| *** masber has joined #openstack-keystone | 00:03 | |
| *** masber has quit IRC | 00:03 | |
| *** masber has joined #openstack-keystone | 00:03 | |
| *** masber has quit IRC | 00:04 | |
| *** masber has joined #openstack-keystone | 00:04 | |
| *** masber has quit IRC | 00:04 | |
| *** masber has joined #openstack-keystone | 00:06 | |
| *** masber has quit IRC | 00:06 | |
| *** masber has joined #openstack-keystone | 00:08 | |
| *** germs has joined #openstack-keystone | 00:37 | |
| *** germs has quit IRC | 00:37 | |
| *** germs has joined #openstack-keystone | 00:37 | |
| *** felipemonteiro__ has joined #openstack-keystone | 00:41 | |
| *** germs has quit IRC | 00:41 | |
| *** felipemonteiro__ has quit IRC | 00:47 | |
| *** edmondsw has joined #openstack-keystone | 01:07 | |
| *** wxy has joined #openstack-keystone | 01:10 | |
| *** edmondsw has quit IRC | 01:11 | |
| *** edmondsw has joined #openstack-keystone | 01:22 | |
| *** edmondsw has quit IRC | 01:34 | |
| *** edmondsw has joined #openstack-keystone | 01:35 | |
| *** edmondsw has quit IRC | 01:37 | |
| *** edmondsw has joined #openstack-keystone | 01:37 | |
| *** edmondsw has quit IRC | 01:38 | |
| *** eschwartz is now known as o_O | 01:41 | |
| *** o_O is now known as Guest44118 | 01:41 | |
| *** Guest44118 is now known as eschwartz | 01:42 | |
| adriant | lbragstad, cmurphy: I'm just working on the auth receipt code and realised that we've still got references to uuid tokens in places: | 02:13 |
|---|---|---|
| adriant | https://github.com/openstack/keystone/blob/master/keystone/conf/token.py#L68 | 02:13 |
| adriant | Not that it's even remotely urgent, but it's probably something I can look at cleaning up if no one has grabbed that work. | 02:13 |
| cmurphy | adriant: we'll take any help we can get with cleanup work :) | 02:19 |
| adriant | cmurphy: cool, after this patch is at least in a mostly ready for review stage I'll do some extra cleanup stuff. :) | 02:20 |
| adriant | the auth receipt code is actually not looking as terrifying as I originally expected it to be :) | 02:20 |
| *** xinran__ has joined #openstack-keystone | 02:30 | |
| wxy | adriant: this is maybe what you want https://review.openstack.org/#/c/567110/ | 02:33 |
| adriant | wxy: that's the one :) | 02:33 |
| adriant | that's why I asked, awesome will help review! | 02:34 |
| wxy | cool ;) | 02:34 |
| *** germs has joined #openstack-keystone | 02:38 | |
| *** germs has quit IRC | 02:42 | |
| *** threestrands has joined #openstack-keystone | 03:59 | |
| *** liuzz_ has joined #openstack-keystone | 03:59 | |
| *** threestrands_ has joined #openstack-keystone | 04:00 | |
| *** threestrands_ has quit IRC | 04:01 | |
| *** threestrands_ has joined #openstack-keystone | 04:02 | |
| *** liuzz has quit IRC | 04:02 | |
| *** threestrands has quit IRC | 04:04 | |
| *** kevinbenton has joined #openstack-keystone | 04:05 | |
| *** hoonetorg has quit IRC | 04:21 | |
| *** germs has joined #openstack-keystone | 04:39 | |
| *** hoonetorg has joined #openstack-keystone | 04:39 | |
| *** threestrands_ has quit IRC | 04:39 | |
| *** pooja_jadhav has joined #openstack-keystone | 04:40 | |
| *** links has joined #openstack-keystone | 04:41 | |
| *** germs has quit IRC | 04:44 | |
| *** xinran__ has quit IRC | 04:49 | |
| *** pcichy has joined #openstack-keystone | 04:54 | |
| *** threestrands has joined #openstack-keystone | 05:04 | |
| *** threestrands has quit IRC | 05:07 | |
| kmalloc | adriant: I figured the receipt code wouldn't be that bad, but that is why we iterated on the design before hand. | 05:24 |
