |
| *** tqtran has quit IRC | 07:51 | |
| *** rcernin has joined #openstack-keystone | 08:00 | |
| *** rcernin has quit IRC | 08:13 | |
| *** sdake_ has quit IRC | 08:18 | |
| *** sdake has joined #openstack-keystone | 08:18 | |
| *** sheel has quit IRC | 08:35 | |
| *** rcernin has joined #openstack-keystone | 08:40 | |
| *** rcernin has quit IRC | 08:49 | |
| *** sdake_ has joined #openstack-keystone | 08:50 | |
| *** sdake__ has joined #openstack-keystone | 08:53 | |
| *** sdake has quit IRC | 08:54 | |
| *** sdake_ has quit IRC | 08:56 | |
| *** EinstCrazy has quit IRC | 09:04 | |
| *** EinstCrazy has joined #openstack-keystone | 09:08 | |
| *** agireud has quit IRC | 09:11 | |
| *** agireud has joined #openstack-keystone | 09:13 | |
| openstackgerrit | yolanda.robla proposed openstack/keystoneauth: Create custom serializer for keystoneauth and betamax https://review.openstack.org/305937 | 09:16 |
| *** sdake__ has quit IRC | 09:16 | |
| *** EinstCrazy has quit IRC | 09:17 | |
| *** EinstCrazy has joined #openstack-keystone | 09:33 | |
| *** ekarlso has quit IRC | 09:38 | |
| *** e0ne has joined #openstack-keystone | 09:50 | |
| *** e0ne has quit IRC | 09:52 | |
| *** e0ne has joined #openstack-keystone | 10:08 | |
| *** e0ne has quit IRC | 10:08 | |
| *** EinstCrazy has quit IRC | 10:32 | |
| *** lhcheng has joined #openstack-keystone | 10:35 | |
| *** ChanServ sets mode: +v lhcheng | 10:35 | |
| *** lhcheng has quit IRC | 10:40 | |
| *** stingaci has joined #openstack-keystone | 11:37 | |
| *** markvoelker has joined #openstack-keystone | 11:46 | |
| *** ekarlso has joined #openstack-keystone | 11:50 | |
| *** markvoelker has quit IRC | 11:51 | |
| *** henrynash has quit IRC | 12:11 | |
| *** alex_xu has quit IRC | 12:40 | |
| *** alex_xu has joined #openstack-keystone | 12:42 | |
| *** stingaci has quit IRC | 13:01 | |
| *** daemontool has joined #openstack-keystone | 13:02 | |
| *** alex_xu has quit IRC | 13:11 | |
| *** alex_xu has joined #openstack-keystone | 13:14 | |
| openstackgerrit | Merged openstack/oslo.policy: Imported Translations from Zanata https://review.openstack.org/306699 | 13:25 |
| *** sheel has joined #openstack-keystone | 13:26 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystone: Updated from global requirements https://review.openstack.org/305187 | 13:26 |
| *** real56 has joined #openstack-keystone | 13:27 | |
| *** e0ne has joined #openstack-keystone | 13:29 | |
| *** dave-mccowan has joined #openstack-keystone | 13:30 | |
| *** daemontool has quit IRC | 13:33 | |
| *** e0ne has quit IRC | 13:36 | |
| *** fhubik has joined #openstack-keystone | 13:39 | |
| *** orzel has joined #openstack-keystone | 13:40 | |
| orzel | hello. With latest version (mitaka, 9.0.0), it seems to be needed to use a wsgi server. Great move. But it seems that only apache/uwsgi is supported. The main problems being that 1) there's no easy way to the the wsgi "application" 2) the /usr/bin/keystone-wsgi-* contain hyphens, which is a very bad practice in python | 13:41 |
| orzel | and basically, prevent using gunicorn, for exemple | 13:42 |
| orzel | i can't find any information on the web on using gunicorn + keystone | 13:42 |
| *** e0ne has joined #openstack-keystone | 13:44 | |
| orzel | well, actually i'm not even sure uwsgi is supported, documentation only mentions apache mod_wsgi | 13:47 |
| *** e0ne has quit IRC | 13:49 | |
| *** daemontool has joined #openstack-keystone | 13:50 | |
| *** dave-mccowan has quit IRC | 13:51 | |
| *** sdake has joined #openstack-keystone | 14:02 | |
| *** markvoelker has joined #openstack-keystone | 14:07 | |
| *** sdake_ has joined #openstack-keystone | 14:13 | |
| *** henrynash has joined #openstack-keystone | 14:16 | |
| *** ChanServ sets mode: +v henrynash | 14:16 | |
| *** sdake has quit IRC | 14:16 | |
| *** mylu has joined #openstack-keystone | 14:17 | |
| *** sdake_ has quit IRC | 14:27 | |
