| jamielennox | gyee: replied | 00:00 |
|---|---|---|
| gyee | jamielennox, ok, good with me then | 00:01 |
| gyee | jamielennox, so keystoneclient is going to oslo? | 00:02 |
| gyee | or was I hearing it wrong this morning | 00:02 |
| jamielennox | gyee: submit it to oslo? | 00:02 |
| *** andreaf has joined #openstack-keystone | 00:03 | |
| jamielennox | i don't think so - if anything i would like oslo.apiclient to die | 00:05 |
| gyee | jamielennox, eavesdrop.openstack.org/meetings/keystone/2014/keystone.2014-03-18-18.01.log.txt | 00:05 |
| gyee | <ayoung> our client stuff maybe can be merged with the Oslo common client | 00:05 |
| gyee | ayoung, ^^^ can you clarify? | 00:05 |
| jamielennox | gyee: oh i think i missed like the first 10 min | 00:06 |
| openstackgerrit | A change was merged to openstack/keystone: Move test .conf files to keystone/tests/config_files https://review.openstack.org/79526 | 00:07 |
| gyee | jamielennox, looks like dolphm just drop you a bunch of coments :) | 00:08 |
| jamielennox | yea | 00:08 |
| *** wchrisj has quit IRC | 00:08 | |
| gyee | jamielennox, this one is interesting https://review.openstack.org/#/c/74720/ | 00:09 |
| jamielennox | gyee: turns out there is a much bigger problem behind that one | 00:09 |
| gyee | I also think connection pool should be done at the application layer | 00:10 |
| morganfainberg | dolphm, ok fixed, running tests again and will post as soon as that's done | 00:14 |
| openstackgerrit | Jamie Lennox proposed a change to openstack/python-keystoneclient: Handle URLs via the session and auth_plugins https://review.openstack.org/60752 | 00:19 |
| jamielennox | dolphm: Done | 00:19 |
| *** leseb_ has joined #openstack-keystone | 00:22 | |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/keystone: Do not expose internal data on UnexpectedError https://review.openstack.org/81385 | 00:27 |
| *** leseb_ has quit IRC | 00:27 | |
| morganfainberg | wow, 81k reviews | 00:27 |
| morganfainberg | 81k+ | 00:27 |
| morganfainberg | erm patchsets | 00:27 |
| gyee | morganfainberg, for icehouse? | 00:33 |
| morganfainberg | gyee, total | 00:33 |
| morganfainberg | in gerrit | 00:33 |
| gyee | wow | 00:33 |
| morganfainberg | 81385 as of my review ^ | 00:33 |
| morganfainberg | erm patchset | 00:33 |
| gyee | all those rechecks :) | 00:33 |
| gyee | killing jenkins | 00:33 |
| morganfainberg | yeah | 00:33 |
| morganfainberg | 7+hr gate | 00:34 |
| *** dpk__ has joined #openstack-keystone | 00:41 | |
| dpk__ | hello, can we add meta information to an openstack user? | 00:41 |
| *** andreaf has quit IRC | 00:41 | |
| *** wchrisj has joined #openstack-keystone | 00:52 | |
| *** marcoemorais has quit IRC | 00:52 | |
| *** devlaps has quit IRC | 00:52 | |
| *** nkinder has joined #openstack-keystone | 01:02 | |
| *** derek_c has joined #openstack-keystone | 01:15 | |
| *** dpk__ has quit IRC | 01:19 | |
| *** wchrisj has quit IRC | 01:25 | |
| *** richm has quit IRC | 01:40 | |
| openstackgerrit | Richard Megginson proposed a change to openstack/keystone: better handling for empty/None ldap values https://review.openstack.org/76002 | 01:42 |
| openstackgerrit | Zane Bitter proposed a change to openstack/python-keystoneclient: Sync openstack.common.memorycache with oslo-incubator https://review.openstack.org/81414 | 01:50 |
| openstackgerrit | Zane Bitter proposed a change to openstack/python-keystoneclient: Don't register memcached_server option in two groups https://review.openstack.org/81415 | 01:51 |
| *** wchrisj has joined #openstack-keystone | 02:00 | |
| *** dstanek has quit IRC | 02:02 | |
| *** wchrisj has quit IRC | 02:04 | |
| *** dstanek has joined #openstack-keystone | 02:06 | |
| *** zhiyan_ is now known as zhiyan | 02:09 | |
| *** shakamunyi has joined #openstack-keystone | 02:28 | |
| *** devlaps has joined #openstack-keystone | 02:29 | |
| *** gtt116 has joined #openstack-keystone | 02:31 | |
| *** gtt116 has quit IRC | 02:33 | |
| *** chandan_kumar has joined #openstack-keystone | 02:37 | |
| *** amcrn has quit IRC | 02:49 | |
| *** mberlin has joined #openstack-keystone | 02:57 | |
| *** mberlin1 has quit IRC | 02:59 | |
| *** gyee has quit IRC | 02:59 | |
| *** harlowja is now known as harlowja_away | 03:23 | |
| *** devlaps has quit IRC | 03:30 | |
| *** chandan_kumar has quit IRC | 03:41 | |
| *** thiagop has quit IRC | 03:58 | |
| *** thiagop has joined #openstack-keystone | 04:00 | |
| *** amcrn has joined #openstack-keystone | 04:53 | |
| *** wchrisj has joined #openstack-keystone | 05:00 | |
| *** andreaf has joined #openstack-keystone | 05:35 | |
| *** wchrisj has quit IRC | 05:37 | |
| * jamielennox is sick of oslo.config | 05:40 | |
| *** andreaf has quit IRC | 05:48 | |
| *** saju_m has joined #openstack-keystone | 05:58 | |
| openstackgerrit | A change was merged to openstack/keystone: Uses explicit imports for _ https://review.openstack.org/58766 | 06:06 |
| *** stevemar has quit IRC | 06:08 | |
| morganfainberg | jamielennox, why dare i ask? | 06:23 |
| jamielennox | morganfainberg: because it acts differently depending on whether you read config options from a file or stub them from tests | 06:25 |
| morganfainberg | you mean the config fixture? | 06:26 |
| jamielennox | you can preload data from a file, then register the option and read it straight away | 06:26 |
| jamielennox | if you don't register the option though you can't set an override | 06:26 |
| morganfainberg | ok sure. | 06:26 |
| morganfainberg | sure, but if you know how the option will be registered you can register it w/ the conf fixture | 06:26 |
| jamielennox | so if you have a function that is generating the names of the config options you want to use then you can't stub it | 06:26 |
| jamielennox | morganfainberg: right, but then when the function goes to register than options you get a conflict cause it was registered twice | 06:27 |
| morganfainberg | not if it's registered the same exact way | 06:27 |
| morganfainberg | if it's registered with the same parameters you can register it as many times as you want | 06:27 |
| morganfainberg | otherwise you get the conflict | 06:28 |
| jamielennox | morganfainberg: i don't have that choice | 06:28 |
| jamielennox | it's generating the options to use for config | 06:28 |
| morganfainberg | jamielennox, how do you not have that choice? | 06:28 |
| * morganfainberg is confused. | 06:28 | |
| jamielennox | morganfainberg: i'm looking at loading auth plugins from a conf file | 06:28 |
| jamielennox | so the options i need aren't known until the auth plugin is loaded | 06:29 |
| morganfainberg | ok so you know how you'd mangle the name up | 06:29 |
| morganfainberg | and you know what the options are going to be | 06:29 |
| morganfainberg | or you mean you need the auth_plugin to register the opt? | 06:29 |
| jamielennox | morganfainberg: kind of, the auth plugin gives a list of parameters it will accept | 06:30 |
| jamielennox | the function turns that list into a list of cfg.StrOpt | 06:30 |
| morganfainberg | ok, so you'd test it by doing the same work to stub out the opt | 06:30 |
| morganfainberg | load that info, turn it into str opt | 06:30 |
