| *** aojea has joined #openstack-keystone | 00:06 | |
| *** jrist has quit IRC | 00:07 | |
| *** aojea has quit IRC | 00:10 | |
| *** aojea has joined #openstack-keystone | 00:15 | |
| *** aojea has quit IRC | 00:19 | |
| *** markvoelker has quit IRC | 00:20 | |
| *** edmondsw has joined #openstack-keystone | 00:20 | |
| *** jrist has joined #openstack-keystone | 00:21 | |
| *** aojea has joined #openstack-keystone | 00:24 | |
| *** edmondsw has quit IRC | 00:25 | |
| *** thorst has joined #openstack-keystone | 00:29 | |
| *** aojea has quit IRC | 00:29 | |
| *** ducttape_ has joined #openstack-keystone | 00:33 | |
| *** aojea has joined #openstack-keystone | 00:33 | |
| *** thorst has quit IRC | 00:33 | |
| *** markvoelker has joined #openstack-keystone | 00:35 | |
| *** aojea has quit IRC | 00:37 | |
| *** aojea has joined #openstack-keystone | 00:42 | |
| *** aojea has quit IRC | 00:47 | |
| *** ducttape_ has quit IRC | 00:50 | |
| *** aojea has joined #openstack-keystone | 00:51 | |
| *** aojea has quit IRC | 00:55 | |
| *** thorst has joined #openstack-keystone | 00:59 | |
| *** aojea has joined #openstack-keystone | 01:00 | |
| *** aojea has quit IRC | 01:05 | |
| *** ducttape_ has joined #openstack-keystone | 01:06 | |
| *** thorst has quit IRC | 01:07 | |
| *** aojea has joined #openstack-keystone | 01:09 | |
| *** aojea has quit IRC | 01:14 | |
| *** ducttape_ has quit IRC | 01:15 | |
| *** aojea has joined #openstack-keystone | 01:18 | |
| *** lucasxu has joined #openstack-keystone | 01:22 | |
| *** aojea has quit IRC | 01:23 | |
| *** lucasxu has quit IRC | 01:24 | |
| *** aojea has joined #openstack-keystone | 01:27 | |
| *** harlowja has quit IRC | 01:29 | |
| *** aojea has quit IRC | 01:32 | |
| *** Shunli has joined #openstack-keystone | 01:32 | |
| *** aojea has joined #openstack-keystone | 01:36 | |
| *** aojea has quit IRC | 01:41 | |
| *** aojea has joined #openstack-keystone | 01:46 | |
| *** aojea has quit IRC | 01:50 | |
| *** aojea has joined #openstack-keystone | 01:55 | |
| *** aojea has quit IRC | 01:59 | |
| *** aojea has joined #openstack-keystone | 02:04 | |
| *** aojea has quit IRC | 02:08 | |
| *** edmondsw has joined #openstack-keystone | 02:08 | |
| *** otleimat has quit IRC | 02:11 | |
| *** aojea has joined #openstack-keystone | 02:13 | |
| *** edmondsw has quit IRC | 02:13 | |
| *** aojea has quit IRC | 02:17 | |
| *** prashkre has joined #openstack-keystone | 02:38 | |
| *** aojea has joined #openstack-keystone | 02:40 | |
| *** aojea has quit IRC | 02:45 | |
| *** aojea has joined #openstack-keystone | 02:49 | |
| *** aojea has quit IRC | 02:53 | |
| *** prashkre has quit IRC | 02:56 | |
| *** aojea has joined #openstack-keystone | 02:58 | |
| *** aojea has quit IRC | 03:02 | |
| *** aojea has joined #openstack-keystone | 03:07 | |
| *** aojea has quit IRC | 03:11 | |
| *** aojea has joined #openstack-keystone | 03:16 | |
| *** aojea has quit IRC | 03:21 | |
| *** aojea has joined #openstack-keystone | 03:25 | |
| *** zsli_ has joined #openstack-keystone | 03:26 | |
| *** Shunli has quit IRC | 03:28 | |
| *** zsli_ has quit IRC | 03:30 | |
| *** aojea has quit IRC | 03:30 | |
| *** links has joined #openstack-keystone | 03:30 | |
| *** links has quit IRC | 03:30 | |
| *** aojea has joined #openstack-keystone | 03:35 | |
| *** aojea has quit IRC | 03:39 | |
| *** aojea has joined #openstack-keystone | 03:44 | |
| *** aojea has quit IRC | 03:48 | |
| *** aojea has joined #openstack-keystone | 03:53 | |
| *** edmondsw has joined #openstack-keystone | 03:57 | |
| *** aojea has quit IRC | 03:57 | |
| *** edmondsw has quit IRC | 04:01 | |
| *** nicolasbock has joined #openstack-keystone | 04:04 | |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add new tags attribute to project https://review.openstack.org/470317 | 04:21 |
|---|---|---|
| *** deep-book-gk_ has joined #openstack-keystone | 04:24 | |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add new tags attribute to project https://review.openstack.org/470317 | 04:24 |
| *** deep-book-gk_ has left #openstack-keystone | 04:25 | |
| *** oomichi has quit IRC | 04:29 | |
| *** oomichi has joined #openstack-keystone | 04:30 | |
| *** kornicameister has joined #openstack-keystone | 04:32 | |
| *** gongysh has joined #openstack-keystone | 04:34 | |
| *** harlowja has joined #openstack-keystone | 04:40 | |
| *** gongysh has quit IRC | 04:40 | |
| *** thorst has joined #openstack-keystone | 05:04 | |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add new tags attribute to project https://review.openstack.org/470317 | 05:07 |
| *** thorst has quit IRC | 05:10 | |
| *** ducttape_ has joined #openstack-keystone | 05:11 | |
| *** markvoelker has quit IRC | 05:13 | |
| *** ducttape_ has quit IRC | 05:16 | |
| *** harlowja has quit IRC | 05:39 | |
| *** gongysh has joined #openstack-keystone | 05:40 | |
| *** aojea has joined #openstack-keystone | 05:41 | |
| *** edmondsw has joined #openstack-keystone | 05:45 | |
| *** aojea has quit IRC | 05:46 | |
| *** edmondsw has quit IRC | 05:49 | |
| openstackgerrit | Tin Lam proposed openstack/python-keystoneclient master: Add project tags to keystoneclient https://review.openstack.org/481223 | 05:53 |
| *** prashkre has joined #openstack-keystone | 05:56 | |
| *** Shunli has joined #openstack-keystone | 05:59 | |
| *** rcernin has joined #openstack-keystone | 06:06 | |
| *** aojea has joined #openstack-keystone | 06:09 | |
| *** aojea has quit IRC | 06:14 | |
| *** aojea has joined #openstack-keystone | 06:18 | |
| *** aojea has quit IRC | 06:23 | |
| *** aojea has joined #openstack-keystone | 06:27 | |
| *** thorst has joined #openstack-keystone | 06:29 | |
| *** aojea has quit IRC | 06:32 | |
| *** thorst has quit IRC | 06:33 | |
| *** aojea has joined #openstack-keystone | 06:36 | |
| *** blake has joined #openstack-keystone | 06:39 | |
| *** aojea has quit IRC | 06:41 | |
| *** aojea has joined #openstack-keystone | 06:45 | |
| *** aojea has quit IRC | 06:50 | |
| *** blake has quit IRC | 06:51 | |
| *** aojea has joined #openstack-keystone | 06:54 | |