| *** jdennis1 has joined #openstack-keystone | 14:28 | |
| *** jdennis has quit IRC | 14:30 | |
| *** toddnni has quit IRC | 14:40 | |
| *** toddnni has joined #openstack-keystone | 14:40 | |
| *** sdake has joined #openstack-keystone | 14:59 | |
| *** chlong has joined #openstack-keystone | 15:00 | |
| *** mylu has quit IRC | 15:05 | |
| *** e0ne has joined #openstack-keystone | 15:20 | |
| *** real56 has quit IRC | 15:21 | |
| *** mylu has joined #openstack-keystone | 15:26 | |
| *** e0ne has quit IRC | 15:28 | |
| *** e0ne has joined #openstack-keystone | 15:32 | |
| *** e0ne has quit IRC | 15:40 | |
| *** e0ne has joined #openstack-keystone | 15:46 | |
| *** e0ne has quit IRC | 15:52 | |
| rodrigods | mylu, there? any success with k2k tests? | 16:06 |
| mylu | rodrigods: not really, I haven't much time to try more things with it yet the plan is to play with it this weekend haha | 16:07 |
| rodrigods | mylu, ok | 16:09 |
| *** mylu has quit IRC | 16:10 | |
| *** mylu has joined #openstack-keystone | 16:11 | |
| *** rcernin has joined #openstack-keystone | 16:25 | |
| *** mylu has quit IRC | 16:25 | |
| *** mylu has joined #openstack-keystone | 16:28 | |
| *** rcernin has quit IRC | 16:46 | |
| *** mylu has quit IRC | 16:51 | |
| *** mylu has joined #openstack-keystone | 16:51 | |
| *** mylu has quit IRC | 16:54 | |
| *** fhubik has quit IRC | 17:01 | |
| *** e0ne has joined #openstack-keystone | 17:07 | |
| *** e0ne has quit IRC | 17:19 | |
| *** e0ne has joined #openstack-keystone | 17:23 | |
| *** e0ne has quit IRC | 17:26 | |
| *** raddaoui has quit IRC | 17:37 | |
| *** mylu has joined #openstack-keystone | 17:48 | |
| *** mylu_ has joined #openstack-keystone | 17:54 | |
| *** mylu has quit IRC | 17:54 | |
| *** jdennis has joined #openstack-keystone | 17:56 | |
| *** mtreinish has quit IRC | 17:56 | |
| *** mtreinish has joined #openstack-keystone | 17:56 | |
| *** jdennis1 has quit IRC | 17:57 | |
| morgan | orzel: uwsgi is supported | 18:21 |
| morgan | orzel: we have a gate job for it, and it works better (in general) than eventlet did | 18:21 |
| morgan | orzel: gunicorn is not currently tested, we would be happy to have changes proposed to make it work. i am also disappointed that gunicorn can't handle '-' (uwsgi is python, cython specifically, based and can handle the dashes) via cli arguments | 18:23 |
| morgan | orzel: the wsgi files are meant to be file-system not nessicarily consumed within python as non-strings. | 18:23 |
| *** edtubill has joined #openstack-keystone | 18:30 | |
| morgan | orzel: i think i have a way for you to use gunicorn | 18:33 |
| morgan | orzel: gunicorn keystone.server.wsgi:initialize_admin_application | 18:35 |
| morgan | orzel: the "wsgi" files created basically just reference that entry point and are created for mod_wsgi etc. | 18:36 |
| *** edtubill has quit IRC | 19:08 | |
| *** roxanaghe has joined #openstack-keystone | 19:11 | |
| *** roxanaghe has quit IRC | 19:16 | |
| *** roxanaghe has joined #openstack-keystone | 19:18 | |
| *** roxanaghe has quit IRC | 19:23 | |
| *** zqfan has quit IRC | 19:42 | |
| openstackgerrit | Merged openstack/python-keystoneclient: Updated from global requirements https://review.openstack.org/303156 | 19:53 |
| openstackgerrit | Merged openstack/keystoneauth: Fix documentation example again https://review.openstack.org/306570 | 20:09 |
| orzel | morgan: er.. i know uwsgi is supported. i even said "it seems that only apache/uwsgi is supported"... ?? | 20:38 |
| morgan | orzel: wasgoing on "orzel | well, actually i'm not even sure uwsgi is supported, documentation only mentions apache mod_wsgi" | 20:38 |
| orzel | morgan: i tried creating my own wsgi file, but initialize_admin_application() doesn't return a wsgi application it seems | 20:38 |
| orzel | morgan: ah, ok, indeed | 20:39 |
| morgan | orzel: the wsgi files we create are very specifically for the wsgi runnners that can't load directly like gunicorn | 20:39 |
