| openstackgerrit | Steve Martinelli proposed a change to openstack/python-openstackclient: Create a whole slew of functional tests for identity https://review.openstack.org/125216 | 00:06 |
|---|---|---|
| *** rmcall has quit IRC | 00:38 | |
| openstackgerrit | Steve Lewis proposed a change to stackforge/python-openstacksdk: Handle booleans that are strings in APIs https://review.openstack.org/124209 | 01:22 |
| *** bknudson has quit IRC | 01:27 | |
| *** bknudson has joined #openstack-sdks | 01:27 | |
| openstackgerrit | Steve Martinelli proposed a change to openstack/python-openstackclient: Create a whole slew of functional tests for identity https://review.openstack.org/125216 | 01:33 |
| *** jamiehannaford has joined #openstack-sdks | 01:42 | |
| dtroyer | stevemar: that's close to what I was thinking and works for me. BTW, s/shell/shell script/ | 02:38 |
| stevemar | dtroyer, OK cool, that's what i thought :P | 02:39 |
| stevemar | dtroyer, i don't mind changing it up, whatever you think is best, i'm indifferent on this one | 02:39 |
| *** jamiehannaford has quit IRC | 02:43 | |
| dtroyer | stevemar: I'm working on two things: a) this should all be together in a subdir, and 2) that subdir shouldn't be in the project namespace. I could be wrong on both for all I know… | 02:43 |
| dtroyer | I feel like we need to be mindful of keeping top-level namespaces neat. | 02:45 |
| stevemar | valid concerns | 02:45 |
| stevemar | OTOH, we can always move these around | 02:46 |
| stevemar | it's not that much of a PITA (yet) | 02:46 |
| stevemar | dtroyer, re: running on other clouds... the harpoonrc file is sourced before the run, so any env. variables added there would be picked up, so as long as OS_AUTH_URL is there, we should be good... | 02:47 |
| dtroyer | I know but I'm getting impatient with the gate, right now fewer reviews feels more productive. | 02:47 |
| stevemar | where were you a few weeks ago during FF? | 02:48 |
| stevemar | that was a real clusterf*** | 02:48 |
| dtroyer | stevemar: I'm thinking about adding a localrc to harpoonrc similar to what devstack does so we don't have to modify a file from the repo for this to work | 02:48 |
| dtroyer | stevemar: oh I was here…or mostly over there chasing devstack and whatnot | 02:49 |
| stevemar | dtroyer, hmmm | 02:50 |
| stevemar | dtroyer, localrc --> to setup a devstack? | 02:55 |
| stevemar | dtroyer, i am thinking... maybe just `harpoon.sh <path_to_openrc>` | 02:56 |
| dtroyer | maybe, once we do that. right now, just to set the auth, or in my first use case, to set DEVSTACK_DIR to ~/devstack so I can use the user files it created | 02:56 |
| dtroyer | doing it as an arg would work for the current case, but with a localrc I could just set the creds up directly | 02:57 |
| dtroyer | but being able to switch quickly would be nice too (thinking out loud now) | 02:57 |
| stevemar | dtroyer, i do account of DEVSTACK_DIR in the script | 02:58 |
| dtroyer | I liked localrc (or local.conf?) because it is a usage pattern folks are already familiar with | 02:58 |
| *** briancurtin has quit IRC | 02:58 | |
| stevemar | dtroyer, mull it over; i'm just glad the tests are running when they are supposed to | 02:59 |
| stevemar | moving things around, or changing the behaviour, is not that much of an issue IMO, since it's mostly an automated job that cares about this :P | 03:00 |
| dtroyer | sure, just thinking how to make this work with my current setup (devstack in $HOME and $DEST=/opt/stack) | 03:00 |
| stevemar | dtroyer, set DEVSTACK_DIR in harpoonrc | 03:01 |
| dtroyer | the point was I didn't want to modify a file from the repo just to run it. now I have a changed file that shouldn't be checked in | 03:01 |
| stevemar | ohhhhh | 03:01 |
| stevemar | i see i see | 03:02 |
| stevemar | *now the wheels are turning* | 03:02 |
| dtroyer | localrc is never distributed. I left that bit out maybe | 03:02 |
| stevemar | i guess harpoonrc doesn't need to be checked in | 03:02 |
| stevemar | we could source localrc, if it exists | 03:03 |
| dtroyer | yes it does, there will always be something that needs to be twiddled with that should have a default | 03:03 |
| dtroyer | again, I'm maintaining the usage patterns from devstack and genade | 03:03 |