| *** aojea has quit IRC | 06:59 | |
| *** aojea has joined #openstack-keystone | 07:03 | |
| *** belmoreira has joined #openstack-keystone | 07:05 | |
| *** aojea has quit IRC | 07:08 | |
| *** mvpnitesh has joined #openstack-keystone | 07:08 | |
| *** aojea has joined #openstack-keystone | 07:13 | |
| *** ducttape_ has joined #openstack-keystone | 07:13 | |
| *** markvoelker has joined #openstack-keystone | 07:14 | |
| *** aojea has quit IRC | 07:14 | |
| *** aojea has joined #openstack-keystone | 07:14 | |
| *** ducttape_ has quit IRC | 07:17 | |
| *** phalmos has quit IRC | 07:23 | |
| *** edmondsw has joined #openstack-keystone | 07:33 | |
| *** edmondsw has quit IRC | 07:37 | |
| *** jaosorior has quit IRC | 07:43 | |
| *** dims has quit IRC | 07:45 | |
| *** tobberydberg has joined #openstack-keystone | 07:48 | |
| *** markvoelker has quit IRC | 07:48 | |
| *** dims has joined #openstack-keystone | 07:49 | |
| *** jaosorior has joined #openstack-keystone | 07:57 | |
| *** jaosorior has quit IRC | 07:58 | |
| *** jaosorior has joined #openstack-keystone | 07:59 | |
| *** zsli_ has joined #openstack-keystone | 08:00 | |
| *** jaosorior has quit IRC | 08:02 | |
| *** Shunli has quit IRC | 08:03 | |
| *** jaosorior has joined #openstack-keystone | 08:03 | |
| *** jaosorior has quit IRC | 08:03 | |
| *** jaosorior has joined #openstack-keystone | 08:04 | |
| *** zsli__ has joined #openstack-keystone | 08:08 | |
| *** zsli_ has quit IRC | 08:11 | |
| *** gongysh has quit IRC | 08:12 | |
| *** jaosorior has quit IRC | 08:13 | |
| *** zsli__ has quit IRC | 08:15 | |
| *** aselius has quit IRC | 08:15 | |
| *** Shunli has joined #openstack-keystone | 08:18 | |
| *** zsli_ has joined #openstack-keystone | 08:23 | |
| *** Shunli has quit IRC | 08:26 | |
| *** thorst has joined #openstack-keystone | 08:30 | |
| *** mvpnitesh has quit IRC | 08:32 | |
| *** mvpnitesh has joined #openstack-keystone | 08:33 | |
| *** thorst has quit IRC | 08:34 | |
| *** stevemar has quit IRC | 08:36 | |
| *** stevemar has joined #openstack-keystone | 08:36 | |
| *** markvoelker has joined #openstack-keystone | 08:45 | |
| *** tobberydberg has quit IRC | 08:49 | |
| *** tobberydberg has joined #openstack-keystone | 08:49 | |
| *** jaosorior has joined #openstack-keystone | 08:53 | |
| *** tobberydberg has quit IRC | 08:54 | |
| *** jaosorior has quit IRC | 09:03 | |
| *** tobberydberg has joined #openstack-keystone | 09:13 | |
| *** ducttape_ has joined #openstack-keystone | 09:14 | |
| *** markvoelker has quit IRC | 09:19 | |
| *** john5223_ has quit IRC | 09:19 | |
| *** timburke has quit IRC | 09:20 | |
| *** timburke has joined #openstack-keystone | 09:20 | |
| *** jaosorior has joined #openstack-keystone | 09:21 | |
| *** ducttape_ has quit IRC | 09:21 | |
| *** gongysh has joined #openstack-keystone | 09:21 | |
| *** edmondsw has joined #openstack-keystone | 09:22 | |
| *** iurygregory has quit IRC | 09:22 | |
| *** cristicalin has joined #openstack-keystone | 09:22 | |
| *** odyssey4me has quit IRC | 09:23 | |
| *** iurygregory has joined #openstack-keystone | 09:24 | |
| *** tobberydberg has quit IRC | 09:25 | |
| *** edmondsw has quit IRC | 09:25 | |
| *** tobberydberg has joined #openstack-keystone | 09:26 | |
| *** odyssey4me has joined #openstack-keystone | 09:27 | |
| *** mvpnitesh has quit IRC | 09:27 | |
| *** zsli_ has quit IRC | 09:28 | |
| *** tobberydberg has quit IRC | 09:30 | |
| *** mvk has quit IRC | 09:31 | |
| *** cristicalin has quit IRC | 09:31 | |
| *** aojea has quit IRC | 09:32 | |
| *** aojea has joined #openstack-keystone | 09:32 | |
| *** aojea has quit IRC | 09:57 | |
| *** thorst has joined #openstack-keystone | 09:58 | |
| *** aojea has joined #openstack-keystone | 09:59 | |
| *** mvk has joined #openstack-keystone | 10:04 | |
| *** thorst has quit IRC | 10:06 | |
| *** thorst has joined #openstack-keystone | 10:07 | |
| *** thorst has quit IRC | 10:11 | |
| *** markvoelker has joined #openstack-keystone | 10:16 | |
| *** jaosorior has quit IRC | 10:38 | |
| *** clarkb has quit IRC | 10:45 | |
| *** clarkb has joined #openstack-keystone | 10:45 | |
| *** markvoelker has quit IRC | 10:48 | |
| *** edmondsw has joined #openstack-keystone | 11:09 | |
| *** edmondsw has quit IRC | 11:13 | |
| *** ducttape_ has joined #openstack-keystone | 11:15 | |
| *** ducttape_ has quit IRC | 11:23 | |
| *** markvoelker has joined #openstack-keystone | 11:45 | |
| *** gongysh has quit IRC | 11:47 | |
| *** raildo has joined #openstack-keystone | 12:01 | |
| *** Dinesh_Bhor is now known as Dinesh_Bhor|away | 12:08 | |
| *** Dinesh_Bhor|away is now known as Dinesh_Bhor | 12:09 | |
| *** edmondsw has joined #openstack-keystone | 12:13 | |
| *** chlong has joined #openstack-keystone | 12:16 | |
| *** thorst has joined #openstack-keystone | 12:16 | |
| *** markvoelker has quit IRC | 12:19 | |
| *** catintheroof has joined #openstack-keystone | 12:21 | |
| openstackgerrit | Dinesh Bhor proposed openstack/python-keystoneclient master: Add Response class to return request-id to caller https://review.openstack.org/329913 | 12:21 |
| openstackgerrit | Dinesh Bhor proposed openstack/python-keystoneclient master: Add return-request-id-to-caller function(v3) https://review.openstack.org/267456 | 12:22 |
| openstackgerrit | Dinesh Bhor proposed openstack/python-keystoneclient master: Add return-request-id-to-caller function(v3/contrib) https://review.openstack.org/268003 | 12:22 |
| *** markvoelker has joined #openstack-keystone | 12:28 | |
| *** ducttape_ has joined #openstack-keystone | 12:40 | |
| dims | mordred lbragstad : seen this? https://bugs.launchpad.net/keystoneauth/+bug/1706538 | 12:42 |
| openstack | Launchpad bug 1706538 in python-troveclient ""ValueError: Expecting a string None" with keystoneauth 3.0.1" [Undecided,New] | 12:42 |
| *** amrith has joined #openstack-keystone | 12:43 | |
| amrith | dims ... here | 12:43 |
| dims | efried : ^ | 12:43 |
| *** ducttape_ has quit IRC | 12:45 | |
| *** coolsvap has joined #openstack-keystone | 12:45 | |
| *** coolsvap has left #openstack-keystone | 12:45 | |