| adriant | kmalloc: the most annoying part is just getting the provider logic down. I've pretty much duplicated a chunk from tokens and am stripping it of non-essential parts. | 05:29 |
| adriant | The auth controller code on the other hand is tiny | 05:29 |
| adriant | kmalloc: I should have a working WIP review up hopefully next week but without unit tests. | 05:31 |
| kmalloc | Cool. | 05:54 |
| *** belmoreira has joined #openstack-keystone | 06:08 | |
| *** pcaruana has joined #openstack-keystone | 06:31 | |
| *** annp has joined #openstack-keystone | 06:34 | |
| *** germs has joined #openstack-keystone | 06:40 | |
| *** germs has quit IRC | 06:40 | |
| *** germs has joined #openstack-keystone | 06:40 | |
| *** germs has quit IRC | 06:44 | |
| *** martinus__ has joined #openstack-keystone | 06:49 | |
| *** tesseract has joined #openstack-keystone | 07:10 | |
| *** rcernin has quit IRC | 07:39 | |
| *** AlexeyAbashkin has joined #openstack-keystone | 07:46 | |
| *** rpittau has joined #openstack-keystone | 07:53 | |
| *** kevinbenton has quit IRC | 07:56 | |
| *** annp has quit IRC | 07:57 | |
| *** annp has joined #openstack-keystone | 07:58 | |
| *** belmoreira has quit IRC | 08:01 | |
| *** belmoreira has joined #openstack-keystone | 08:04 | |
| *** alex_xu has quit IRC | 08:32 | |
| *** alex_xu has joined #openstack-keystone | 08:33 | |
| *** srihas has joined #openstack-keystone | 08:40 | |
| srihas | hi guys, I have just installed Openstack with JUJU. When I try to login from horizon, I am getting an error "Unable to establish connection to http://127.0.0.1:5000/v2.0/tokens: HTTPConnectionPool(host='127.0.0.1', port=5000): Max retries exceeded with url: /v2.0/tokens (Caused by NewConnectionError('<requests.packages.urllib3.connection.HTTPConnection object at 0x7f805b125f90>: Failed to establish a new connection: [Errno 111] Connection refu | 08:56 |
| srihas | local_settings.py has the OPENSTACK_HOST set to the IP of keystone though | 08:57 |
| srihas | can someone help? | 08:57 |
| *** annp has quit IRC | 09:04 | |
| *** annp has joined #openstack-keystone | 09:05 | |
| *** masber has quit IRC | 09:06 | |
| *** jaosorior has joined #openstack-keystone | 09:56 | |
| *** xinran__ has joined #openstack-keystone | 10:05 | |
| *** nicolasbock has joined #openstack-keystone | 10:29 | |
| *** edmondsw has joined #openstack-keystone | 11:03 | |
| *** links has quit IRC | 11:05 | |
| *** links has joined #openstack-keystone | 11:17 | |
| *** nicolasbock has quit IRC | 11:49 | |
| *** links has quit IRC | 11:52 | |
| *** links has joined #openstack-keystone | 11:52 | |
| *** raildo has joined #openstack-keystone | 12:02 | |
| *** gyankum has joined #openstack-keystone | 12:11 | |
| *** nicolasbock has joined #openstack-keystone | 12:12 | |
| *** doxa has joined #openstack-keystone | 12:20 | |
| doxa | good day | 12:21 |
| doxa | I am looking into using totp auth. When I use the https://docs.openstack.org/keystone/queens/advanced-topics/auth-totp.html info | 12:22 |
| doxa | I get error {"error": {"message": "The request you have made requires authentication.", "code": 401, "title": "Unauthorized"}} | 12:22 |
| doxa | any thoughts ? | 12:22 |
| *** xinran__ has quit IRC | 12:25 | |
| *** gyankum has quit IRC | 12:42 | |
| *** Shilpa has joined #openstack-keystone | 12:44 | |
| Shilpa | cmurphy: Hi | 12:47 |
| *** dklyle has quit IRC | 13:02 | |
| *** dklyle has joined #openstack-keystone | 13:03 | |
| *** felipemonteiro__ has joined #openstack-keystone | 13:04 | |