| orzel | morgan: let me try nonetheless | 20:39 |
| morgan | orzel: and those also support wsgiref for debug purposes | 20:41 |
| morgan | since the code that is run for loading via mod_wsgi etc is http://paste.openstack.org/show/494317/ it would be easy to load that directly | 20:41 |
| morgan | orzel: it *seemed* to work with the ^^ above gunicorn example (i didn't dig too deep because in the middle of something else) | 20:42 |
| morgan | orzel: it *should* return a wsgi application | 20:43 |
| morgan | since that is how it's called. | 20:43 |
| orzel | morgan: ok, let me do the test. It kinda seem to work from cli, but not from supervisord script, probably a problem of my own, let me check that | 20:44 |
| morgan | ah i haven't tried at all with suporvisord, that could be stomping on something important | 20:44 |
| orzel | i've seen the code you've posted, and i had tried to use it, but without success so far | 20:44 |
| *** sheel has quit IRC | 20:45 | |
| morgan | unfortunately, for cases like yours, if it isn't tested (especially in openstack) it is typically broken :( | 20:45 |
| morgan | i know we had weird issues with gunicorn a while ago | 20:45 |
| openstackgerrit | Steve Martinelli proposed openstack/keystone: Deprecation reason for domain_id_immutable https://review.openstack.org/306021 | 20:45 |
| morgan | stevemar: oh hai | 20:46 |
| stevemar | morgan: howdy partner :) | 20:47 |
| stevemar | also hi to lin who is somewhere in the ether | 20:47 |
| morgan | stevemar: and doesn't seem to be on IRC | 20:47 |
| stevemar | morgan: yep | 20:47 |
| orzel | morgan: now i recall, it starts successfully, but it fails on first hit with TypeError: initialize_admin_application() takes no arguments (2 given). You need to have a script calling it, storing the result in a variable and use this variable ... | 20:48 |
| orzel | I've never seen that, and i have installed/created quite a lot of (django mainly) wsgi applications | 20:49 |
| morgan | orzel: ah. you need a wrapper then. | 20:49 |
| morgan | oh right | 20:49 |
| morgan | cause initialize call jsut builds the admin or main app | 20:49 |
| morgan | it is because we have some "pre-application" code | 20:49 |
| morgan | so you'd need a simple file that does exactly waht is in that paste ^ | 20:50 |
| morgan | or similar | 20:50 |
| *** lamt has joined #openstack-keystone | 20:50 | |
| morgan | orzel: you could also use __import__(<file name>) if you wanted to use the version PBR ships | 20:51 |
| morgan | but our wsgi-files are built to be more just executables which ends up being a little odd. | 20:52 |
| morgan | orzel: now, i can guarnatee you'd have at least a 2 cycle warning if the initialize_admin_application/initialize_public_application/initialize_application changes | 20:53 |
| morgan | orzel: so it should be safe to assume those are there for you to hook into for your own wsgi file for gunicorn | 20:53 |
| morgan | "app file" | 20:53 |
| morgan | stevemar: are we done with eventlet yet? | 20:54 |
| morgan | stevemar: or do i need to chase down the thing dstanek was working on? | 20:54 |
| orzel | ok, the initialize_admin_application() returns a paste.urlmap.URLMap | 20:54 |
| morgan | so our tests are also eventlet-free | 20:54 |
| morgan | orzel: yeah. | 20:55 |
| orzel | this is why it fails and keeps rebooting | 20:55 |
| morgan | orzel: that should work? | 20:55 |
| morgan | if not, then gunicorn is expecting something different than mod_wsgi/uwsgi | 20:55 |
| orzel | i guess gunicorn expects something like a wsgi application | 20:56 |
| morgan | (and either mod_wsgi/uwsgi are sloppy or gunicorn is not adhering to the spec) | 20:56 |
| morgan | it could go either way | 20:56 |
| orzel | django.core.handlers.wsgi.WSGIHandler is what django returns for an "application" | 20:57 |
| orzel | something that has an API with "get_response()" and such | 20:57 |