| amrith | ok, I have no scrollback, let me go find eavesdrop | 12:45 |
| amrith | or someone can tell me what's transpired so far; who killed whom with what and where | 12:46 |
| amrith | dims, looks like the issue is that we're not sending up an 'expires' in the token request | 12:47 |
| amrith | is that the issue | 12:47 |
| amrith | is this a new thing that keystone wants? | 12:48 |
| amrith | why's there no default? | 12:48 |
| amrith | or did I miss something else ... dims ^^ | 12:48 |
| dims | amrith : i haven't looked, just got pointed to the bug report and i added keystoneauth to it | 12:48 |
| lbragstad | dims: checking | 12:58 |
| openstackgerrit | Dinesh Bhor proposed openstack/python-keystoneclient master: Add return-request-id-to-caller function(v3) https://review.openstack.org/267456 | 13:00 |
| openstackgerrit | Dinesh Bhor proposed openstack/python-keystoneclient master: Add return-request-id-to-caller function(v3/contrib) https://review.openstack.org/268003 | 13:00 |
| openstackgerrit | Dinesh Bhor proposed openstack/python-keystoneclient master: Add Response class to return request-id to caller https://review.openstack.org/329913 | 13:00 |
| lbragstad | amrith: expires is an attribute of the token itself | 13:02 |
| lbragstad | its strange that its not there, since it always exists in the token reference when keystone issues and validates tokens | 13:03 |
| *** dave-mccowan has joined #openstack-keystone | 13:09 | |
| mordred | dims: I hadn't - but I have now | 13:15 |
| lbragstad | it's because of the mock in troveclient | 13:15 |
| mordred | morning amrith | 13:16 |
| mordred | AH | 13:16 |
| amrith | g'morning mordred lbragstad | 13:16 |
| *** dave-mcc_ has joined #openstack-keystone | 13:16 | |
| mordred | incomplete mock then perhaps? | 13:16 |
| lbragstad | yeah - i think so? | 13:17 |
| amrith | lbragstad if you have it narrowed down, would you submit a patch | 13:17 |
| lbragstad | working on a patch now | 13:17 |
| mordred | lbragstad: I feel like maybe we should write up a little doc to tell people how to use the requests_mock mocks of tokens and stuff that are in keystoneauth | 13:17 |
| amrith | thx lbragstad | 13:17 |
| lbragstad | v2.0 doesn't have an expires_at attribute https://github.com/openstack/python-troveclient/blob/master/troveclient/tests/test_shell.py#L77 | 13:17 |
| *** dave-mccowan has quit IRC | 13:19 | |
| mordred | oh, that reminds me, I need to figure out the intersection of requests_mock and cachecontrol | 13:19 |
| *** jaosorior has joined #openstack-keystone | 13:24 | |
| lbragstad | amrith: https://review.openstack.org/#/c/487417/ | 13:25 |
| lbragstad | that passes for me locally | 13:25 |
| amrith | looking | 13:25 |
| amrith | @lbragstad change looks fine, if it passes that's great. how come it broke? this line wasn't added recently in the trove client. did something else change? | 13:31 |
| *** jistr is now known as jistr|mtg | 13:31 | |
| lbragstad | amrith: a lot of code in keystoneauth changed with version 3 | 13:33 |
| amrith | as you can see https://github.com/openstack/python-troveclient/blame/master/troveclient/tests/test_shell.py#L77, this has been around for a while. | 13:33 |
| amrith | that may be, but is this a change that breaks a contract? | 13:33 |
| lbragstad | my guess is that the ordering of how token expiration is evaluated changed | 13:33 |
| amrith | I'm fine with your fix but if it wasn't given the 1 version grace period, I'm not thrilled with your change | 13:34 |
| amrith | what else broke in the same way? | 13:34 |
| amrith | ok, trove stubbed its toes on this one, what other things are similar? | 13:34 |
| amrith | me, personally, I'd hold your change will we know what the actual thing is that changed to make this happen. | 13:35 |
| lbragstad | nothing in the contract changed | 13:35 |
| amrith | then why did this happen? | 13:35 |
| lbragstad | it's how python-troveclient built the mock | 13:35 |
| amrith | sorry, I don't follow. was 'expires_at' valid in the past? did it do anything? | 13:36 |
| amrith | was 'expires_at' deprecated? | 13:37 |
| lbragstad | amrith: no | 13:38 |
| lbragstad | amrith: this is an example authentication response from v2.0 http://paste.openstack.org/show/616582/ | 13:38 |
| lbragstad | amrith: this is an example authentication response from v3 http://paste.openstack.org/show/616583/ | 13:39 |
| lbragstad | python-troveclient was using a mock for v2.0 but referencing an expiration attribute only found in v3 | 13:39 |
| lbragstad | expires_at in v3 and expires in v2.0 relay the same information | 13:40 |
| lbragstad | they tell you when the token is going to expire | 13:40 |
| lbragstad | keystoneauth does some processing to determine if the token being used to talk to keystone is going to expire soon | 13:41 |
| amrith | i'm not following you. here is what I'm seeing, https://bugs.launchpad.net/python-troveclient/+bug/1706538 | 13:41 |
| openstack | Launchpad bug 1706538 in python-troveclient ""ValueError: Expecting a string None" with keystoneauth 3.0.1" [Undecided,In progress] - Assigned to Lance Bragstad (lbragstad) | 13:41 |
| amrith | keystone auth has some code in keystoneauth1/_utils.py which is looking for 'exires' | 13:41 |
| amrith | yes? | 13:41 |
| amrith | that's the reason for this failure | 13:41 |
| amrith | the mock (or something) was giving it expires_at | 13:41 |
| lbragstad | keystoneauth is looking for expires but the token reference doesn't have that attribute | 13:42 |
| lbragstad | and that is causing the failure | 13:42 |
| amrith | your v3 paste http://paste.openstack.org/show/616583/ has (at line 33) an expires_at | 13:42 |
| amrith | it has no expires | 13:42 |
| lbragstad | amrith: yes - that's because that's a v3 response | 13:42 |
| lbragstad | what python-troveclient is using is v2.0 | 13:43 |
| lbragstad | which is a different API | 13:43 |
| amrith | so how did this test pass in the past? | 13:43 |
| amrith | clearly, for some period of time in the past 3 years, this test has been passing, I think | 13:43 |
| amrith | I get that your fix makes it pass again, I don't understand why/how | 13:44 |