| stevemar | which is a good thing | 03:03 |
| dtroyer | I could be convinced that isn't appropriate | 03:03 |
| dtroyer | ok | 03:03 |
| stevemar | dtroyer, i dunno, sleep on it, my only request is that this gets in before paris :P | 03:04 |
| dtroyer | of course! actually, I think I'm happy with what you have now. I haven't tried it yet but looks good at a glance | 03:05 |
| stevemar | dtroyer, the main issue is there are 2 main use cases: 1) running against a predictable jenkins job, and 2) running against any cloud. | 03:06 |
| dtroyer | yup. you're worrying about the first and I'm worrying about the second. I think they're totally compatible | 03:07 |
| stevemar | the first is done, and though we don't have the second completed, we haven't worked ourselves into a corner | 03:07 |
| *** openstackgerrit has quit IRC | 03:08 | |
| stevemar | the 2nd requires better tests too, almost every argument passed in has to be a uuid value, or we're going to have possible duplicate errors | 03:08 |
| stevemar | doesn't complicate things or anything, just more of a heads up | 03:09 |
| dtroyer | I haven't thought that far ahead yet so thanks. that means I also haven't really looked at your follow-on review with all the tests in it | 03:10 |
| dtroyer | case 2 is secondary though | 03:10 |
| dtroyer | heh, go figure | 03:10 |
| *** openstackgerrit has joined #openstack-sdks | 03:10 | |
| dtroyer | ah, nice. just setting DEVSTAK_DIR was enough and it works | 03:11 |
| stevemar | hooray | 03:11 |
| stevemar | another thought - the jenkins job is based on devstack-gate i think... which always has the same setup | 03:12 |
| stevemar | which i think doesn' | 03:13 |
| stevemar | doesn't include neutron | 03:13 |
| stevemar | or swift? | 03:13 |
| dtroyer | It can include it, I think that is a jjb setting. swift should be there by default? <checking> | 03:13 |
| *** briancurtin has joined #openstack-sdks | 03:14 | |
| *** briancurtin has joined #openstack-sdks | 03:14 | |
| dtroyer | I see swift logs in the -full job | 03:14 |
| stevemar | oh thats good | 03:14 |
| dtroyer | http://logs.openstack.org/09/122609/5/check/check-tempest-dsvm-full/488a7d5/logs/localrc.txt.gz | 03:14 |
| stevemar | full might be different | 03:14 |
| dtroyer | and functional: http://logs.openstack.org/09/122609/5/gate/gate-osc-dsvm-functional/35376e4/logs/localrc.txt.gz | 03:15 |
| stevemar | niceeee | 03:15 |
| dtroyer | looks like the sane ENABLED_SERVICES | 03:15 |
| dtroyer | nova-net | 03:15 |
| stevemar | oh trove too, neat | 03:16 |
| dtroyer | so we might need a second one at some point to get neutron | 03:16 |
| stevemar | yep | 03:16 |
| stevemar | more work for terry! | 03:16 |
| dtroyer | because we need to support both net interfaces for quite some time | 03:16 |
| stevemar | i can toss up a patch for that when needed | 03:16 |
| dtroyer | so I'm +2 on that, let's see if terrylhowe has some input yet | 03:17 |
| stevemar | awesomo | 03:17 |
| stevemar | i imagine it'll still need some work around the edges, especially for use case 2, but we'll get there | 03:17 |
| *** openstackgerrit has quit IRC | 03:18 | |
| stevemar | blah, tempest fail | 03:18 |
| *** openstackgerrit has joined #openstack-sdks | 03:18 | |
| stevemar | i already found some bugs with the functional tests too :) | 03:19 |
| dtroyer | I just pushed 122331 back into the queue. I though maybe a new +A would get it straight into the gate queue but no…start from the top… | 03:19 |
| dtroyer | since it's getting late I'm going to do 122334 so they can live or die together | 03:19 |
| stevemar | lots of the v3 identity commands have a 'links' field in the returned object, i'll push another patch to fix those | 03:20 |
| stevemar | poor gate :( | 03:20 |
| dtroyer | so the 'whole slew' (I assume that's a Canadian term?) ran for me… | 03:29 |
| dtroyer | I see what you mean now about running against random clouds, maybe at some point we need a user-level vs admin-level distinction? | 03:30 |
| stevemar | dtroyer, i didn't think it was until now ... http://www.itworldcanada.com/article/all-things-weird-and-wonderful/35961 maybe "whole slew" is used more often up here | 03:33 |
| dtroyer | I've heard it before, but not often. | 03:34 |