| morganfainberg | then register | 06:31 |
| jamielennox | right, but because the option is generated from a list when you go and register it again i get a conflict | 06:31 |
| morganfainberg | remember the conf_fixture will destroy all opts registered with it between tests | 06:31 |
| morganfainberg | then you're somehow getting mismatched options. | 06:31 |
| jamielennox | i'ts not smart enough to check that the option is the same name/type it checks that "optA is optB" | 06:32 |
| jamielennox | so because the opt instance is different it conflicts | 06:32 |
| morganfainberg | jamielennox, http://paste.openstack.org/show/73798/ | 06:34 |
| jamielennox | morganfainberg: hmm | 06:34 |
| jamielennox | morganfainberg: damn - i need to get it exact | 06:35 |
| morganfainberg | jamielennox, http://paste.openstack.org/show/73799/ as you can see if it's not exact it does fail | 06:35 |
| morganfainberg | jamielennox, yeah. i've been working in that code for a bit | 06:36 |
| jamielennox | ok, so i was ignoring some of the descriptions in places | 06:36 |
| morganfainberg | ah, yeah the comparison is very picky | 06:36 |
| jamielennox | morganfainberg: ok, that gives me something to go on | 06:43 |
| morganfainberg | jamielennox, happy to help | 06:43 |
| jamielennox | morganfainberg: i'm not necessarily any less annoyed with it, but at least i know what's up | 06:43 |
| jamielennox | morganfainberg: test pass :) | 06:47 |
| morganfainberg | i need to propose a change to oslo.config that lets me unregister groups | 06:47 |
| morganfainberg | for the same reason (testing) i think... | 06:47 |
| openstackgerrit | Jamie Lennox proposed a change to openstack/python-keystoneclient: Plugin loading from config objects https://review.openstack.org/79542 | 07:02 |
| *** flaper87|afk is now known as flaper87 | 07:14 | |
| *** morganfainberg is now known as morganfainberg_Z | 07:43 | |
| *** derek_c has quit IRC | 07:54 | |
| *** shakamunyi has quit IRC | 08:07 | |
| *** leseb_ has joined #openstack-keystone | 08:15 | |
| *** leseb_ has quit IRC | 08:22 | |
| *** leseb_ has joined #openstack-keystone | 08:22 | |
| *** uvirtbot has joined #openstack-keystone | 08:25 | |
| *** leseb_ has quit IRC | 08:27 | |
| openstackgerrit | Jamie Lennox proposed a change to openstack/python-keystoneclient: Plugin loading from config objects https://review.openstack.org/79542 | 08:29 |
| *** shakamunyi has joined #openstack-keystone | 08:33 | |
| *** henrynash has joined #openstack-keystone | 08:34 | |
| *** chandankumar_ has quit IRC | 08:37 | |
| *** shakamunyi has quit IRC | 08:38 | |
| *** chandan_kumar has joined #openstack-keystone | 08:39 | |
| *** amcrn has quit IRC | 08:42 | |
| *** henrynash has quit IRC | 08:43 | |
| *** andreaf has joined #openstack-keystone | 08:54 | |
| *** henrynash has joined #openstack-keystone | 09:20 | |
| *** leseb_ has joined #openstack-keystone | 09:33 | |
| *** shakamunyi has joined #openstack-keystone | 09:34 | |
| *** leseb_ has quit IRC | 09:37 | |
| *** shakamunyi has quit IRC | 09:38 | |
| *** shakamunyi has joined #openstack-keystone | 10:35 | |
| *** shakamunyi has quit IRC | 10:39 | |
| *** flaper87 is now known as flaper87|afk | 11:17 | |
| openstackgerrit | henry-nash proposed a change to openstack/keystone: Make domain_id immutable by default https://review.openstack.org/81499 | 11:28 |
| *** flaper87|afk is now known as flaper87 | 12:08 | |
| *** chandan_kumar has quit IRC | 12:23 | |
| *** henrynash has quit IRC | 12:33 | |
| *** leseb_ has joined #openstack-keystone | 12:33 | |
| *** shakamunyi has joined #openstack-keystone | 12:36 | |
| *** leseb_ has quit IRC | 12:40 | |
| *** leseb_ has joined #openstack-keystone | 12:41 | |
| *** raildo_ has joined #openstack-keystone | 12:41 | |
| *** shakamunyi has quit IRC | 12:41 | |
| *** leseb_ has quit IRC | 12:42 | |
| *** leseb_ has joined #openstack-keystone | 12:42 | |
| *** dims has quit IRC | 12:50 | |
| *** chandan_kumar has joined #openstack-keystone | 12:56 | |
| dstanek | hmmm....i don't see any tests for the templated catalog | 13:30 |
| dstanek | dolphm: will we not be supported the templated catalog after v2 is gone? | 13:30 |
| *** bknudson has joined #openstack-keystone | 13:32 | |
| *** henrynash has joined #openstack-keystone | 13:37 | |
| dolphm | dstanek: i honestly don't know how many people are using it, but it's completely broken for v3.. | 13:38 |
| dstanek | dolphm: yeah, i noticed that because i just fixed a bug in get_catalog and noticed that get_v3_catalog raises not implemented | 13:39 |
| *** browne has joined #openstack-keystone | 13:46 | |
| *** david-lyle has quit IRC | 13:48 | |
| openstackgerrit | David Stanek proposed a change to openstack/keystone: Ignore broken endpoints in get_v3_catalog https://review.openstack.org/81527 | 13:52 |
| openstackgerrit | David Stanek proposed a change to openstack/keystone: Ingore broken endpoints in get_catalog https://review.openstack.org/81528 | 13:52 |
| *** wchrisj has joined #openstack-keystone | 13:54 | |
| *** lbragstad__ is now known as lbragstad | 13:58 | |
| *** thiagop has quit IRC | 14:00 | |
| henrynash | dolphm: can you remind me, do we support upgrading from Grizzly to IceHouse directly (i.e. without upgrading to the interim of Havana)? | 14:02 |
| dolphm | henrynash: in terms of migrations? yes | 14:03 |
| henrynash | dolphm: is that true across the OpenStack projects do you know, that you can upgrade from N to N+2 ? | 14:03 |
| *** thiagop has joined #openstack-keystone | 14:05 | |
| dstanek | dolphm: does this impact string freeze? https://review.openstack.org/#/c/81385/ | 14:06 |
| dolphm | henrynash: for integrated projects, i believe that's the community's goal | 14:06 |
| henrynash | dolphm: ok, thx, that's what I thought too | 14:07 |
| dolphm | dstanek: ooh... | 14:07 |
| dolphm | dstanek: it only introduces a new string, so it should be fine according to my understanding of string freeze | 14:07 |
| dolphm | dstanek: but i believe morganfainberg_Z mentioned that a change failed gating yesterday because it changed a string? i didn't get a chance to follow up on that | 14:08 |
| dstanek | dolphm: ok, i thought i'd ask :-) | 14:09 |
| *** chandan_kumar has quit IRC | 14:13 | |
| *** stevemar has joined #openstack-keystone | 14:14 | |
| *** chandan_kumar has joined #openstack-keystone | 14:17 | |
| *** shakamunyi has joined #openstack-keystone | 14:18 | |
| dolphm | dstanek: so, i asked in -dev, and i'll be pinging the i18n folks asap | 14:24 |
| *** henrynash has quit IRC | 14:25 | |
| dolphm | dstanek: go ahead and approve... i'm going to ask for forgiveness, give them a heads up on this specific change, and find out what they want the process to be moving forward | 14:25 |
| dolphm | dstanek: we've never actually *had* to change a string during string freeze, so i've never been through an exception process before | 14:26 |
| dstanek | dolphm: ok. once jenkins weighs in i'll +2 it | 14:26 |
| dstanek | dolphm: about your change here: https://review.openstack.org/#/c/71674/ | 14:27 |
| dstanek | dolphm: where would be the best place for this kind of documentation? | 14:27 |