| amrith | it was that it passed in the past and stopped now | 13:44 |
| amrith | there has got to be something else that changed, and I can't see what that is. | 13:44 |
| *** catintheroof has quit IRC | 13:44 | |
| amrith | I get what you are saying, that expires_at is a v3 field and expires is a v2 field. | 13:44 |
| amrith | but v3 didn't exist 3 years ago when this code was written | 13:45 |
| amrith | (did it?) | 13:45 |
| lbragstad | v3 has been around for that long | 13:45 |
| *** catintheroof has joined #openstack-keystone | 13:46 | |
| amrith | ok, so something changed that made this test fail | 13:46 |
| amrith | ok, trove's test was wrong | 13:46 |
| amrith | what was the other change? | 13:46 |
| amrith | were there other things that similarly used to work but now don't | 13:46 |
| amrith | I don't know, I'm happy to hold my nose and +2 your change. | 13:46 |
| amrith | just strikes me as a shortcut | 13:47 |
| *** jistr|mtg is now known as jistr | 13:47 | |
| lbragstad | amrith: something in keystoneauth that moved the logic to check token expiration up a bit somewhere https://github.com/openstack/keystoneauth/blob/master/keystoneauth1/access/access.py#L87-L98 | 13:48 |
| *** catintheroof has quit IRC | 13:48 | |
| lbragstad | fwiw - trove has the same exact test for v3 https://github.com/openstack/python-troveclient/blob/555ba5ed690027088358288c9b0252d5a58bdc0d/troveclient/tests/test_shell.py#L330 | 13:48 |
| *** catintheroof has joined #openstack-keystone | 13:48 | |
| lbragstad | and it didn't break with the new version of keystoneauth https://github.com/openstack/python-troveclient/blob/555ba5ed690027088358288c9b0252d5a58bdc0d/troveclient/tests/test_shell.py#L249 | 13:48 |
| lbragstad | because it was using the right token attributes | 13:49 |
| lbragstad | in the mock | 13:49 |
| amrith | lbragstad I would not be so picky if this was a request, but this is a response. in other words in the past, a v2 response was saying expires_at. | 13:50 |
| amrith | is that correct? | 13:50 |
| amrith | (the mocked one) | 13:50 |
| lbragstad | no - v2.0 has never issued expires_at as an attribute of token responses | 13:50 |
| amrith | my point exactly, some code someplace was ok with this bogus mock | 13:51 |
| *** catintheroof has quit IRC | 13:51 | |
| *** catintheroof has joined #openstack-keystone | 13:52 | |
| *** rcernin has quit IRC | 13:54 | |
| efried | lbragstad mordred Is it worth making ksa more lenient with missing attrs? Seems like not - seems like the troveclient test simply suffered from RBB here. | 13:55 |
| lbragstad | efried: expires and expires_at are required attributes of tokens | 13:56 |
| * amrith listens, doesn't follow efried because this isn't a request attr but a response attr and the failure is *in keystone* invoked by troveclient | 13:57 | |
| efried | amrith I think the point here is that ksa would never give you back that response. | 13:58 |
| efried | ksa would give you a response with `expires`, not `expires_at`. | 13:58 |
| amrith | but it seemed to be fine with it in the past as part of a response | 13:58 |
| amrith | and didn't fail | 13:58 |
| efried | Luck :) | 13:58 |
| amrith | and clearly did something with it? | 13:59 |
| amrith | or the test is bogus (also possible) | 13:59 |
| efried | amrith Right, it may be that the response is now getting funnelled back through ksa in some path that expects the `expires` attr, which wasn't happening before. | 13:59 |
| efried | amrith Yahyah, I see it. The point of the test wasn't to validate the auth; it was to validate versions. So the token setup was incidental, and that token was getting passed back to ksa in order to do the version discovery. That's when ksa blew up on the invalid token. | 14:04 |
| efried | Why it wasn't going through that code path before is an academic question. | 14:04 |
| *** dave-mcc_ has quit IRC | 14:05 | |
| * amrith wonders | 14:05 | |
| efried | But certainly ksa3 has changed the get_endpoint plumbing to always go through the new get_endpoint_data. | 14:05 |
| efried | mordred could probably answer better there. | 14:06 |
| *** ppiela has left #openstack-keystone | 14:06 | |
| lbragstad | the new version of keystoneauth (3.0.1) ends up actually checking the token to see if it needs to be refreshed | 14:08 |
| lbragstad | https://github.com/openstack/keystoneauth/blob/master/keystoneauth1/identity/base.py#L304 | 14:08 |
| lbragstad | compared to previous versions https://github.com/openstack/keystoneauth/blob/stable/ocata/keystoneauth1/identity/base.py#L160 | 14:08 |
| lbragstad | which is the difference in how the get_endpoint() method changed with version 3.0.1 | 14:09 |
| efried | Yup - get_endpoint_data is doing get_project_id, which get_endpoint wasn't doing before. | 14:09 |
| lbragstad | right | 14:10 |
| lbragstad | because it's using get_access() it automatically checks if the token needs to be refreshed | 14:10 |
| efried | lbragstad mordred Is that actually a regression? Like, before, you could get an endpoint from the catalog (a catalog you already had) without rechecking the token? | 14:11 |
| lbragstad | efried: the older version of get_endpoint() also relied on get_access(), which does the same thing https://github.com/openstack/keystoneauth/blob/stable/ocata/keystoneauth1/identity/base.py#L212 | 14:12 |
| lbragstad | which means this same thing could have happened ^ if that path was invoked in versions of keystoneauth < 3 | 14:14 |
| efried | lbragstad That's interesting; cause get_access didn't change. | 14:15 |
| lbragstad | efried: right - but this case must have not been executed https://github.com/openstack/keystoneauth/blob/stable/ocata/keystoneauth1/identity/base.py#L203-L216 | 14:17 |
| lbragstad | otherwise - you would have seen the same failure we saw today | 14:17 |
| *** aojea has quit IRC | 14:31 | |
| *** otleimat has joined #openstack-keystone | 14:31 | |
| *** ducttape_ has joined #openstack-keystone | 14:41 | |
| prashkre | Hi. Anyone for core team, could you please review https://review.openstack.org/#/c/485302/ for another +2 and workflow. | 14:45 |