| *** dklyle has quit IRC | 13:06 | |
| *** felipemonteiro_ has joined #openstack-keystone | 13:11 | |
| *** dklyle has joined #openstack-keystone | 13:12 | |
| *** felipemonteiro__ has quit IRC | 13:15 | |
| *** mvk has quit IRC | 13:16 | |
| *** dklyle has quit IRC | 13:17 | |
| *** felipemonteiro__ has joined #openstack-keystone | 13:27 | |
| *** felipemonteiro_ has quit IRC | 13:27 | |
| *** superdan is now known as dansmith | 13:33 | |
| hrybacki | o/ | 13:33 |
| *** belmorei_ has joined #openstack-keystone | 13:33 | |
| *** belmoreira has quit IRC | 13:35 | |
| lbragstad | o/ | 13:35 |
| *** mvk has joined #openstack-keystone | 13:42 | |
| *** devx has joined #openstack-keystone | 13:55 | |
| *** devx has quit IRC | 13:56 | |
| *** jgwentworth is now known as melwitt | 13:57 | |
| *** devx has joined #openstack-keystone | 13:59 | |
| *** xinran__ has joined #openstack-keystone | 14:01 | |
| *** links has quit IRC | 14:03 | |
| *** r-daneel has joined #openstack-keystone | 14:05 | |
| *** spilla has joined #openstack-keystone | 14:09 | |
| kmalloc | o/ | 14:18 |
| gagehugo | o/ | 14:33 |
| *** gyankum has joined #openstack-keystone | 14:34 | |
| *** gyankum has quit IRC | 14:34 | |
| prometheanfire | I think keystone is the only project left that webob-1.8.1 breaks things on https://bugs.launchpad.net/openstack-requirements/+bug/1765748 http://logs.openstack.org/39/568039/1/check/cross-keystone-py27/0fb13ba/testr_results.html.gz | 14:36 |
| openstack | Launchpad bug 1765748 in OpenStack Global Requirements "webob-1.8.1 breaks projects" [High,In progress] - Assigned to Matthew Thode (prometheanfire) | 14:36 |
| openstackgerrit | Lance Bragstad proposed openstack/keystone master: Update tests to work with WebOb 1.8.1 https://review.openstack.org/568304 | 14:36 |
| lbragstad | prometheanfire: ^ | 14:36 |
| lbragstad | fast ping->bug fix ever! | 14:36 |
| lbragstad | fastest* | 14:36 |
| prometheanfire | lol | 14:38 |
| prometheanfire | ya | 14:38 |
| lbragstad | prometheanfire: i'm not sure if you've been noticing a specific pattern with projects that have been affected by this | 14:38 |
| lbragstad | but i just replaced our uuid usage with 'en' | 14:38 |
| * lbragstad shrugs | 14:38 | |
| lbragstad | since i don't suppose we're all that interested in testing how webob deals with that header, i just replaced it with something that passes the new regex | 14:39 |
| lbragstad | https://github.com/Pylons/webob/blob/master/src/webob/acceptparse.py#L3630-L3643 | 14:39 |
| prometheanfire | lbragstad: sure, all I know offhand is that it was a quick fix for them too | 14:40 |
| openstackgerrit | Lance Bragstad proposed openstack/keystone master: Update tests to work with WebOb 1.8.1 https://review.openstack.org/568304 | 14:44 |
| lbragstad | ok - updated so that it will be easier to ever tell if the mock stop works on a system with 'en' by default (which i assume would be wide range) | 14:45 |
| lbragstad | stops working* | 14:45 |
| *** belmorei_ has quit IRC | 14:54 | |
| *** pcaruana has quit IRC | 15:02 | |
| *** dklyle has joined #openstack-keystone | 15:14 | |
| *** dklyle has quit IRC | 15:22 | |
| *** rajalokan has joined #openstack-keystone | 15:22 | |
| *** panbalag has joined #openstack-keystone | 15:31 | |
| *** panbalag has left #openstack-keystone | 15:31 | |
| *** felipemonteiro__ has quit IRC | 15:34 | |
| kmalloc | lbragstad: hm. you know if henrynash been around recently? | 15:42 |
| kmalloc | lbragstad: had a question for him | 15:42 |