| morgan | right, but this is python - so if it "looks" like a WSGIHandler it is a wsgi handler | 20:57 |
| morgan | it also might be gunicorn is very narrow | 20:57 |
| orzel | er... the URLMap doesn't have get_response() for example | 20:58 |
| orzel | that doesn't look like 'narrow' to fail on it | 20:58 |
| orzel | (it it?) | 20:58 |
| morgan | it might be. i;'m looking now | 20:58 |
| orzel | it's just a map | 20:58 |
| orzel | [((None, '/v2.0'), <oslo_middleware.cors.CORS object at 0x7fe595e71290>), ((None, '/v3'), <oslo_middleware.cors.CORS object at 0x7fe595c75390>), ((None, ''), <oslo_middleware.cors.CORS obje | 20:59 |
| orzel | ct at 0x7fe595bd3790>)] | 20:59 |
| orzel | not sure there's much besides this | 20:59 |
| *** lhcheng has joined #openstack-keystone | 21:00 | |
| *** ChanServ sets mode: +v lhcheng | 21:00 | |
| orzel | i'm far from a wsgi specialist, but that seems radically different from all wsgi 'applications' i've seen so far | 21:00 |
| *** lhcheng has quit IRC | 21:00 | |
| morgan | ahhhhh | 21:03 |
| morgan | i see the issue | 21:03 |
| * orzel wonders ! | 21:03 | |
| morgan | at least *an* issue | 21:04 |
| orzel | that's still something :) | 21:04 |
| morgan | gunicorn passes on cli arguments to the wsgi application | 21:04 |
| morgan | which keystone explicitly rejects | 21:04 |
| morgan | gunicorn i'm getting: | 21:04 |
| morgan | http://paste.openstack.org/show/494319/ | 21:05 |
| morgan | basicallt oslo.config is stomping on gunicorn and vice versa | 21:05 |
| morgan | it looks like keystone cannot support gunicorn in this case without work. | 21:05 |
| morgan | this is on master btw. | 21:06 |
| morgan | orzel: =( | 21:06 |
| stevemar | morgan: we're not done with eventlet yet, i haven't bothered to rebase my patch | 21:07 |
| stevemar | was going to do that today | 21:07 |
| morgan | stevemar: ++ | 21:07 |
| stevemar | morgan: it's still going to be... icky | 21:08 |
| morgan | orzel: sorry. i am not sure how we can best resolve this. | 21:10 |
| morgan | orzel: we might need to ignore unknown options on the "cli" for keystone. | 21:11 |
| morgan | but it might still stomp all over gunicorn's option handler | 21:11 |
| orzel | morgan: er, not sure to get it all. this is keystone saying "unrecognized arguments" ? | 21:11 |
| morgan | orzel: basically yes. keystone's option parser is stomping on gunicorn's | 21:11 |
| morgan | and is seeing gunicron's cli options | 21:12 |
| morgan | which are "unrecognized" | 21:12 |
| orzel | cli and arguments don't look very wsgi to me, actually | 21:12 |
| morgan | these are hold overs from eventlet. | 21:12 |
| morgan | we have had to support both CLI and non-CLI forms of option passing | 21:12 |
| morgan | so in short, gunicorn definitely not supportable right now, it might be fixable with some work to isolate the cli option passing, it might be a lot easier once eventlet code is removed | 21:13 |
| morgan | it is also related to the wsgiref support. | 21:14 |
| orzel | i'm suprised gunicorn passes the whole args to the worker, i'll double check this | 21:14 |
| morgan | orzel: i think it's because argparse is being used in both places | 21:15 |
| morgan | so we end up with conflicting arg parsers. | 21:15 |
| morgan | this may also mean that gunicorn is pretty unsupportable across all of openstack (even where wsgi app, non-eventlet | 21:16 |
| morgan | works (except horizon, since it's django) | 21:16 |
| orzel | it works with horizon, i do use gunicorn for it | 21:17 |
| morgan | yeah horizon being the special case since it's really just a django app | 21:17 |
| orzel | i would be very surprised that keystone is really wsgi "compliant" with such a behaviour. But i wont have time to check this :( | 21:17 |
| morgan | this looks like gunicorn specific incompatibility | 21:18 |
| morgan | if we were truely not wsgi compliant i doubt mod_wsgi and uwsgi would work | 21:18 |