| *** ducttape_ has quit IRC | 14:52 | |
| *** prashkre has quit IRC | 14:55 | |
| *** blake has joined #openstack-keystone | 15:00 | |
| openstackgerrit | Lance Bragstad proposed openstack/keystone master: Move performance documentation to admin-guide https://review.openstack.org/481383 | 15:01 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystone master: Updated from global requirements https://review.openstack.org/487465 | 15:02 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystonemiddleware master: Updated from global requirements https://review.openstack.org/487466 | 15:02 |
| *** ducttape_ has joined #openstack-keystone | 15:09 | |
| lbragstad | gagehugo: lamt are tags sanitized in other projects? | 15:18 |
| lbragstad | e.g. is `myTag` converted to `mytag`? | 15:18 |
| lamt | lbragstad : nova's tags are case sensitive - that caused some issue in the the implementation | 15:19 |
| lamt | so myTag != mytag for tag name | 15:19 |
| lbragstad | so it's possible to have `myTag` and `mytag` as two separate tags on a project | 15:20 |
| gagehugo | yes | 15:20 |
| *** rcernin has joined #openstack-keystone | 15:21 | |
| *** jaosorior has quit IRC | 15:28 | |
| *** belmoreira has quit IRC | 15:29 | |
| *** rcernin has quit IRC | 15:34 | |
| *** aselius has joined #openstack-keystone | 15:36 | |
| *** aojea has joined #openstack-keystone | 15:42 | |
| *** aojea has quit IRC | 15:46 | |
| *** efried has quit IRC | 15:49 | |
| *** mvk has quit IRC | 15:49 | |
| *** Drankis has joined #openstack-keystone | 15:55 | |
| *** ducttape_ has quit IRC | 15:59 | |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add new tags attribute to project https://review.openstack.org/470317 | 16:00 |
| *** efried has joined #openstack-keystone | 16:02 | |
| *** tristanC has quit IRC | 16:03 | |
| *** rcernin has joined #openstack-keystone | 16:03 | |
| samueldmq | lbragstad: lamt: gagehugo: few more comments in the initial patches for project tags | 16:04 |
| *** ducttape_ has joined #openstack-keystone | 16:05 | |
| gagehugo | samueldmq sure | 16:05 |
| samueldmq | gagehugo: we can remove the id from the table and make tag+project_id the pk for the table, correct? | 16:05 |
| samueldmq | I think we discussed that before | 16:05 |
| lamt | samueldmq yeah, I think we accidentally patched that back in | 16:06 |
| gagehugo | samueldmq yes, the current PS for db migration got rebased to an old one, fixing that atm | 16:06 |
| gagehugo | rebasing 4 patches on top of one another is interesting | 16:06 |
| lbragstad | ok - so we don't need the id? | 16:07 |
| lbragstad | was having the id related to being case sensitive somehow? | 16:07 |
| *** chlong has quit IRC | 16:08 | |
| gagehugo | I don't think so, the main issue we had with case sensitivity was the unicode | 16:08 |
| samueldmq | The main one is not too far either | 16:08 |
| samueldmq | lbragstad: no | 16:08 |
| samueldmq | So nice, the first 3 patches are really close then | 16:08 |
| samueldmq | The main one is not too far either | 16:08 |
| lbragstad | ok - good deal | 16:08 |
| lbragstad | i thought i saw a comment somewhere about that | 16:09 |
| *** harlowja has joined #openstack-keystone | 16:15 | |
| samueldmq | lbragstad: let's try to get all the chain approved before we workflow+1 it all at once, if that makes sense | 16:21 |
| lbragstad | sure - that's fine | 16:22 |
| *** mvk has joined #openstack-keystone | 16:23 | |
| samueldmq | gagehugo: lamt: so ... for the project tag implementation | 16:27 |
| samueldmq | I argue we should return the project entity WITH tags from the manager layer already. | 16:27 |
| samueldmq | so we dont need to do it manually all over | 16:27 |
| samueldmq | and the manager will return the full representation of a project (as it does today) with the new attribute (tags) | 16:28 |
| samueldmq | that should make the code simpler too, see comment inline and let me know what you think | 16:28 |
| lamt | samueldmq I agree, we found out the hard way there are other things that can return a project entity not under /v3/projects/..., e.g. the endpoint filtering | 16:29 |
| gagehugo | the issue though is domains are weird | 16:30 |
| lamt | so some unit tests failed because of the missing tags=[] | 16:30 |
| lbragstad | knikolla: is there a specific goal you had in mind for https://blueprints.launchpad.net/keystone/+spec/devstack-plugin ? | 16:30 |
| gagehugo | but we could do that | 16:31 |
| lbragstad | knikolla: wondering what to put for the series goal there - since the specification is still ongoing | 16:31 |
| knikolla | lbragstad: not really. it's a CI thing, so not really user facing and release specific | 16:34 |
| lbragstad | ok - i'll leave it unset then for now | 16:35 |
| lbragstad | stepping away for lunch quick | 16:37 |
| *** spilla_ has joined #openstack-keystone | 16:40 | |
| *** tristanC has joined #openstack-keystone | 16:48 | |
| openstackgerrit | Rohan Arora proposed openstack/python-keystoneclient master: Add project tags to keystoneclient https://review.openstack.org/481223 | 16:49 |
| *** blake has quit IRC | 16:52 | |
| samueldmq | gagehugo: lamt: domains are projects. so there is no issue on domains inheriting tags too | 16:55 |
| samueldmq | it makes total sense to me | 16:55 |
| gagehugo | some of the logic for domains seemed weird in the manager, but I will take another look | 16:56 |
| samueldmq | not sure if others think differently, it might be worth it hear from others if you disagree | 16:56 |
| gagehugo | I would think domains can be tagged | 16:57 |
| samueldmq | gagehugo: ok. so imo you should just add the tags, regardless any logic for domains or projects | 16:57 |
| gagehugo | ok | 16:57 |
| samueldmq | might be just before the return statement. add the tags and return | 16:57 |
| samueldmq | should not be hard | 16:57 |
| *** aojea has joined #openstack-keystone | 16:58 | |
| *** aojea has quit IRC | 17:02 | |
| *** ducttape_ has quit IRC | 17:06 | |