| *** lbragstad has quit IRC | 14:28 | |
| dolphm | dstanek: that probably needs to go into keystone.common.config now? and if keystone-manage pki_setup --help is supported, it should be doc'd there too | 14:28 |
| dstanek | dolphm: k. going to tack that onto my todo list. | 14:29 |
| *** daneyon has joined #openstack-keystone | 14:29 | |
| *** dims has joined #openstack-keystone | 14:43 | |
| *** daneyon has quit IRC | 14:46 | |
| *** daneyon has joined #openstack-keystone | 14:47 | |
| *** thedodd has joined #openstack-keystone | 14:49 | |
| *** devlaps has joined #openstack-keystone | 14:51 | |
| *** dolphm changes topic to "[ Icehouse RC blockers https://launchpad.net/keystone/+milestone/icehouse-rc1 ][ Icehouse RC Target Date: March 27th, 2014 ][ No new strings during string freeze (dolphm was wrong!) - ping dolphm concerning exceptions ]" | 14:52 | |
| *** shakamunyi has quit IRC | 14:52 | |
| *** shakamunyi has joined #openstack-keystone | 14:52 | |
| *** david-lyle has joined #openstack-keystone | 14:58 | |
| *** lbragstad has joined #openstack-keystone | 14:59 | |
| *** daneyon has quit IRC | 15:04 | |
| *** daneyon has joined #openstack-keystone | 15:04 | |
| dolphm | dstanek: pinged the -i18n mailing list concerning the above review (it also failed jenkins, so i did a recheck) | 15:04 |
| dstanek | dolphm: any response? | 15:05 |
| dolphm | dstanek: not yet - just sent it | 15:05 |
| dolphm | dstanek: sounds like we just need to be conservative about additions, but i'll do my best to keep them informed and if they have any feedback ("too many new strings!", etc) i'll pass it along | 15:06 |
| dstanek | do we have more than just the one so far? | 15:06 |
| dolphm | dstanek: i just looked through the open RC-targeted changes, and that's the only one | 15:07 |
| dolphm | dstanek: if there are more in other changes we want to land, just give me a heads up | 15:07 |
| ayoung | dolphm, so, the issue with deleting tokens from Horizon has me thinking that we have a disconnect in how Horizon views things and Keystone. Horizon has the "web" view of things that, if the user continues to be active, you should extend the session. Keystone has implicitly supported this with long lived tokens, but the shorter ones are going to break it | 15:07 |
| ayoung | what if... | 15:07 |
| ayoung | we had a "web token" | 15:07 |
| ayoung | and a web token is unscoped, can only be extended...and...must be bound to another authentication mechanism | 15:08 |
| dolphm | ayoung: morganfainberg_Z suggested 2 hours as a more realistic default token duration, i suspect for that reason | 15:08 |
| ayoung | dolphm, if the session times out, they will be prompted for userid and password anyway | 15:09 |
| ayoung | most people have that cached in their browser | 15:09 |
| ayoung | so right now at the one hour mark, they will get another redirect to the login page | 15:10 |
| ayoung | its not a big deal...I'm more thinking in terms of how to get Keystone and Horizon to have the same view of authorization and authentication lifespans | 15:10 |
| *** flaper87 is now known as flaper87|afk | 15:11 | |
| *** gokrokve has joined #openstack-keystone | 15:11 | |
| ayoung | lets say we make token lifespan 10 minutes, and horizon were to make session timeout 10 minutes | 15:11 |
| ayoung | token is unscoped, so it can only be used with keystone, but the "expires" time gets extended each time it gets passed back. If we did that with no additional check, it could lead to abuse. | 15:13 |
| ayoung | but, if the "extend" call had to come from the same source as the original token request (IP address check?) would that be OK? | 15:13 |
| dstanek | ayoung: this sounds similar to why oauth refresh tokens were created | 15:15 |
| ayoung | dstanek, yep. And maybe oauth is the right solution for Horizon, | 15:17 |
| ayoung | dstanek, I'd love it if we coud redirect to Keystone directly, get a new token, and then redirect back to horizon | 15:17 |
| *** flaper87|afk is now known as flaper87 | 15:18 | |
| ayoung | Then the Keystone operation could be done using a crypto-secure method | 15:18 |
| *** pcargnel_ has joined #openstack-keystone | 15:31 | |
| openstackgerrit | Lance Bragstad proposed a change to openstack/keystone: Enforce required parameters for V3 Regions https://review.openstack.org/76444 | 15:32 |
| *** leseb_ has quit IRC | 15:32 | |
| *** leseb has joined #openstack-keystone | 15:32 | |
| *** leseb has quit IRC | 15:39 | |
| *** leseb has joined #openstack-keystone | 15:40 | |
| *** leseb has quit IRC | 15:44 | |
| *** richm has joined #openstack-keystone | 15:48 | |
| openstackgerrit | Steve Martinelli proposed a change to openstack/keystone: POC - Add openID Connect auth plugin https://review.openstack.org/61662 | 15:54 |
| *** chandankumar_ has joined #openstack-keystone | 15:54 | |
| *** chandan_kumar has quit IRC | 15:56 | |
| *** marekd|away is now known as marekd | 15:58 | |
| *** dims has quit IRC | 15:59 | |
| *** henrynash has joined #openstack-keystone | 16:02 | |
| marekd | dstanek: https://review.openstack.org/#/c/81277/ - replied to your comments. | 16:07 |
| marekd | dstanek: i don't really want to get into academic discussions where a method should be placed, so if you really want to, I can move this serialized test directly to test_v3_federation. | 16:08 |
| marekd | dstanek: im not sure whether you asked for that or just stated this loudly :-) | 16:09 |
| *** flaper87 is now known as flaper87|afk | 16:11 | |
| *** marcoemorais has joined #openstack-keystone | 16:13 | |
| openstackgerrit | Marek Denis proposed a change to openstack/keystone: Store groups ids objects list in the OS-FEDERATION object. https://review.openstack.org/81277 | 16:19 |
| openstackgerrit | Ilya Pekelny proposed a change to openstack/keystone: Sync test_migrations https://review.openstack.org/80618 | 16:25 |
| openstackgerrit | Ilya Pekelny proposed a change to openstack/keystone: Comparision of database models and migrations. https://review.openstack.org/80630 | 16:25 |
| *** daneyon has quit IRC | 16:33 | |
| dstanek | marekd: i was wondering until i saw that you had to another class to inherit from to get it | 16:33 |
| *** leseb has joined #openstack-keystone | 16:33 | |
| dstanek | marekd: that adds a bunch more tests (basically duplicated) i think | 16:33 |
| *** jagee has joined #openstack-keystone | 16:34 | |
| marekd | dstanek: so move this test to federation tests... | 16:36 |
| *** shakamunyi has quit IRC | 16:40 | |
| *** gyee has joined #openstack-keystone | 16:43 | |
| openstackgerrit | Marek Denis proposed a change to openstack/keystone: Store groups ids objects list in the OS-FEDERATION object. https://review.openstack.org/81277 | 16:48 |
| marekd | dstanek: ^^ moved | 16:48 |
| *** dstanek_afk has joined #openstack-keystone | 16:48 | |
| *** dstanek is now known as Guest64110 | 16:49 | |
| *** dstanek_afk is now known as dstanek | 16:49 | |
| dstanek | marekd: you may have an issue there because you are using methods from test_v3_federation | 16:49 |
| *** browne has quit IRC | 16:50 | |
| *** leseb has quit IRC | 16:50 | |
| marekd | dstanek: tox -ep27 test_v3_federation passes.... | 16:50 |