| stevemar | yeah... i try to cover the case for admin/user level in harpoonrc, specifying the user and tenant in there too | 03:34 |
| dtroyer | so right now, running as demo/demo, I get a number of failures, as expected. | 03:36 |
| dtroyer | should we trap the HTTP 403 and mark that as skipped maybe? that's the right return for not-admin, and really the only way we have of knowing if the account has privs or not | 03:37 |
| stevemar | dtroyer, probably easier to create a a non-admin specific suite of tests | 03:59 |
| stevemar | and check at the on setup if we should skip a test or not | 03:59 |
| stevemar | just talking aloud | 03:59 |
| dtroyer | That would be good especially where you don't want to go banging on Rax or HP with admin requests and 'get flagged'. Although, arguably, we don't have access to their admin endpoints, at least for Identity. | 04:01 |
| dtroyer | but v3… | 04:01 |
| dtroyer | anyway, that's for later. | 04:02 |
| dtroyer | I think I'm about done for the night… | 04:02 |
| *** briancurtin has quit IRC | 05:07 | |
| *** terrylhowe has quit IRC | 05:34 | |
| *** stevemar has quit IRC | 06:20 | |
| openstackgerrit | A change was merged to openstack/python-openstackclient: Add low-level API base class https://review.openstack.org/122331 | 06:43 |
| openstackgerrit | A change was merged to openstack/python-openstackclient: Move object-store commands to low-level API https://review.openstack.org/122334 | 06:43 |
| *** jamiehannaford has joined #openstack-sdks | 07:59 | |
| *** mfer has joined #openstack-sdks | 08:04 | |
| openstackgerrit | Marek Denis proposed a change to openstack/python-openstackclient: Implement CRUD operations for Mapping objects https://review.openstack.org/86912 | 08:06 |
| *** mfer has quit IRC | 08:09 | |
| *** fifieldt has joined #openstack-sdks | 12:37 | |
| *** terrylhowe has joined #openstack-sdks | 13:08 | |
| openstackgerrit | Marek Denis proposed a change to openstack/python-openstackclient: CRUD operations for federated protocols https://review.openstack.org/125627 | 13:19 |
| openstackgerrit | Marek Denis proposed a change to openstack/python-openstackclient: CRUD operations for federated protocols https://review.openstack.org/125627 | 13:27 |
| tellesnobrega | terrylhowe, hey, can you give a little overview of what are the plans for the unified sdks from now on? what are the major needs? | 13:32 |
| terrylhowe | well tellesnobrega brian curtain is the acting ptl on the project | 13:32 |
| terrylhowe | he doesn’t seem to be on now | 13:33 |
| terrylhowe | the biggest thing to me is get through the code that is in the pipe and fill it out. It is still missing version discovery and support of compute v2.1, orchestration, probably some trove stuff. I’d love to see cloud foundry support. | 13:34 |
| terrylhowe | if that is even possible, I don’t know what the cloud foundry api looks like | 13:34 |
| terrylhowe | stevelle: was working on trove stuff, so that might be taken care of | 13:35 |
| terrylhowe | originally, there was talk about a higher level api that was version agnostic, not sure if that is realistically going to work out. I kind of think it is impossible when you get down to it. | 13:37 |
| terrylhowe | there are still gaps in identity, but jamie lennox was working on that | 13:37 |
| terrylhowe | tellesnobrega: whatever support you need to get rolling, I can provide | 13:40 |
| *** mfer has joined #openstack-sdks | 13:41 | |
| tellesnobrega | terrylhowe, thanks. we worked on a bug for getting the authentication version from the auth url. submitted a patch yesterday | 13:42 |
| tellesnobrega | i want to know the priorities so we can organize what and figure out where we can be helpful | 13:42 |
| terrylhowe | compute v2.1, orchestration and cloud foundry are least likely to step on work in the pipe, so those would be the most important I think | 13:44 |
| terrylhowe | tellesnobrega: I just noticed no one has worked on block storage, that is a huge gap | 13:46 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:49 | |
| tellesnobrega | terrylhowe, thanks for the help. Hopefully we will be able to get some work done here | 13:51 |
| terrylhowe | that would be awesome tellesnobrega | 13:52 |