| morgan | neither of those pass on cli arguments | 21:18 |
| orzel | morgan: ok, but then, this is a very different problem of application being a wsgi handler in one case (django) and a dictionnary / url / mapping stuff with keystone ? Or is it related ? | 21:19 |
| morgan | i am not sure what you're hitting i can't even get that far | 21:19 |
| morgan | i'm guessing you are pulling in something else not quite the same way | 21:20 |
| orzel | i'm just "from myapp import application" in the module where i've put the application script | 21:20 |
| orzel | script being import + application = wsgi_server.initialize_application("admin") | 21:20 |
| orzel | seems to work, i got my usual keystone warnings about deprecated options | 21:20 |
| orzel | but i got a map, not a wsgi handler | 21:20 |
| morgan | so you are adding in a second layer? | 21:21 |
| morgan | so module:application then another wsgi file explicitly for importing application? | 21:21 |
| orzel | just a cli is easire to do things like type(application) or application.<tab> | 21:21 |
| orzel | nono, i'm doing the import from 'bpython' | 21:22 |
| orzel | not another script | 21:22 |
| morgan | so you aren't doing: gunicorn -w 4 myapp:app | 21:22 |
| morgan | basically? | 21:22 |
| orzel | if i do that it keeps crashing/restarting | 21:22 |
| orzel | i'm trying to debug this | 21:22 |
| morgan | toss up a gist/paste of what you're doing | 21:23 |
| orzel | ( i can see crashes/restart from log, but nothing more) | 21:23 |
| morgan | and i can try and follow along | 21:23 |
| orzel | you right | 21:23 |
| morgan | but i admit i had bad experiences making gunicorn work, so i gravitated to uwsgi | 21:23 |
| orzel | http://pastebin.com/ZEyMzyZj | 21:24 |
| morgan | never saw bpython until today | 21:24 |
| morgan | fwiw ;) | 21:24 |
| orzel | like ipython but better | 21:25 |
| orzel | ipython like normal python CLI, but better :) | 21:25 |
| morgan | i only use cpython interpreters | 21:25 |
| morgan | ok let me see what i can do with your pastebin | 21:25 |
| orzel | it gets you colors and, most importantly, auto-completion | 21:25 |
| orzel | i feel naked in a cpython cli now :/ | 21:25 |
| orzel | like using vi(not vim) in black&white on some old sys-V | 21:26 |
| morgan | orzel: lol | 21:26 |
| *** henrynash has quit IRC | 21:32 | |
| orzel | er... double-crash in my kernel. Playing with evenlet/gunicorn on some containerized env doesn't play nice | 21:35 |
| morgan | ftr i dislike paste. | 21:41 |
| morgan | running into some odd issues getting the paste_deploy file registered in my environment | 21:42 |
| orzel | ok, it almost work | 21:43 |
| morgan | ugh. | 21:45 |
| morgan | oslo.config doesn't search in a logical place for my configs... grumble. | 21:46 |
| morgan | orzel: so, paste.urlmap.URLMap is a callable | 21:48 |
| orzel | i got problems with config too, now | 21:48 |
| orzel | File "/usr/lib64/python2.7/site-packages/keystone/common/config.py", line 1176, in find_paste_config | 21:48 |
| orzel | raise exception.ConfigFileNotFound(config_file=paste_config_value) | 21:48 |
| orzel | not sure what it means | 21:48 |
| morgan | orzel: that means URLMap can be called as an application | 21:49 |
| orzel | yes, apparently | 21:49 |
| morgan | since it has a __call__ method | 21:49 |
| morgan | the find paste_config is what i just worked around. | 21:50 |
| morgan | basically i just put my config files in ~/ | 21:50 |
| morgan | because i am not putting them in /etc/ | 21:50 |
| morgan | on this machine | 21:50 |
| morgan | it could also (in this case) be ~/keystone | 21:50 |
| morgan | or /etc/keystone | 21:50 |
| *** mylu_ has quit IRC | 21:51 | |
| morgan | orzel: Heh https://review.openstack.org/#/c/288216/10 we have a bug specifically for making this easier | 21:53 |
| patchbot | morgan: patch 288216 - keystone - Customize config file location when run as wsgi app. | 21:53 |