| *** leseb has joined #openstack-keystone | 16:51 | |
| marekd | -epy27 | 16:51 |
| *** tstevenson has quit IRC | 16:52 | |
| *** stevemar has quit IRC | 16:53 | |
| dstanek | marekd: i misread what you said :-) just +2ed it. thanks for the quick updates. | 16:54 |
| *** amcrn has joined #openstack-keystone | 16:55 | |
| marekd | dstanek: thanks! | 16:56 |
| *** browne has joined #openstack-keystone | 16:58 | |
| *** andreaf has quit IRC | 16:59 | |
| *** chandan_kumar has joined #openstack-keystone | 17:05 | |
| *** jamielennox is now known as jamielennox|away | 17:07 | |
| *** harlowja_away is now known as harlowja | 17:09 | |
| *** morganfainberg_Z is now known as morganfainberg | 17:15 | |
| *** leseb has quit IRC | 17:16 | |
| *** leseb has joined #openstack-keystone | 17:16 | |
| *** dstanek has quit IRC | 17:16 | |
| *** chandan_kumar has quit IRC | 17:17 | |
| *** daneyon has joined #openstack-keystone | 17:21 | |
| *** leseb has quit IRC | 17:22 | |
| *** dims has joined #openstack-keystone | 17:23 | |
| morganfainberg | henrynash, nova only does one release | 17:23 |
| morganfainberg | henrynash, afaik | 17:23 |
| henrynash | morganfainberq: agh, right | 17:23 |
| morganfainberg | henrynash, we did upgrades recently skipping a release, requires checking out the missing release for nova and doing migrate, then doing the next one as well | 17:24 |
| henrynash | morganfainberg: yep, can imagine how to do that... | 17:26 |
| *** harlowja has quit IRC | 17:26 | |
| *** stevemar has joined #openstack-keystone | 17:26 | |
| morganfainberg | dolphm, going to press +A on https://review.openstack.org/#/c/81385/ since Jenkins said "good" unless there is a reason to wait | 17:30 |
| morganfainberg | dolphm, oh and dstanek is right, about string freeze =/ | 17:31 |
| morganfainberg | should i respin that and leave %(exception)s in? | 17:31 |
| morganfainberg | just ensure it's blank in _build_message? | 17:32 |
| *** dstanek has joined #openstack-keystone | 17:32 | |
| morganfainberg | dstanek, you're right on stringfreeze btw. | 17:32 |
| morganfainberg | dstanek, need dolphm to weigh in if we need to fix that. | 17:32 |
| morganfainberg | dstanek, the right answer might be to leave %(exception)s in and just ensure it's empty. | 17:33 |
| dstanek | He is communicating with the i18n team about it | 17:33 |
| morganfainberg | cool | 17:33 |
| morganfainberg | dstanek, the string that caused gate failure was because _ wasn't imported in trust controller | 17:35 |
| morganfainberg | dstanek, we added a new string to a file that benefited from the magic _ patch | 17:35 |
| morganfainberg | dstanek, so the patch that make _ imports explicit failed | 17:35 |
| morganfainberg | effectively it was an order of merge/rebase issue | 17:36 |
| morganfainberg | not a string check | 17:36 |
| dstanek | Was that merged already? | 17:36 |
| *** harlowja has joined #openstack-keystone | 17:36 | |
| morganfainberg | all that stuff has gone through | 17:36 |
| morganfainberg | _ is now expliticly imported | 17:36 |
| morganfainberg | looks like all the test conf stuff went through too | 17:37 |
| morganfainberg | and the ENV test config is gating | 17:37 |
| openstackgerrit | A change was merged to openstack/keystone: Rename scope_to_bad_project() to test_scope_to_bad_project(). https://review.openstack.org/81395 | 17:37 |
| dstanek | Yay! I didn't realize that all of that is in | 17:37 |
| dolphm | morganfainberg: dstanek: just +A'd it myself | 17:37 |
| morganfainberg | dolphm, okie | 17:37 |
| openstackgerrit | A change was merged to openstack/keystone: Make LIVE Tests configurable with ENV https://review.openstack.org/80409 | 17:37 |
| morganfainberg | dstanek, ^ | 17:38 |
| dolphm | morganfainberg: dstanek: i sent an email to openstack-i18n@ list with a pointer to it | 17:38 |
| morganfainberg | dstanek, on the topic of LIVE test changes | 17:38 |
| morganfainberg | dolphm, cool | 17:38 |
| dolphm | morganfainberg: i had no clue that *new* strings violated string freeze -- according to my understanding, string freeze was all about not wasting translators time | 17:39 |
| dolphm | morganfainberg: giving them more work was fair game ;) | 17:39 |
| morganfainberg | dolphm, my understanding was logs were ok exceptions weren't | 17:39 |
| morganfainberg | dolphm, i guess we had different understandings | 17:39 |
| dolphm | morganfainberg: anyway, it sounds like horizon was first on their priority list anyway, and they're going to start gating stable branches against string changes too | 17:40 |
| morganfainberg | dolphm, interesting | 17:41 |
| dolphm | morganfainberg: i assume that can be applied to master branches during string freeze next cycle :) | 17:42 |
| morganfainberg | dolphm, :) | 17:42 |
| dolphm | can anyone that speaks ldap fluently double check my understanding here? https://review.openstack.org/#/c/78521/4/keystone/common/ldap/core.py | 17:45 |
| dolphm | (cc- richm: ^) | 17:45 |
| lbragstad | bknudson: quick question on https://review.openstack.org/#/c/81040/2. What happens if some where down the line we try validating the API for users? That would lie in the controllers I think, but wouldn't that prevent someone from including additional attribute mappings? | 17:45 |
| *** dstanek is now known as dstanek_afk | 17:46 | |
| bknudson | lbragstad: the user api now allows any attributes if using sql | 17:46 |
| richm | dolphm: updated | 17:47 |
| bknudson | lbragstad: description is an attribute defined in the spec. | 17:47 |
| dolphm | richm: thanks! i'll post a revised patch to update that inline comment in a minute | 17:49 |
| lbragstad | bknudson: ok, cool | 17:52 |
| lbragstad | just wanted to double check, thanks! | 17:52 |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Ability to turn off ldap referral chasing https://review.openstack.org/78521 | 17:54 |
| *** daneyon has quit IRC | 17:54 | |
| dolphm | bknudson: richm: revised https://review.openstack.org/#/c/78521/5/keystone/common/ldap/core.py | 17:55 |
| dolphm | lbragstad: ^ | 17:55 |
| richm | dolphm: Did you still want to log some sort of message if there were only referral objects and no entry objects in the search results? | 17:56 |
| lbragstad | I was thinking that too | 17:57 |
| dolphm | richm: i was thinking you'd be skipping every result in that case, but it sounds like that's not very likely? | 17:57 |
| lbragstad | at line 601 | 17:57 |
| openstackgerrit | Steve Martinelli proposed a change to openstack/keystone: Add user_id when calling populate_roles_for_groups https://review.openstack.org/81587 | 17:58 |
| richm | dolphm: yes, you would be skipping every result in that case - It's not likely with open source ldap servers, but AD uses referrals/continuation references quite a lot | 17:58 |
| bknudson | if your server is returning referrals and you're not following them then you're not getting all the results that would be expected. | 17:59 |
| richm | and you need to know what you need to do in order to get all of the results you are expecting | 18:00 |
| *** daneyon has joined #openstack-keystone | 18:00 | |