| *** blake has joined #openstack-keystone | 17:06 | |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add new tags attribute to project https://review.openstack.org/470317 | 17:16 |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add JSON schema validation for project tags https://review.openstack.org/484483 | 17:16 |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add database migration for project tags https://review.openstack.org/484456 | 17:16 |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add policy for project tags https://review.openstack.org/486757 | 17:16 |
| mordred | efried, lbragstad: reading scrollback | 17:16 |
| *** prashkre has joined #openstack-keystone | 17:18 | |
| mordred | lbragstad, efried: yah - I don't think it's a regression - I think that behavior differences that arise from ways in which people have mocked something incompletely aren't a thing we can possibly deal with | 17:18 |
| *** ducttape_ has joined #openstack-keystone | 17:18 | |
| mordred | efried, lbragstad: HOWEVER - I think that we can definitely give people good reusable ksa mocks (and do already) along with some docs - so that it's not people's job to mock keystoneauth themselves | 17:19 |
| mordred | since writing mocks that provide appropriate token data, catalogs and discovery data is ... complex | 17:20 |
| lbragstad | mordred: yeah - i would agree with that | 17:23 |
| lbragstad | documentation on sound mocks would be a good start | 17:23 |
| *** spilla_ has quit IRC | 17:25 | |
| *** chlong has joined #openstack-keystone | 17:25 | |
| mordred | lbragstad: ++ | 17:25 |
| mordred | lbragstad: I can't get to that this week - but I might be able to next week or the week after | 17:25 |
| lbragstad | mordred: understandable | 17:26 |
| *** chlong has quit IRC | 17:27 | |
| *** chlong has joined #openstack-keystone | 17:28 | |
| *** aojea has joined #openstack-keystone | 17:31 | |
| *** aojea has quit IRC | 17:32 | |
| *** aojea has joined #openstack-keystone | 17:34 | |
| knikolla | cmurphy: for the broken tempest. the errors seems to have disappeared :/ i see no errors on the logs but it still didn't work. | 17:35 |
| openstackgerrit | Lance Bragstad proposed openstack/keystone-specs master: Bump application credentials to backlog https://review.openstack.org/487533 | 17:37 |
| openstackgerrit | Lance Bragstad proposed openstack/keystone-specs master: Bump support for federated attributes to backlog https://review.openstack.org/487534 | 17:39 |
| *** prashkre has quit IRC | 17:42 | |
| *** harlowja has quit IRC | 17:46 | |
| *** aojea has quit IRC | 17:47 | |
| openstackgerrit | Lance Bragstad proposed openstack/keystone master: Fix typo in index documentation https://review.openstack.org/479998 | 17:48 |
| openstackgerrit | Eric Fried proposed openstack/keystone master: Fix typo in index documentation https://review.openstack.org/479998 | 17:52 |
| openstackgerrit | Merged openstack/keystone master: Added new subsections to developer docs https://review.openstack.org/476635 | 17:58 |
| openstackgerrit | Kristi Nikolla proposed openstack/keystone master: In the devstack plugin, restart keystone after modifying conf https://review.openstack.org/487202 | 17:58 |
| openstackgerrit | Merged openstack/keystone master: Move performance documentation to admin-guide https://review.openstack.org/481383 | 18:04 |
| openstackgerrit | Merged openstack/keystone master: Make the devstack plugin more configurable for federation https://review.openstack.org/484480 | 18:04 |
| *** aojea has joined #openstack-keystone | 18:16 | |
| *** ducttape_ has quit IRC | 18:27 | |
| knikolla | lbragstad: cmurphy: this time it's really fixed https://review.openstack.org/#/c/487202/ | 18:27 |
| knikolla | apache had to be restarted also. | 18:27 |
| lbragstad | knikolla: awesome - looks good | 18:28 |
| amrith | lbragstad https://review.openstack.org/487417 has merged | 18:29 |
| *** aojea has quit IRC | 18:30 | |
| lbragstad | amrith: cool - let me know if you have any other issues with ksa | 18:31 |
| *** ducttape_ has joined #openstack-keystone | 18:31 | |
| *** ducttape_ has quit IRC | 18:33 | |
| *** ducttape_ has joined #openstack-keystone | 18:34 | |
| amrith | thx @lbragstad | 18:41 |
| *** ducttape_ has quit IRC | 18:41 | |
| *** ducttape_ has joined #openstack-keystone | 18:44 | |
| *** prashkre has joined #openstack-keystone | 18:51 | |
| *** aojea has joined #openstack-keystone | 18:56 | |
| openstackgerrit | Lance Bragstad proposed openstack/keystone master: WIP: Remove deprecate default domain id compatibility https://review.openstack.org/487561 | 18:57 |
| *** aojea has quit IRC | 19:01 | |
| *** dave-mccowan has joined #openstack-keystone | 19:02 | |
| *** dave-mccowan has quit IRC | 19:04 | |
| *** nicolasbock has quit IRC | 19:06 | |
| *** jamiec has quit IRC | 19:12 | |
| openstackgerrit | Samuel Pilla proposed openstack/python-keystoneclient master: Add project tags to keystoneclient https://review.openstack.org/481223 | 19:13 |
| *** jamiec has joined #openstack-keystone | 19:19 | |
| *** harlowja has joined #openstack-keystone | 19:21 | |
| *** ducttap__ has joined #openstack-keystone | 19:24 | |
| *** ducttape_ has quit IRC | 19:24 | |
| *** harlowja has quit IRC | 19:25 | |
| *** rcernin has quit IRC | 19:27 | |
| *** bknudson has joined #openstack-keystone | 19:27 | |
| *** bknudson has quit IRC | 19:28 | |
| *** chlong has quit IRC | 19:43 | |
| *** chlong has joined #openstack-keystone | 19:49 | |
| *** chlong has quit IRC | 19:52 | |
| *** chlong has joined #openstack-keystone | 19:53 | |
| prashkre | samueldmq: Hi. Could you please review https://review.openstack.org/#/c/485302/? | 19:55 |
| prashkre | breton:cmurphy: Hi. If you have sometime, could you please help me by reviewing https://review.openstack.org/#/c/485302/. | 19:58 |
| *** blake has quit IRC | 20:07 | |
| *** prashkre has quit IRC | 20:16 | |
| *** raildo has quit IRC | 20:16 | |
| *** raildo has joined #openstack-keystone | 20:17 | |
| openstackgerrit | Lance Bragstad proposed openstack/keystone master: WIP: Remove deprecate default domain id compatibility https://review.openstack.org/487561 | 20:19 |