| *** kmalloc sets mode: -o kmalloc | 15:43 | |
| lbragstad | kmalloc: i have not seen him in some time | 15:44 |
| lbragstad | last i talked to him was in dublin | 15:44 |
| kmalloc | lbragstad: ok | 15:44 |
| *** nicolasbock has quit IRC | 15:50 | |
| *** ayoung has joined #openstack-keystone | 15:53 | |
| *** nicolasbock has joined #openstack-keystone | 15:53 | |
| kmalloc | lbragstad: replacing the header is fine | 16:02 |
| kmalloc | lbragstad: it's a silly test we're doing there... "does webob work" | 16:02 |
| kmalloc | ... if it doesn't... | 16:02 |
| kmalloc | why are we using it. | 16:02 |
| lbragstad | yeah | 16:03 |
| kmalloc | lbragstad: so... nit on structure for my current thing | 16:04 |
| kmalloc | lbragstad: keystone.flask or keystone.common.wsgi.flask or...? | 16:04 |
| kmalloc | lbragstad: any preferance? | 16:04 |
| * kmalloc leans towards keystone.flask | 16:05 | |
| kmalloc | but...... | 16:05 |
| *** rmascena has joined #openstack-keystone | 16:06 | |
| *** raildo has quit IRC | 16:08 | |
| lbragstad | why keystone.flask? | 16:08 |
| kmalloc | or keystone.server.flask | 16:09 |
| kmalloc | trying to avoid keystone.common dumping ground | 16:09 |
| lbragstad | gotcha | 16:10 |
| kmalloc | i'll use keystone.server.flask | 16:10 |
| lbragstad | yeah - that's fine | 16:10 |
| lbragstad | i'm not sure if i have a strong preference? | 16:10 |
| lbragstad | i feel like it should be in common, but at the same time we also have things like keystone.exception, keystone.notification, etc.. | 16:11 |
| kmalloc | right, we have wsgi initialization stuff in keystone.server | 16:11 |
| kmalloc | (paste deploy, etc) | 16:11 |
| kmalloc | so, i figure that is the right place to keep this stuff | 16:11 |
| *** gyee has joined #openstack-keystone | 16:13 | |
| lbragstad | sure - that works | 16:13 |
| *** pcaruana has joined #openstack-keystone | 16:20 | |
| *** dave-mccowan has joined #openstack-keystone | 16:23 | |
| *** dave-mccowan has quit IRC | 16:28 | |
| *** jmlowe has quit IRC | 16:29 | |
| *** dave-mccowan has joined #openstack-keystone | 16:32 | |
| *** felipemonteiro has joined #openstack-keystone | 16:32 | |
| lbragstad | stepping away to get a run in over lunch quick | 16:47 |
| *** AlexeyAbashkin has quit IRC | 16:57 | |
| *** mugsie has quit IRC | 17:06 | |
| *** mugsie has joined #openstack-keystone | 17:06 | |
| *** mugsie has quit IRC | 17:06 | |
| *** mugsie has joined #openstack-keystone | 17:06 | |
| *** mugsie has quit IRC | 17:08 | |
| *** tesseract has quit IRC | 17:11 | |
| *** mugsie has joined #openstack-keystone | 17:12 | |
| *** mugsie has quit IRC | 17:12 | |
| *** mugsie has joined #openstack-keystone | 17:12 | |
| *** xinran__ has quit IRC | 17:21 | |
| *** rajalokan has quit IRC | 17:31 | |
| kmalloc | lbragstad: man, converting to flask is a bunch of work. just getting out from under paste is weird. | 17:33 |
| *** germs has joined #openstack-keystone | 17:33 | |
| kmalloc | lbragstad: i think i have a pretty smooth path forward here. it's starting to shape up, and it looks like the biggest impact is just going to be convering how we do routes (from mapper/dispatch) to flask-specific [which, btw, is much better for validation of the URI elements] | 17:33 |
| lbragstad | ok - cool]\ | 17:36 |
| lbragstad | the mapper stuff has always been a little obtuse to me, but that's a personal opinion | 17:36 |
| lbragstad | curious to see what that looks like in flask | 17:37 |
| kmalloc | lbragstad: well, i'm mirroring it mostly right now, basically new application_factory loads in the routers and calls .add_routes() | 17:44 |