| bknudson | seems like keystone just doesn't work with referrals... unless you're able to use anonymous bind. | 18:01 |
| *** amcrn has quit IRC | 18:02 | |
| morganfainberg | bknudson, i think python-ldap has issues with it | 18:02 |
| dolphm | bknudson: richm: i'm fine to skip the extra log message for the moment, considering it's a bit of an edge case and would introduce another new translatable string :P | 18:17 |
| dolphm | bknudson: richm: happy to propose a second patch though, preferably for juno | 18:18 |
| *** gokrokve has quit IRC | 18:18 | |
| bknudson | dolphm: the only way they'll get referrals back is if they configure their system to not chase referrals | 18:19 |
| bknudson | and if you do that then they should be expecting to get partial results. | 18:19 |
| bknudson | so not sure if a log is required. | 18:19 |
| bknudson | might be nice to have a debug log that the results included a referral that wasn't resolved | 18:20 |
| morganfainberg | bknudson, ++ debug would be cool, but not much else is likely needed. | 18:22 |
| dolphm | bknudson: ? http://pasteraw.com/4kv54uol5nc0fd34ea976vdb0e3gico | 18:23 |
| bknudson | dolphm: that's fancy! | 18:24 |
| dolphm | bknudson: would you prefer that to be in icehouse then? | 18:24 |
| dolphm | morganfainberg: ^ | 18:24 |
| morganfainberg | dolphm, hm. i like it. | 18:25 |
| bknudson | dolphm: I like it as is | 18:25 |
| morganfainberg | dolphm, but no need to add it | 18:25 |
| lbragstad | dolphm: ++ on the LOG.debug | 18:25 |
| bknudson | with the log or without it, I'm fine either way | 18:26 |
| morganfainberg | dolphm, if you're rolling another patchset, add it, if not, no need | 18:26 |
| lbragstad | same here, my comments were addressed | 18:26 |
| gyee | you guys know why we default the notification driver to rabbit? | 18:28 |
| gyee | used to be no op | 18:28 |
| gyee | TRACE keystone.notifications DriverLoadFailure: Failed to load transport driver "rabbit": No 'oslo.messaging.drivers' driver found, looking for 'rabbit' | 18:28 |
| lbragstad | gyee: was the a change that happened when we moved to oslo.messaging? | 18:28 |
| gyee | I believe so | 18:29 |
| gyee | looking at the code, we should default it to "fake" | 18:29 |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Ability to turn off ldap referral chasing https://review.openstack.org/78521 | 18:29 |
| dolphm | lbragstad: morganfainberg: bknudson: threw it in since the sample.conf needed an update too ^ | 18:29 |
| lbragstad | dolphm: ++ thanks! +1 | 18:29 |
| *** leseb has joined #openstack-keystone | 18:31 | |
| *** stevemar has quit IRC | 18:31 | |
| *** andreaf has joined #openstack-keystone | 18:33 | |
| bknudson | do we still have tests.dirs.tests('test_overrides.conf') ? | 18:34 |
| *** marekd is now known as marekd|away | 18:36 | |
| *** stevemar has joined #openstack-keystone | 18:37 | |
| *** gokrokve has joined #openstack-keystone | 18:53 | |
| *** vhoward has quit IRC | 19:01 | |
| dolphm | bknudson: that got removed, no? | 19:02 |
| dolphm | morganfainberg: ^ | 19:03 |
| dolphm | bknudson: now it's just "def config_overrides(..." | 19:03 |
| *** leseb has quit IRC | 19:17 | |
| lbragstad | dolphm: bknudson doing a quick grep on the latest master, and I'm not seeing it | 19:18 |
| *** saju_m has quit IRC | 19:21 | |
| *** saju_m has joined #openstack-keystone | 19:23 | |
| *** leseb has joined #openstack-keystone | 19:26 | |
| *** marcoemorais has quit IRC | 19:31 | |
| *** marcoemorais has joined #openstack-keystone | 19:34 | |
| *** leseb has quit IRC | 19:36 | |
| *** petertoft has joined #openstack-keystone | 19:42 | |
| *** petertoft has quit IRC | 19:43 | |
| dolphm | bknudson: just realized why you were asking :) | 19:46 |
| *** lbragstad has left #openstack-keystone | 19:49 | |
| *** lbragstad has joined #openstack-keystone | 19:49 | |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Ability to turn off ldap referral chasing https://review.openstack.org/78521 | 19:50 |
| *** amcrn has joined #openstack-keystone | 19:52 | |
| *** devlaps has quit IRC | 19:53 | |
| *** leseb has joined #openstack-keystone | 19:55 | |
| morganfainberg | bknudson, test_overrides is gone | 20:06 |
| morganfainberg | bknudson, yeah just config_overrides. | 20:07 |
| bknudson | ok, I need to run this by someone... | 20:07 |
| *** devlaps has joined #openstack-keystone | 20:07 | |
| *** saju_m has quit IRC | 20:08 | |
| bknudson | so I configure auth_token with an admin token... | 20:08 |
| bknudson | http://git.openstack.org/cgit/openstack/python-keystoneclient/tree/keystoneclient/middleware/auth_token.py?id=0.2.3#n317 | 20:08 |
| bknudson | (this is using old auth_token, I doubt it's changed) | 20:08 |
| *** saju_m has joined #openstack-keystone | 20:08 | |
| bknudson | Looks like my configured admin token can get overridden... | 20:09 |
| bknudson | http://git.openstack.org/cgit/openstack/python-keystoneclient/tree/keystoneclient/middleware/auth_token.py?id=0.2.3#n989 | 20:09 |
| bknudson | if verify_uuid_token fails with 401 | 20:09 |
| morganfainberg | bknudson, right. a verify of 401 should only occur if the admin token is invalid | 20:10 |
| morganfainberg | bknudson, a 404 should occur if the subject-token is invalid | 20:10 |
| morganfainberg | bknudson, i _think_ that fix went in for Havana | 20:10 |
| bknudson | morganfainberg: ah, so maybe this is fixed in more recent releases. | 20:11 |
| morganfainberg | bknudson, https://review.openstack.org/#/c/46123/ | 20:12 |
| *** browne has quit IRC | 20:12 | |
| morganfainberg | bknudson, that was specifically fixed to limit churn of tokens / fix admin token | 20:12 |
| morganfainberg | now if the admin token is invalid, yes, you'd overrite it | 20:13 |
| *** browne has joined #openstack-keystone | 20:13 | |
| bknudson | I'm trying to figure out if there's another case where the initial token would get overwritten. | 20:14 |
| bknudson | there's also the case where fetching the revocation list errors with 401... | 20:15 |
| morganfainberg | 401, if it would "expire soon" | 20:15 |
| morganfainberg | adaict | 20:15 |
| morganfainberg | the revocation list generating a 401 should be the same case, the admin token is invalid | 20:15 |
| morganfainberg | s/adaict/afict | 20:15 |
| bknudson | hm, this is weird... customer system it works the first time and fails otherwise. | 20:16 |
| morganfainberg | so if the admin token wouldn't "expire soon" and is valid, i don't see how it could be overridden | 20:16 |
| morganfainberg | grizzly? | 20:16 |
| morganfainberg | or havana+ | 20:16 |
| bknudson | yes, grizzly | 20:16 |
| morganfainberg | ok grizzly doesn't have the 401 fix | 20:16 |
| morganfainberg | if the subject token is invalid, you'll have that issue | 20:16 |
| bknudson | so the first time looks like it uses the admin token and then it tries to get a nova token (but their nova password is incorrect). | 20:16 |
| dolphm | duration for a full test run has almost double again for me this week | 20:16 |
| morganfainberg | dolphm, :( | 20:16 |
| bknudson | so the obvious fix is to use the right nova password. | 20:17 |
| bknudson | but I'm wondering why it's happening anyways. | 20:17 |