| morgan | orzel: anyway, so, this still seems like gunicorn passing the arguments down because of how argparse works | 21:54 |
| morgan | orzel: and it means gunicorn wont be usable with keystone as of today | 21:54 |
| morgan | stevemar: ^ i think we need to see how CLI args are passed and handle this in a more sane way. also the patch 288216 needs eyes | 21:55 |
| patchbot | morgan: https://review.openstack.org/#/c/288216/ - keystone - Customize config file location when run as wsgi app. | 21:55 |
| *** jasonsb has joined #openstack-keystone | 22:08 | |
| morgan | stevemar: want me to rebase your eventlet delete patch? | 22:12 |
| morgan | stevemar: or you got it? /me would like to see that land soon. | 22:12 |
| *** roxanaghe has joined #openstack-keystone | 22:15 | |
| stevemar | morgan: yo, let me try that now | 22:30 |
| * morgan nods. | 22:30 | |
| stevemar | morgan: was running around all morning and afternoon | 22:30 |
| morgan | phsaw. running | 22:30 |
| stevemar | morgan: finally got home and ate, watching the jays lose now :( | 22:31 |
| morgan | hehe. | 22:31 |
| stevemar | morgan: it was more driving around, but in traffic | 22:31 |
| stevemar | and going to 5 different locations :P | 22:31 |
| morgan | so: could have walked there faster? :P | 22:31 |
| stevemar | nah | 22:31 |
| *** david-lyle has quit IRC | 22:39 | |
| *** david-lyle has joined #openstack-keystone | 22:39 | |
| openstackgerrit | Steve Martinelli proposed openstack/keystone: Remove eventlet support https://review.openstack.org/249486 | 22:42 |
| stevemar | morgan: le rebase ^ | 22:42 |
| stevemar | blah, the release notes need to be fixed | 22:43 |
| morgan | did you cover bknudson's comments? | 22:43 |
| stevemar | morgan: no, strictly a rebase | 22:43 |
| morgan | ok | 22:43 |
| morgan | will hold for the update then to cover reno/comments | 22:44 |
| morgan | it looks reasonable barring those things. | 22:44 |
| openstackgerrit | Merged openstack/keystone: Deprecation reason for domain_id_immutable https://review.openstack.org/306021 | 22:50 |
| orzel | morgan: i've removed the rags (sys.argv=sys.argv[:1]). Not it runs, but i have those config stuff errors. Does keystone _requires_ some cli arg ? I can't find any in the apache/keystone official documentation | 22:52 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystone: Updating sample configuration file https://review.openstack.org/306785 | 22:52 |
| morgan | orzel: no we dont require it for wsgi, we do require it for eventlet which is not gone yet | 22:53 |
| morgan | and we have some cases where we should allow it for wsgiref | 22:53 |
| morgan | (standalone mode) | 22:53 |
| *** john5223_ has joined #openstack-keystone | 22:57 | |
| orzel | morgan: how can i debug those then ? http://pastebin.com/fkZ6fH3r | 23:01 |
| orzel | i dont know where to start | 23:01 |
| *** sdake has quit IRC | 23:14 | |
| *** roxanaghe has quit IRC | 23:19 | |
| *** mylu has joined #openstack-keystone | 23:40 | |
| orzel | I've narrowed down the problem to CONF.find_file(paste_config) with paste_config="keystone-paste.ini", and this returns None | 23:53 |
| orzel | it seems to have found keystone.conf though ... | 23:53 |
| *** mylu has quit IRC | 23:53 | |
| *** roxanaghe has joined #openstack-keystone | 23:55 | |
| orzel | ah, i'm not alone. It's know issue https://bugzilla.redhat.com/show_bug.cgi?id=1174795 | 23:56 |
| openstack | bugzilla.redhat.com bug 1174795 in openstack-selinux "keystone fails to start: raise exception.ConfigFileNotFound(config_file=paste_config_value)" [Unspecified,New] - Assigned to lhh | 23:56 |
| *** roxanaghe has quit IRC | 23:56 | |
| orzel | old, unfixed, the env is different, but the stack do look very similar | 23:56 |
| *** roxanaghe has joined #openstack-keystone | 23:56 | |
| *** mylu has joined #openstack-keystone | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!