| kmalloc | in the future i'll convert these to flask blueprints [future patch] which will make it more explicit | 17:44 |
| kmalloc | but we need to address @protected etc | 17:44 |
| *** germs has quit IRC | 17:46 | |
| lbragstad | yeah | 17:53 |
| lbragstad | removing that decorator is going to make fixing https://bugs.launchpad.net/keystone/+bugs?field.tag=policy easier | 17:54 |
| lbragstad | too | 17:54 |
| *** jmlowe has joined #openstack-keystone | 17:56 | |
| kmalloc | flask likes to use the app.route() decorator for each item, but i don't think that works for our architecture (we don't want to initialize the app early) | 17:56 |
| kmalloc | not too bad though | 17:56 |
| kmalloc | honestly | 17:56 |
| kmalloc | the JSON_HOME bits are going to be the hardest part | 17:56 |
| kmalloc | and maybe json_schema | 17:56 |
| *** pcichy has quit IRC | 17:58 | |
| kmalloc | euuw. flask wants regparse instead of json-schema... yeah i'll just implement json-schema support directly | 17:58 |
| *** r-daneel_ has joined #openstack-keystone | 18:00 | |
| kmalloc | oh nvm, we already od it, we could move the json-schema bits higher up in flask though, nice | 18:00 |
| lbragstad | cool - make api validation happen closer to the edge of the app then | 18:00 |
| *** r-daneel has quit IRC | 18:01 | |
| *** r-daneel_ is now known as r-daneel | 18:01 | |
| kmalloc | lbragstad: yeah. | 18:09 |
| *** harlowja has joined #openstack-keystone | 18:12 | |
| *** mvk has quit IRC | 18:22 | |
| *** idlemind has joined #openstack-keystone | 18:27 | |
| kmalloc | lbragstad: ... how does our json_home thing even work? | 18:27 |
| cmurphy | magic | 18:28 |
| kmalloc | cmurphy: yeah... it looks like it | 18:34 |
| kmalloc | cmurphy: =/ | 18:34 |
| lbragstad | i haven't dug into the details of it in a while but it uses the routes bits to build the document, then just emits that when content-type: application/json | 18:37 |
| lbragstad | iirc brant did a bunch of that stuff | 18:38 |
| *** Guest16323 is now known as mgagne | 18:40 | |
| *** mgagne has joined #openstack-keystone | 18:40 | |
| *** markvoelker_ has joined #openstack-keystone | 18:42 | |
| *** markvoelker has quit IRC | 18:43 | |
| lbragstad | i'm not sure i have all the context on this yet, but is there any reason why we don't use oslo.service? | 18:44 |
| lbragstad | or is it just because oslo.service came after we had something that did pretty much the same thing? | 18:44 |
| *** dklyle has joined #openstack-keystone | 18:49 | |
| *** markvoelker has joined #openstack-keystone | 18:49 | |
| *** markvoelker_ has quit IRC | 18:50 | |
| *** mvk has joined #openstack-keystone | 18:52 | |
| *** r-daneel has quit IRC | 18:59 | |
| *** r-daneel_ has joined #openstack-keystone | 18:59 | |
| *** r-daneel_ is now known as r-daneel | 19:01 | |
| *** jmlowe has quit IRC | 19:13 | |
| *** jmlowe has joined #openstack-keystone | 19:13 | |
| *** jmlowe has quit IRC | 19:13 | |
| lbragstad | kmalloc: i think we're digging close to the same area of code actually | 19:19 |
| lbragstad | i'm looking at what it would take to implement the community goal for this release, without having to intercept SIGUP signals | 19:20 |
| lbragstad | one of the suggestions is to setup a listener to see if the configuration file changes during run time, and just reload the logger | 19:21 |
| lbragstad | that'd would have to take place after https://github.com/openstack/keystone/blob/6bb59948e956b67a6b5b5d17f17c7e7eb291ef7e/keystone/server/wsgi.py#L72 wouldn't it? | 19:22 |