| openstackgerrit | Lance Bragstad proposed openstack/keystone master: WIP: Remove deprecate default domain id compatibility https://review.openstack.org/487561 | 20:25 |
| lbragstad | knikolla: nice job! https://review.openstack.org/#/c/487202/2 | 20:25 |
| knikolla | lbragstad: thanks! | 20:28 |
| lbragstad | it's nice to see those green again | 20:28 |
| knikolla | yep. were an eyesore. | 20:29 |
| *** sjain has joined #openstack-keystone | 20:37 | |
| *** Drankis has quit IRC | 20:39 | |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add new tags attribute to project https://review.openstack.org/470317 | 20:43 |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add JSON schema validation for project tags https://review.openstack.org/484483 | 20:43 |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add database migration for project tags https://review.openstack.org/484456 | 20:43 |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add policy for project tags https://review.openstack.org/486757 | 20:43 |
| openstackgerrit | Kristi Nikolla proposed openstack/keystone master: WIP - Clean lingering roles when purging mappings https://review.openstack.org/487579 | 20:43 |
| knikolla | lbragstad: thoughts on ^^ | 20:43 |
| openstackgerrit | Samuel Pilla proposed openstack/python-keystoneclient master: Add project tags to keystoneclient https://review.openstack.org/481223 | 20:44 |
| lbragstad | knikolla: i can look | 20:45 |
| *** aojea has joined #openstack-keystone | 20:45 | |
| openstackgerrit | Lance Bragstad proposed openstack/keystone master: Clarify documentation on whitelists and blacklists https://review.openstack.org/487583 | 20:49 |
| *** thorst has quit IRC | 20:50 | |
| openstackgerrit | Samriddhi proposed openstack/keystone master: Improve docs about token persistence drivers https://review.openstack.org/476384 | 20:51 |
| *** sjain has quit IRC | 20:56 | |
| *** catintheroof has quit IRC | 21:01 | |
| openstackgerrit | Eric Fried proposed openstack/keystoneauth master: WIP: get_conf_options can exclude deprecated opts https://review.openstack.org/487590 | 21:04 |
| *** raildo has quit IRC | 21:06 | |
| *** sjain has joined #openstack-keystone | 21:10 | |
| *** sjain has quit IRC | 21:14 | |
| *** chlong has quit IRC | 21:14 | |
| openstackgerrit | Eric Fried proposed openstack/keystoneauth master: get_conf_options can exclude deprecated opts https://review.openstack.org/487590 | 21:16 |
| *** blake has joined #openstack-keystone | 21:19 | |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add new tags attribute to project https://review.openstack.org/470317 | 21:22 |
| *** blake has quit IRC | 21:27 | |
| *** thorst has joined #openstack-keystone | 21:32 | |
| openstackgerrit | Gage Hugo proposed openstack/keystone master: Add new tags attribute to project https://review.openstack.org/470317 | 21:33 |
| *** ducttape_ has joined #openstack-keystone | 21:34 | |
| *** thorst has quit IRC | 21:37 | |
| *** ducttap__ has quit IRC | 21:37 | |
| *** ducttap__ has joined #openstack-keystone | 21:49 | |
| *** ducttape_ has quit IRC | 21:52 | |
| *** edmondsw has quit IRC | 21:58 | |
| *** harlowja has joined #openstack-keystone | 22:00 | |
| *** aojea has quit IRC | 22:23 | |
| cmurphy | knikolla: so it needed apache restarted too? | 22:25 |
| cmurphy | knikolla: i guess that makes some sense because shibboleth | 22:25 |
| *** thorst has joined #openstack-keystone | 22:25 | |
| *** thorst has quit IRC | 22:26 | |
| *** phalmos has joined #openstack-keystone | 22:32 | |
| *** blake has joined #openstack-keystone | 22:34 | |
| rm_work | is there a definitive list of the current options for keytone_authtoken config section for services using the middleware? | 22:36 |
| rm_work | everything i find says something different and/or is super old | 22:36 |
| *** blake has quit IRC | 22:38 | |
| rm_work | https://github.com/openstack/keystonemiddleware/blob/master/keystonemiddleware/auth_token/_opts.py | 22:41 |
| rm_work | may be my best bet i guess lol | 22:41 |
| *** blake has joined #openstack-keystone | 22:41 | |
| breton | rm_work: the options in your link + auth plugins' options | 22:49 |
| rm_work | hmm k | 22:49 |
| rm_work | i'm trying to disable cacheing of tokens | 22:49 |
| rm_work | because it is correct that the inmemory cache is bad | 22:49 |
| rm_work | but i don't want to replace it with memcached either at the moment | 22:49 |
| rm_work | i just want to have *no cache* | 22:49 |
| rm_work | i set the cache time to -1 which it says will disable cacheing | 22:50 |
| rm_work | but it still prints the warning about the inmemory thing everywhere | 22:50 |
| rm_work | i guess it's just not smart enough to realize i disabled cacheing? | 22:50 |
| breton | rm_work: you should probably know that we tried that | 22:52 |
| breton | rm_work: and had to revert the patch | 22:52 |
| breton | rm_work: https://review.openstack.org/#/c/271352/ | 22:52 |
| rm_work | hmm | 22:52 |
| rm_work | it was that bad? | 22:52 |
| rm_work | why is it THAT bad | 22:52 |
| breton | rm_work: i don't know :) you can try reverting that revert if you feel brave | 22:53 |
| rm_work | breton: so what it seems i'm running into (with the octavia api), is that i get slow requests every first request and every 5 minutes, which SEEMS to be related to the token cache | 22:53 |
| rm_work | if i mess with the cache time and other related settings, it seems to change accordingly | 22:54 |
| breton | rm_work: it's probably cache, right | 22:54 |
| rm_work | but it's like ... 10s request every time it misses the cache for a token | 22:54 |
| rm_work | which is bonkers | 22:54 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystone master: Updated from global requirements https://review.openstack.org/487465 | 22:55 |
| breton | rm_work: is cache enabled for keystone? | 22:59 |
| rm_work | on the keystone server itself? I don't actually know, but i can find out | 22:59 |
| breton | rm_work: yes | 23:01 |
| rm_work | we do not have cache enabled in keystone | 23:02 |
| *** ducttap__ has quit IRC | 23:03 | |