| morganfainberg | bknudson, because they are validating an invalid subject token | 20:17 |
| morganfainberg | bknudson, which also invalidates the admin token | 20:17 |
| morganfainberg | bknudson, if you backport that fix i linked, it should resolve it | 20:17 |
| morganfainberg | or convince them to move to Havana :P | 20:17 |
| dolphm | bknudson: yeah, the admin_token config for auth_token is completely broken/useless afaict | 20:18 |
| morganfainberg | dolphm, wonder if i can ressurect the parallel testing fix easily | 20:18 |
| morganfainberg | dolphm, as far as i know it might be usable now with the other test fixes. | 20:19 |
| dolphm | morganfainberg: i don't remember the list of blockers, but i bet we're damn close | 20:19 |
| morganfainberg | dolphm, the biggest blocker that i ran across was calling load_backends multiple times, but we may have solved a lot of that with the config_fixture | 20:19 |
| morganfainberg | dolphm, i still want to make load_backends called 1 time and must be called | 20:19 |
| morganfainberg | but.... | 20:19 |
| dolphm | bug 1240052 | 20:20 |
| uvirtbot | Launchpad bug 1240052 in keystone "Hardcoded paths prevent tests from being parallelized" [Low,In progress] https://launchpad.net/bugs/1240052 | 20:20 |
| morganfainberg | i am pretty sure with all the other fixes it might work now. | 20:20 |
| morganfainberg | OOH! uvirtbot YOU'RE BACK! | 20:20 |
| dolphm | uvirtbot: welcome | 20:20 |
| uvirtbot | dolphm: Error: "welcome" is not a valid command. | 20:20 |
| morganfainberg | dolphm, yeah i think _think_ most of the issues have been mostly resolved | 20:21 |
| morganfainberg | conf_fixture makes things sticky (option wise) and not loading things from conf files makes life way better | 20:21 |
| morganfainberg | dolphm, i'll bet this takes 2-3 minor changes. | 20:22 |
| dolphm | morganfainberg: do we still need the grouping here? https://review.openstack.org/#/c/65870/5/.testr.conf | 20:22 |
| dolphm | morganfainberg: i imagine we still need test_cert_setup | 20:22 |
| morganfainberg | yes we should still group, i think those. | 20:22 |
| morganfainberg | because we don't want to download keystoneclients multiple times | 20:22 |
| morganfainberg | and it think i can make the concurrency an ENV var. | 20:23 |
| morganfainberg | let me see. | 20:23 |
| dolphm | morganfainberg: why would you not want to use concurrency if it worked? | 20:24 |
| morganfainberg | dolphm, non-sqlite backends | 20:24 |
| morganfainberg | we need code to make dynamic dbs / ldap trees | 20:25 |
| morganfainberg | we don't have that. | 20:25 |
| morganfainberg | so in those cases you want to make it non-parallel | 20:25 |
| *** dstanek_afk has quit IRC | 20:25 | |
| dolphm | morganfainberg: fair enough | 20:25 |
| morganfainberg | also, https://bitbucket.org/hpk42/tox/pull-request/86/support-optional-env-variable-substitution/diff this will make it so we can make run_tests just call tox across the board | 20:25 |
| *** dstanek_afk has joined #openstack-keystone | 20:25 | |
| morganfainberg | if we can get that released and make the new version of tox our minimum. | 20:25 |
| morganfainberg | :) | 20:25 |
| morganfainberg | right now we can't support the --failing etc options. | 20:26 |
| morganfainberg | dolphm, ok i think i know what i need to do to make this "work". let me see how hard it will be to get parallel resurrected | 20:26 |
| *** vhoward has joined #openstack-keystone | 20:27 | |
| *** dstanek has joined #openstack-keystone | 20:31 | |
| *** dstanek_afk has quit IRC | 20:31 | |
| bknudson | --failing works for me. | 20:31 |
| morganfainberg | bknudson, with tox? | 20:32 |
| morganfainberg | bknudson, tox -epy27 -- --failing? | 20:32 |
| morganfainberg | or how do you run it? | 20:32 |
| bknudson | yes | 20:32 |
| morganfainberg | bknudson, because i couldn't get it to work. | 20:32 |
| morganfainberg | hmmm. i'll revisit | 20:32 |
| bknudson | morganfainberg: what does it do instead? | 20:33 |
| morganfainberg | bknudson, tells me '--failing' isn't a valid option | 20:33 |
| morganfainberg | testr error | 20:33 |
| morganfainberg | i'll revisit / try again | 20:33 |
| dolphm | morganfainberg: if you remove that concurrency line from testr.conf ... you shouldn't have to do anything else to enable concurrency, right? | 20:35 |
| morganfainberg | dolphm, well we need to do grouping and DB backend isolation | 20:35 |
| morganfainberg | dolphm, about 4 lines of code i think | 20:35 |
| morganfainberg | dolphm, i'm trying it now | 20:35 |
| dolphm | morganfainberg: i mean, just to have to run concurrently and crash as a result | 20:36 |
| morganfainberg | basically each worker needs it's own tmpdir | 20:36 |
| morganfainberg | oh right | 20:36 |
| morganfainberg | yes | 20:36 |
| dstanek | i barely have enough memory to run one process for the tests :-) | 20:39 |
| dstanek | i'm walking through the RC1 bugs and it looks like lots of them are already approved | 20:40 |
| dolphm | morganfainberg: why no crashy http://pasteraw.com/fg4zyjnhn03pv0glxqc40ey7zjm6sbr | 20:43 |
| dolphm | dstanek: lol get a bigger cloud server :P | 20:43 |
| morganfainberg | dolphm, lucky? | 20:44 |
| dolphm | dstanek: this is the last one we don't have a fix for https://bugs.launchpad.net/bugs/1291423 | 20:44 |
| uvirtbot | Launchpad bug 1291423 in keystone "revocation events sync slows responses to all authenticated calls" [High,Triaged] | 20:44 |
| dolphm | morganfainberg: but it was 2 minutes slower than my last attempt :P | 20:44 |
| dolphm | which was definitely non-parallel | 20:44 |
| morganfainberg | dolphm, hmmm... | 20:44 |
| morganfainberg | let me try with my minor changes | 20:44 |
| morganfainberg | and w/o | 20:44 |
| dolphm | dstanek: and i'm hoping this one can start gating soon :) https://review.openstack.org/#/c/78521/ | 20:45 |
| *** gokrokve has quit IRC | 20:45 | |
| *** gokrokve has joined #openstack-keystone | 20:46 | |
| *** gokrokve_ has joined #openstack-keystone | 20:47 | |
| morganfainberg | dolphm, ok trying a run w/o fixes | 20:47 |
| morganfainberg | dolphm, to get a baseline | 20:47 |
| morganfainberg | dolphm, then i'll run parallel, then i'll do it w/ fixes | 20:48 |
| dstanek | dolphm: in https://bugs.launchpad.net/bugs/1291423 do you know what calls are slow? everything? | 20:48 |
| uvirtbot | Launchpad bug 1291423 in keystone "revocation events sync slows responses to all authenticated calls" [High,Triaged] | 20:48 |
| dstanek | ayoung: ^ | 20:48 |
| *** gokrokv__ has joined #openstack-keystone | 20:48 | |
| morganfainberg | dstanek, the syncronize | 20:48 |
| morganfainberg | it's in keystone.contrib.revoke.model | 20:48 |
| morganfainberg | dstanek, since it affects all validation of tokens, it makes things much slower | 20:49 |
| morganfainberg | dstanek, and it is called on _all_ token validation. | 20:49 |
| morganfainberg | so each call to keystone has to validate a token | 20:49 |
| dolphm | it spends a lot of time ensuring the performance optimization is optimized for accuracy | 20:49 |
| morganfainberg | dolphm, ++ | 20:50 |
| morganfainberg | with the patch we got in earlier this week, i think it's less of an issue since revoke is actually disabled | 20:50 |