| lbragstad | becuase application is what is returned to the web server, right? | 19:25 |
| *** jmlowe has joined #openstack-keystone | 19:34 | |
| *** dave-mccowan has quit IRC | 19:36 | |
| *** belmoreira has joined #openstack-keystone | 19:38 | |
| *** felipemonteiro_ has joined #openstack-keystone | 19:38 | |
| *** felipemonteiro has quit IRC | 19:42 | |
| *** devx has quit IRC | 19:43 | |
| *** devx has joined #openstack-keystone | 19:44 | |
| *** dave-mccowan has joined #openstack-keystone | 19:53 | |
| *** jmlowe has quit IRC | 20:06 | |
| *** jmlowe has joined #openstack-keystone | 20:08 | |
| *** felipemonteiro_ has quit IRC | 20:17 | |
| *** felipemonteiro_ has joined #openstack-keystone | 20:17 | |
| openstackgerrit | Brian Rosmaita proposed openstack/keystoneauth master: WIP - fix logging of encoded headers https://review.openstack.org/568365 | 20:25 |
| *** felipemonteiro__ has joined #openstack-keystone | 20:29 | |
| *** devx has quit IRC | 20:30 | |
| *** devx has joined #openstack-keystone | 20:30 | |
| *** devx has quit IRC | 20:31 | |
| *** devx has joined #openstack-keystone | 20:31 | |
| *** dklyle has quit IRC | 20:31 | |
| *** felipemonteiro_ has quit IRC | 20:32 | |
| *** belmoreira has quit IRC | 20:33 | |
| *** devx has quit IRC | 20:34 | |
| *** devx has joined #openstack-keystone | 20:34 | |
| openstackgerrit | Brian Rosmaita proposed openstack/keystoneauth master: WIP - fix logging of encoded headers https://review.openstack.org/568365 | 20:35 |
| *** devx has quit IRC | 20:38 | |
| *** devx has joined #openstack-keystone | 20:38 | |
| *** jmlowe has quit IRC | 20:39 | |
| *** devx has joined #openstack-keystone | 20:39 | |
| *** jmlowe has joined #openstack-keystone | 20:41 | |
| *** belmoreira has joined #openstack-keystone | 20:42 | |
| *** pcaruana has quit IRC | 20:43 | |
| *** felipemonteiro_ has joined #openstack-keystone | 20:46 | |
| *** felipemonteiro__ has quit IRC | 20:46 | |
| *** edmondsw has quit IRC | 20:47 | |
| *** aojea_ has joined #openstack-keystone | 20:47 | |
| *** germs has joined #openstack-keystone | 20:47 | |
| *** germs has quit IRC | 20:47 | |
| *** germs has joined #openstack-keystone | 20:47 | |
| *** edmondsw has joined #openstack-keystone | 20:47 | |
| *** belmoreira has quit IRC | 20:47 | |
| *** germs has quit IRC | 20:51 | |
| *** edmondsw has quit IRC | 20:52 | |
| *** devx has quit IRC | 20:53 | |
| *** devx has joined #openstack-keystone | 20:53 | |
| openstackgerrit | prashkre proposed openstack/python-keystoneclient master: WIP: Allow passing log attribute https://review.openstack.org/568373 | 20:55 |
| *** spilla has quit IRC | 20:57 | |
| kmalloc | lbragstad: ah nice | 21:01 |
| kmalloc | lbragstad: so, as long as we are doing something like inotify, we can have everything check the file for changes | 21:02 |
| kmalloc | but we're going to need to re-work how we handle the cases of instantiating managers | 21:02 |
| lbragstad | dhellmann just had some input on that front | 21:02 |
| kmalloc | because they read from the files and it could be bad(tm) if we reconfigure mid-request. | 21:02 |
| lbragstad | yeah... | 21:03 |
| lbragstad | and it kinda goes against the direction the oslo.config wants to take for pluggable config backends (e.g. secret storage) that aren't file-based | 21:03 |
| kmalloc | yeah | 21:03 |
| lbragstad | http://lists.openstack.org/pipermail/openstack-dev/2018-May/130467.html | 21:03 |
| lbragstad | so maybe not as robust as i was thinking | 21:03 |