| *** ducttape_ has joined #openstack-keystone | 23:05 | |
| *** blake has quit IRC | 23:05 | |
| *** edmondsw has joined #openstack-keystone | 23:06 | |
| *** ducttape_ has quit IRC | 23:09 | |
| *** ducttape_ has joined #openstack-keystone | 23:10 | |
| *** edmondsw has quit IRC | 23:11 | |
| harlowja | breton does caching do anything if u turn it on and say still use the sql token backend | 23:19 |
| harlowja | does using memcache offer any benefits unless u drop that token backend? | 23:20 |
| harlowja | (and switch to the memcache backend) | 23:20 |
| * harlowja wondering cause rm_work and i use the sql backend | 23:20 | |
| harlowja | and we aren't turning on caching, but what would caching offer if we left the sql token backend as is | 23:20 |
| harlowja | like i think we discovered that keystonemiddleware is all sorts of messed up due to the forced-on-memory-cache | 23:21 |
| harlowja | that was super nice to figure out | 23:21 |
| harlowja | lol | 23:21 |
| harlowja | so pretty much memcache is a requirement of anything that uses keystonemiddleware (ie everything) | 23:22 |
| *** ducttape_ has quit IRC | 23:22 | |
| harlowja | ^ not saying i disagree with that, just ummm, was there every messaging to <anyone> about that? | 23:22 |
| harlowja | lol | 23:22 |
| harlowja | like perhaps https://docs.openstack.org/keystonemiddleware/latest/ should say ' u have to really use this with memcache' in bold | 23:23 |
| harlowja | imho it shouldn't even be optional | 23:23 |
| harlowja | it should default to 'u must have memcache' | 23:23 |
| *** thorst has joined #openstack-keystone | 23:27 | |
| rm_work | it does kinda | 23:28 |
| rm_work | when you spin up and a request happens: | 23:28 |
| rm_work | WARNING keystonemiddleware.auth_token [-] Using the in-process token cache is deprecated as of the 4.2.0 release and may be removed in the 5.0.0 release or the 'O' development cycle. The in-process cache causes inconsistent results and high memory usage. When the feature is removed the auth_token middleware will not cache tokens by default which may result in performance issues. It is recommended to use memcache for the | 23:29 |
| rm_work | auth_token token cache by setting the memcached_servers option. | 23:29 |
| harlowja | ya, that's a message say ummmm a little late in the whole game no? | 23:29 |
| rm_work | heh | 23:29 |
| harlowja | u deployed all the things, ummm, and u thought it was all goign to work | 23:30 |
| harlowja | but ya, go back to start and add memcache | 23:30 |
| harlowja | lol | 23:30 |
| *** oomichi has quit IRC | 23:32 | |
| *** thorst has quit IRC | 23:32 | |
| harlowja | unless u like ` inconsistent results and high memory usage.` | 23:34 |
| harlowja | i mean who doesn't like ^ | 23:34 |
| harlowja | lol | 23:34 |
| *** oomichi has joined #openstack-keystone | 23:34 | |
| *** efried is now known as efried_zzz | 23:35 | |
| *** phalmos has quit IRC | 23:37 | |
| breton | harlowja: caching matters with any backend | 23:40 |
| breton | harlowja: we cache a lot of stuff in very interesting ways | 23:40 |
| harlowja | lol | 23:40 |
| harlowja | not sure if impressed or scared | 23:41 |
| harlowja | still deciding, lol | 23:41 |
| breton | harlowja: users, projects, revocations, even full tokens can be cached | 23:41 |
| breton | > imho it shouldn't even be optional > it should default to 'u must have memcache' | 23:42 |
| breton | we can't do that yet because keystonemiddleware is still used by eventlet-based servers | 23:42 |
| openstackgerrit | Merged openstack/keystoneauth master: get_conf_options can exclude deprecated opts https://review.openstack.org/487590 | 23:42 |
| breton | and they work badly with memcache client library | 23:42 |
| harlowja | which lib? | 23:42 |
| harlowja | want me to make a lib that works, lol | 23:42 |
| breton | python-memcached | 23:42 |
| harlowja | don't use that one? | 23:42 |
| harlowja | lol | 23:42 |
| breton | harlowja: write a new one for us and i will be the first to +2 :p | 23:43 |
| harlowja | isn't that pymemcache? | 23:43 |
| harlowja | pretty sure i have some contributions into that, lol | 23:43 |
| breton | pymemcache is good, yes | 23:44 |
| harlowja | so ya, ummmm, whats the hold up then, lol | 23:46 |
| openstackgerrit | Nicolas Helgeson proposed openstack/keystone master: Add new tags attribute to project https://review.openstack.org/470317 | 23:47 |
| *** thorst has joined #openstack-keystone | 23:50 | |
| morgan | harlowja: convert to use oslo.cache and i can write a pymemcache impl in ~10m | 23:52 |
| morgan | (ok ok 1hr w/ tests) | 23:52 |
| harlowja | get er done | 23:52 |
| harlowja | lol | 23:52 |
| morgan | in all seriousness, oslo.cache would make a huge difference and make it easier to convert to pymemcache | 23:54 |
| harlowja | i'd assume so | 23:54 |
| morgan | the biggest issue is actually the swift cache object that is being passed to ksm sometimes. | 23:54 |
| harlowja | https://github.com/openstack/keystonemiddleware/blob/master/keystonemiddleware/auth_token/_cache.py#L59 scared me though, lol | 23:54 |
| harlowja | at least in the keystone middleware code, that scared me | 23:55 |
| harlowja | so i need to go to PTSD or something first | 23:55 |
| openstackgerrit | Merged openstack/keystone master: In the devstack plugin, restart keystone after modifying conf https://review.openstack.org/487202 | 23:55 |
| morgan | well, you do get inconsistent results | 23:55 |
| harlowja | lol | 23:55 |
| morgan | some processes may cache the token, some may not | 23:55 |
| morgan | a token may be valid in one request and not the next to the same endpoint | 23:55 |
| morgan | and then valid again | 23:55 |
| harlowja | yup | 23:55 |
| morgan | and with a ton of requests you burn spots in dict() for the sake of a maybe-use-sometime and a horrible performance bottleneck | 23:56 |
| harlowja | so ya, i need to recover from PTSD first for a little | 23:56 |
| harlowja | lol | 23:56 |
| *** ducttape_ has joined #openstack-keystone | 23:59 | |
| *** ducttape_ has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!