| *** gokrokve has quit IRC | 20:50 | |
| morganfainberg | unless you're using revoke that is | 20:50 |
| morganfainberg | and iirc tests are. | 20:50 |
| dolphm | morganfainberg: maybe that's why test runs are so slow now :P | 20:51 |
| morganfainberg | dolphm, likely | 20:51 |
| *** gokrokve has joined #openstack-keystone | 20:51 | |
| *** jamielennox|away is now known as jamielennox | 20:51 | |
| ayoung | dstanek, it might be nothing | 20:51 |
| dolphm | morganfainberg: does anything graph runtimes of jenkins jobs over time? | 20:51 |
| ayoung | with revocation events truely disabnled now...we might even be able to ignore that | 20:51 |
| *** gokrokve_ has quit IRC | 20:51 | |
| *** harlowja is now known as harlowja_away | 20:52 | |
| *** gokrokv__ has quit IRC | 20:52 | |
| dstanek | i'm going to poke around a bit and see what i can see | 20:52 |
| morganfainberg | dolphm, don't think we have that graph | 20:55 |
| morganfainberg | dolphm, i want that graph though | 20:55 |
| morganfainberg | dolphm, i'll poke around infra data and see what we log, maybe we can make it happen easily | 20:56 |
| bknudson | dolphm: I think migrations are O(n*n) | 20:57 |
| bknudson | or number of migrations * number of tests | 20:58 |
| dolphm | bknudson: morgan will fix that soon :) | 20:58 |
| morganfainberg | bknudson, as soon as RC is cut i'm going to smash those down | 20:59 |
| *** vhoward has left #openstack-keystone | 20:59 | |
| morganfainberg | bknudson, we'll drop everything below 37 (36 will be our starting point) | 20:59 |
| morganfainberg | since 36 was the last grizzly migration iirc | 21:00 |
| morganfainberg | dolphm, i'm seeing a lot slower of a run across the restful tests | 21:00 |
| morganfainberg | dolphm, non parallel: Ran 3163 tests in 814.493s | 21:03 |
| dolphm | morganfainberg: that's about the number i was seeing a week ago | 21:04 |
| morganfainberg | running in parallel, omg splody fail | 21:05 |
| morganfainberg | about 75% failure | 21:05 |
| morganfainberg | dolphm, you in #openstack, talking about a possible RC bug now | 21:05 |
| morganfainberg | might be an issue with TZ in kvs/memcache token | 21:05 |
| morganfainberg | just a heads up | 21:05 |
| dolphm | jenkins +1 on RC blocker :D https://review.openstack.org/#/c/78521/ bknudson lbragstad ayoung | 21:05 |
| *** leseb has quit IRC | 21:05 | |
| dolphm | morganfainberg: joining | 21:06 |
| morganfainberg | dolphm, k | 21:06 |
| morganfainberg | in parallel: | 21:07 |
| morganfainberg | Ran 3167 tests in 166.774s | 21:07 |
| morganfainberg | FAILED (failures=1674) | 21:07 |
| bknudson | almost halfway there. | 21:08 |
| *** dstanek has quit IRC | 21:08 | |
| *** dstanek has joined #openstack-keystone | 21:14 | |
| *** derek_c has joined #openstack-keystone | 21:14 | |
| *** harlowja_away is now known as harlowja | 21:14 | |
| morganfainberg | dolphm, are you running with -- --subunit ? | 21:20 |
| morganfainberg | dolphm, i think that makes it implicily single-worker | 21:20 |
| morganfainberg | dolphm, https://bugs.launchpad.net/keystone/+bug/1294862 i'll see if i can get something rolled by tomorrow for this | 21:21 |
| uvirtbot | Launchpad bug 1294862 in keystone "Token expiration time with memcache->kvs->dogpile is wrong" [Medium,New] | 21:21 |
| dolphm | morganfainberg: no, but i got them to fail with --parallel --concurrency=5 | 21:22 |
| morganfainberg | dolphm, hehe | 21:22 |
| morganfainberg | dolphm, going to try and run this w/o --subunit | 21:22 |
| morganfainberg | i wonder if i somehow still end up with a single-worker somehow... | 21:23 |
| morganfainberg | because i was seeing massive failures before | 21:23 |
| morganfainberg | dolphm, ok --subunit makes it single worker | 21:24 |
| morganfainberg | running 4 workers now | 21:24 |
| morganfainberg | and i am suddenly even happier some people like living on the edge and running master of keystone ;) | 21:26 |
| *** daneyon has quit IRC | 21:26 | |
| *** bvandenh has quit IRC | 21:27 | |
| morganfainberg | dolphm, looks like one-ish failure so far. | 21:28 |
| *** daneyon has joined #openstack-keystone | 21:28 | |
| *** saju_m has quit IRC | 21:28 | |
| *** derek_c has quit IRC | 21:28 | |
| *** pcargnel_ has quit IRC | 21:31 | |
| *** zhiyan is now known as zhiyan_ | 21:36 | |
| morganfainberg | dolphm, Ran 3165 (+1469) tests in 607.362s (+235.483s) FAILED (id=5, failures=4 (+3), skips=816) | 21:36 |
| morganfainberg | well thats 200s better | 21:37 |
| dolphm | morganfainberg: just 4 failures running in parallel? | 21:41 |
| morganfainberg | dolphm, with a couple minor modifications to the tmpdir calc code | 21:41 |
| morganfainberg | dolphm, and the failures are because the test_cert stuff ended up using TESTSDIR not the TMPDIR | 21:41 |
| morganfainberg | it looks like | 21:41 |
| morganfainberg | or my group_regex was also bad | 21:42 |
| morganfainberg | and the last run i'll do this with the DB in tmpfs | 21:43 |
| morganfainberg | once i've got everything working as expected (no failures) | 21:43 |
| morganfainberg | yep group regex is somehow messed up | 21:44 |
| dolphm | morganfainberg: are you going to restore https://review.openstack.org/#/c/65870/ ? | 21:46 |
| morganfainberg | dolphm, yeah i'll toss this review to supplant that one | 21:46 |
| dolphm | morganfainberg: revise the commit message from Partial-Bug to Closes-Bug then :) | 21:46 |
| morganfainberg | yep | 21:47 |
| *** lbragstad has quit IRC | 21:47 | |
| *** leseb has joined #openstack-keystone | 21:49 | |
| ayoung | dolphm, surprised I never formally reveiwed that: I know I've looked at it | 21:52 |
| dolphm | ayoung: you reviewed it enough to block it, with good reason :) | 21:56 |
| *** derek_c has joined #openstack-keystone | 21:59 | |
| ayoung | dolphm, ah...there it is, hadn't looked far enough back in history. Anyway, +2 +A. think it should go in | 22:01 |
| morganfainberg | dolphm, Ran 3164 tests in 186.247s FAILED (id=0, failures=13, skips=816) - with tmpfs | 22:02 |
| morganfainberg | dolphm, also, with parallel, you'll need to nuke your .testrepository, it has bad training data in it | 22:02 |
| morganfainberg | just the first time. | 22:03 |
| *** devlaps has quit IRC | 22:03 | |
| *** marcoemorais has quit IRC | 22:09 | |
| *** marcoemorais has joined #openstack-keystone | 22:09 | |
| openstackgerrit | Jenkins proposed a change to openstack/python-keystoneclient: Updated from global requirements https://review.openstack.org/79695 | 22:10 |
| *** devlaps has joined #openstack-keystone | 22:18 | |
| *** david-lyle has quit IRC | 22:19 | |
| *** bknudson has quit IRC | 22:24 | |
| *** derek_c has quit IRC | 22:27 | |
| morganfainberg | dolphm, using tmpfs back: Ran 3163 (+3106) tests in 186.896s (+178.655s) | 22:32 |
| morganfainberg | dolphm, 4 workers | 22:32 |
| *** marcoemorais has quit IRC | 22:36 | |
| dstanek | dolphm, morganfainberg, ayoung: random thought - could we use a non-blocking lock instead of a blocking mutex? | 22:38 |
| morganfainberg | dstanek, for? | 22:38 |