| kmalloc | my view is we can support something like apachectl reconfigure (SIGHUP) for the parent uwsgi process | 21:03 |
| kmalloc | which should then winddown/cycle the subsequent processes | 21:04 |
| kmalloc | [or have a pipe we can issue a command on[] | 21:04 |
| kmalloc | pipe/socket | 21:04 |
| lbragstad | i think that was along the lines of fungi's suggestion | 21:04 |
| lbragstad | pipe/socket | 21:04 |
| kmalloc | that would be my go-to design | 21:04 |
| kmalloc | want to see the first bits of the paste-deploy-ectomy/flaskification? | 21:05 |
| kmalloc | let me push this review up. | 21:05 |
| lbragstad | sure | 21:05 |
| lbragstad | fwiw - i punted on trying to figure out the mutable config stuff today and put it on the meeting schedule for tomorrow | 21:05 |
| openstackgerrit | Morgan Fainberg proposed openstack/keystone master: Convert Keystone to use Flask https://review.openstack.org/568377 | 21:05 |
| kmalloc | sounds good | 21:05 |
| kmalloc | mutable configs are tough | 21:05 |
| kmalloc | ^ that is the first pass [still needs lots of work] | 21:06 |
| kmalloc | and that wont pass check / gate / pep8 | 21:06 |
| kmalloc | but that is the start. | 21:06 |
| lbragstad | i'm not sure which is best, and i don't know if converting to oslo.service and intercepting SIGUP when it's clearly documented as a no-no against mod_wsgi is a good thing | 21:06 |
| *** r-daneel has quit IRC | 21:07 | |
| kmalloc | yeah, and with mod_Wsgi, you would need apachectl reconfigure anyway | 21:07 |
| *** r-daneel has joined #openstack-keystone | 21:07 | |
| kmalloc | you can't do pipe/socket really | 21:07 |
| *** rmascena has quit IRC | 21:07 | |
| kmalloc | with uwsgi / gunicorn, we are in a better state to do something. | 21:07 |
| kmalloc | but ... still not "great" | 21:07 |
| kmalloc | fwiw, the "_path_prefix" values are temporary | 21:08 |
| *** martinus__ has quit IRC | 21:08 | |
| * fungi is shocked at having had a suggestion... he has a short memory | 21:08 | |
| kmalloc | that is just so i can build the dispatcher map. | 21:08 |
| kmalloc | then we can convert each subsystem into a flask "blueprint" | 21:09 |
| lbragstad | ok | 21:09 |
| fungi | if memory serves, my suggestion was "signal handler or rpc socket" | 21:09 |
| fungi | pretty vague | 21:09 |
| kmalloc | fungi: lol ;) | 21:09 |
| kmalloc | fungi: well rpc-socket would be my choice. | 21:09 |
| kmalloc | though, like i said, under mod_Wsgi, you're better off doing an apachectl reconfigure anyway, since apache owns all the processes. | 21:10 |
| fungi | sure. signal handling is kinda old-school bsd daemon think | 21:10 |
| kmalloc | and some of the wsgi runners do a poooooor job of signal handling. | 21:10 |
| kmalloc | lbragstad: i'm going to enable at least one hook point for custom middleware. | 21:11 |
| kmalloc | it'll be a new config value, ListOpt and it will take stevedore-loadable entry-points | 21:12 |
| kmalloc | so: oslo.middleware:debug | 21:12 |
| kmalloc | and parse those. | 21:12 |
| kmalloc | and load them in. | 21:12 |
| lbragstad | ok | 21:13 |
| *** felipemonteiro_ has quit IRC | 21:20 | |
| kmalloc | lbragstad: do we want to support middleware hook after ours or just before? | 21:20 |
| lbragstad | today we support both, right? | 21:20 |
| kmalloc | e.g. just before healthcheck [if you look at paste-ini now], or just after json_body, or both | 21:20 |
| lbragstad | but we don't guarantee it will work | 21:20 |
| kmalloc | right now, we support anywhere | 21:20 |
| lbragstad | hmm | 21:21 |
| kmalloc | i'm inclined to only support "pre" our middleware | 21:21 |
| lbragstad | i'm inclined to say before? | 21:21 |
| kmalloc | yeah | 21:21 |
| lbragstad | just because once we run our middleware, we should pass it to our app | 21:21 |
| *** jmlowe has quit IRC | 21:21 | |
| kmalloc | that is my inclination | 21:22 |
| lbragstad | supporting the ability to do things in between those events seems like a good way override what we do in middleware | 21:22 |
| kmalloc | but, that has the effect that no one can hook in after we validate the token | 21:22 |
| lbragstad | yeah... | 21:22 |
| lbragstad | what would we want to have people do with the token before passing control to keystone? | 21:23 |
| *** sonuk has joined #openstack-keystone | 21:30 | |
| kmalloc | ifthey wanted to add an extension or something that handles code to keystone | 21:30 |
| kmalloc | their own apis | 21:30 |
| kmalloc | i'm disinclined to support tht | 21:30 |
| kmalloc | that* | 21:30 |
| lbragstad | yeah | 21:33 |
| lbragstad | i'm struggling to think of a good use case for that right now | 21:34 |
| kmalloc | ok, i think... i think i'm now at the point when i need to swap over to the new app factory | 21:35 |
| kmalloc | and replace the "load_app" bit from paste | 21:35 |
| kmalloc | this is kindof awesome. | 21:35 |
| kmalloc | https://www.irccloud.com/pastebin/aNq7s6El/ | 21:36 |
| kmalloc | this might not even be too bad to review | 21:36 |
| *** jmlowe has joined #openstack-keystone | 21:36 | |
| *** sonuk has quit IRC | 21:36 | |
| lbragstad | kmalloc: https://review.openstack.org/#/c/555279/6 might be good for you to look at when you have time | 21:38 |
| *** jmlowe has quit IRC | 21:41 | |
| *** rcernin has joined #openstack-keystone | 21:53 | |
| *** harlowja has quit IRC | 21:56 | |
| *** dklyle has joined #openstack-keystone | 21:59 | |
| *** edmondsw has joined #openstack-keystone | 22:03 | |
| gagehugo | lbragstad I've been meaning to find something relatively stable to test that list_users refactor on | 22:07 |
| *** jistr has quit IRC | 22:07 | |
| *** edmondsw has quit IRC | 22:08 | |
| *** dklyle has quit IRC | 22:08 | |
| *** aojea_ has quit IRC | 22:08 | |
| *** jistr has joined #openstack-keystone | 22:10 | |
| *** threestrands has joined #openstack-keystone | 22:16 | |
| lbragstad | gagehugo: even if it's just a dev box with minimal stuff running | 22:30 |
| lbragstad | and you can abstract the performance improves into percentages | 22:31 |
| lbragstad | that'd be just fine imo | 22:31 |
| gagehugo | lbragstad I have a raspberry pi I could use :) | 22:37 |
| gagehugo | but I may have a dev laptop that I could wipe for testing | 22:38 |
| *** dklyle has joined #openstack-keystone | 22:50 | |
| *** r-daneel has quit IRC | 22:54 | |
| adriant | Any idea why keystonemiddleware would be returning: {"error": "Authentication Invalid"} when delay_auth_decision is true? | 22:56 |
| *** dklyle has quit IRC | 22:56 | |
| kmalloc | not sure. | 22:57 |
| adriant | weird. I'll have to dig further. Am having issues wrapping a little flask app with it | 22:57 |
| adriant | kmalloc: pretty much all I'm doing is: http://paste.openstack.org/show/720962/ and that worked in the past so I'm not sure if I've screwed something up | 23:01 |
| *** dave-mccowan has quit IRC | 23:02 | |
| adriant | NVM, found the issue | 23:21 |
| adriant | it's not a middleware problem... it's our code | 23:21 |
| *** dklyle has joined #openstack-keystone | 23:35 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!