| dstanek | the revoke sync - only one thread really ever needs to do it and there isn't much of a need to block other threads while performing the task | 22:39 |
| morganfainberg | dstanek, well with the new code it does caching instead of an explicit lock/sync | 22:40 |
| morganfainberg | dstanek, so it would use the same locking that whatever caching backend uses. | 22:40 |
| morganfainberg | iirc | 22:40 |
| morganfainberg | should be better than the explicit lock/sync each time | 22:40 |
| morganfainberg | but i think the way the tree is built/tested against is the real issue | 22:40 |
| morganfainberg | ayoung and YorikSar were working on that bit | 22:41 |
| dstanek | morganfainberg: i'm not really a fan of the in process tree index | 22:42 |
| *** thedodd has quit IRC | 22:42 | |
| *** daneyon has quit IRC | 22:43 | |
| dstanek | morganfainberg: i "finished" the blinker poc a few days ago, but didn't have time to debug a few tests - seems like some data is being left hanging about and causing failures | 22:46 |
| *** browne has left #openstack-keystone | 22:46 | |
| *** derek_c has joined #openstack-keystone | 22:49 | |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/keystone: Add support for parallel testr workers in Keystone https://review.openstack.org/65870 | 22:49 |
| morganfainberg | dolphm, dstanek, ^ yes it works | 22:49 |
| morganfainberg | dstanek, dolphm, last run with no tmpfs: Ran 3163 tests in 391.212s (+204.376s) | 22:50 |
| morganfainberg | with tmpfs Ran 3163 (+3106) tests in 186.896s (+178.655s) | 22:51 |
| morganfainberg | that is 4 workers | 22:51 |
| morganfainberg | lets hope jenkins likes it this time | 22:51 |
| ayoung | dstanek, I thought I had submitted a patch to do that | 22:51 |
| *** zhiyan_ is now known as zhiyan | 22:51 | |
| *** lbragstad has joined #openstack-keystone | 22:51 | |
| *** dims has quit IRC | 22:51 | |
| ayoung | but I am not seeing it... | 22:52 |
| morganfainberg | though it appears that use of -- --subunit will make it implictly a single-worker | 22:52 |
| morganfainberg | dstanek, cool | 22:52 |
| morganfainberg | dstanek, i'm sure it wont be hard to smush the bugs out | 22:53 |
| *** zhiyan is now known as zhiyan_ | 22:53 | |
| dstanek | morganfainberg: yeah, let me post the reviews - the first two refactor things a bit to make the third one easier | 22:53 |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/keystone: Add support for parallel testr workers in Keystone https://review.openstack.org/65870 | 22:56 |
| dstanek | ah, that's why i failed to commit the refactorings. there is still some dependencies on a private variable that need to be removed | 22:57 |
| *** harlowja is now known as harlowja_away | 22:59 | |
| morganfainberg | hmmm. ok so i need to move the "vendor" directory | 23:05 |
| morganfainberg | *grumble* | 23:05 |
| *** jagee has quit IRC | 23:07 | |
| morganfainberg | have i mentioned how much i dislike the "checkout_vendor" stuff? | 23:07 |
| *** dims has joined #openstack-keystone | 23:07 | |
| jamielennox | morganfainberg: i can't remember if you have done it, but so far everyone else who has ever played with fixing the tests has had a go | 23:09 |
| morganfainberg | jamielennox, i'm going to just move it into the tmpdir | 23:09 |
| morganfainberg | i think. | 23:09 |
| jamielennox | morganfainberg: i saw something interesting the other day that i haven't been able to track down again | 23:09 |
| jamielennox | (or tried that hard) | 23:09 |
| jamielennox | i think it was related to WSME or pecan or something like that | 23:09 |
| jamielennox | that they do gate checks in reverse | 23:10 |
| morganfainberg | interesting | 23:10 |
| morganfainberg | i mean, don't get me wrong, the tests passed py26 and py27 | 23:10 |
| jamielennox | so not only do you do all off the pecan unit tests but you run a couple of the openstack services with the new pecan to make sure it's not introducing regressions | 23:10 |
| morganfainberg | but we do have a minor race on checkout of keystoneclient | 23:10 |
| openstackgerrit | David Stanek proposed a change to openstack/keystone: More notification unit tests https://review.openstack.org/81659 | 23:11 |
| jamielennox | so i wonder if we could offload those checkout tests to a zuul job instead? | 23:11 |
| openstackgerrit | David Stanek proposed a change to openstack/keystone: Refactor notifications https://review.openstack.org/81660 | 23:11 |
| openstackgerrit | David Stanek proposed a change to openstack/keystone: wip: Implements notifications with blinker https://review.openstack.org/81661 | 23:11 |
| morganfainberg | jamielennox, actually i was thinking of making them their own tox target | 23:11 |
| morganfainberg | jamielennox, and introducing a new test run for jenkins, but ehh... | 23:12 |
| morganfainberg | jamielennox, i don't like that after learning more about how tox works and the magic pyXX envs | 23:12 |
| jamielennox | ... do you have that sort of control in tox? i mean i know there is the -r requirements, but can you then lock down the exact client version | 23:12 |
| morganfainberg | jamielennox, sure. we could ovverride things | 23:12 |
| morganfainberg | jamielennox, but probably not a good idea. | 23:13 |
| morganfainberg | oh oh i know how to fix this. | 23:13 |
| morganfainberg | derrrrp | 23:13 |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/keystone: Add support for parallel testr workers in Keystone https://review.openstack.org/65870 | 23:14 |
| morganfainberg | make the group regex less picky | 23:14 |
| morganfainberg | group both test_keystoneclient test files together | 23:14 |
| morganfainberg | anyway | 23:16 |
| morganfainberg | there we go | 23:16 |
| *** derek_c has quit IRC | 23:19 | |
| jamielennox | grr, no reviews overnight... | 23:20 |
| *** gokrokve_ has joined #openstack-keystone | 23:27 | |
| openstackgerrit | Brant Knudson proposed a change to openstack/python-keystoneclient: Reference docstring for auth_token fields https://review.openstack.org/70617 | 23:27 |
| *** lbragstad has quit IRC | 23:28 | |
| *** gokrokve has quit IRC | 23:30 | |
| openstackgerrit | Brant Knudson proposed a change to openstack/keystone: Clean up config help text https://review.openstack.org/78497 | 23:34 |
| *** huats has quit IRC | 23:35 | |
| *** huats has joined #openstack-keystone | 23:35 | |
| *** huats has quit IRC | 23:35 | |
| *** huats has joined #openstack-keystone | 23:35 | |
| *** bknudson has joined #openstack-keystone | 23:37 | |
| *** marcoemorais has joined #openstack-keystone | 23:37 | |
| *** harlowja_away is now known as harlowja | 23:38 | |
| *** vhoward has joined #openstack-keystone | 23:39 | |
| openstackgerrit | Brant Knudson proposed a change to openstack/keystone: Cleanup config.py https://review.openstack.org/81671 | 23:39 |
| *** marcoemorais has quit IRC | 23:42 | |
| *** stevemar has quit IRC | 23:48 | |
| *** derek_c has joined #openstack-keystone | 23:48 | |
| *** henrynash has quit IRC | 23:51 | |
| openstackgerrit | Brant Knudson proposed a change to openstack/keystone: Remove unused db_sync from extensions https://review.openstack.org/79211 | 23:51 |
| *** leseb has quit IRC | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!