| openstackgerrit | Marek Denis proposed a change to openstack/python-openstackclient: CRUD operations for federated protocols https://review.openstack.org/125627 | 13:54 |
| *** briancurtin has joined #openstack-sdks | 14:01 | |
| *** briancurtin2 has joined #openstack-sdks | 14:02 | |
| *** briancurtin has quit IRC | 14:02 | |
| *** stevemar has joined #openstack-sdks | 14:08 | |
| openstackgerrit | Victor Silva proposed a change to stackforge/python-openstacksdk: Determines version from auth_url when not explicit https://review.openstack.org/125473 | 15:05 |
| *** briancurtin2 is now known as briancurtin | 15:15 | |
| *** briancurtin has joined #openstack-sdks | 15:15 | |
| *** rmcall has joined #openstack-sdks | 15:49 | |
| stevemar | terrylhowe, ping! | 15:59 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 17:03 | |
| *** briancurtin has quit IRC | 17:08 | |
| terrylhowe | what’s up stevemar ? | 17:19 |
| openstackgerrit | Steve Martinelli proposed a change to openstack/python-openstackclient: Pass in domain and project as positional args, not kwargs https://review.openstack.org/125714 | 17:20 |
| stevemar | terrylhowe, sent you a PM, just a request to review the functional test code | 17:20 |
| terrylhowe | ah, cool, I’ll take a look | 17:21 |
| *** rmcall_ has joined #openstack-sdks | 17:23 | |
| *** rmcall has quit IRC | 17:24 | |
| *** rmcall_ is now known as rmcall | 17:24 | |
| stevemar | terrylhowe, dtroyer what do you guys think about defaulting OS_USER_DOMAIN_ID and OS_PROJECT_DOMAIN_ID to "default" ? | 17:24 |
| stevemar | currently for authN'ing with keystone v3 those would need to be supplied... | 17:24 |
| stevemar | getting food | 17:25 |
| terrylhowe | it would be convenient | 17:25 |
| terrylhowe | could always back that out if someone complains | 17:26 |
| terrylhowe | A better message from keystone might help too as I recall it just says 405 malformed | 17:26 |
| *** rmcall_ has joined #openstack-sdks | 17:31 | |
| *** rmcall has quit IRC | 17:31 | |
| *** rmcall has joined #openstack-sdks | 17:34 | |
| *** rmcall_ has quit IRC | 17:35 | |
| openstackgerrit | Victor Silva proposed a change to stackforge/python-openstacksdk: Determines version from auth_url when not explicit https://review.openstack.org/125473 | 17:35 |
| *** briancurtin has joined #openstack-sdks | 17:37 | |
| *** briancurtin has joined #openstack-sdks | 17:37 | |
| dtroyer | stevemar: I'm fine with that, it's unlikely Keystone will change that any time soon, right? | 17:38 |
| victsou | hey briancurtin, thank you for your patience on my patch | 17:46 |
| victsou | a handful of silly mistakes :p | 17:46 |
| briancurtin | victsou: no worries. thank *you* for writing it! | 17:47 |
| victsou | sure, np | 17:47 |
| stevemar | dtroyer, it's highly unlikely that will change - maybe if we move to v4 :P | 17:55 |
| dtroyer | stevemar: as they say, version numbers are cheap, right? | 17:56 |
| stevemar | dtroyer, that explains why windows is skipping 9 | 17:56 |
| dtroyer | so winver will return 7.0, right? No, that's not confusing. I really think its because they want to confuse everyone who knows 'ten' as a Mac OS… | 18:00 |
| *** stevemar has quit IRC | 18:00 | |
| *** stevemar has joined #openstack-sdks | 18:01 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 18:03 | |
| stevemar | dtroyer, ping | 18:23 |
| stevemar | dtroyer, actually, we are calling user.update.assert_called_with() | 18:24 |
| stevemar | https://github.com/openstack/python-openstackclient/blob/master/openstackclient/tests/identity/v3/test_user.py#L849-L852 | 18:24 |
| dtroyer | so did I pick that from the wrong command? | 18:26 |
| dtroyer | that review patches SetUser | 18:26 |
| dtroyer | heh, so the tests check what we send, not what the spec says? also, shouldn't that review break the test? | 18:28 |
| dtroyer | ah, it did break… | 18:28 |
| openstackgerrit | Steve Martinelli proposed a change to openstack/python-openstackclient: Pass in domain and project as positional args, not kwargs https://review.openstack.org/125714 | 18:32 |
| *** jamiehannaford has quit IRC | 18:36 | |
| *** bknudson has quit IRC | 20:07 | |
| *** bknudson has joined #openstack-sdks | 20:13 | |
| openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: compute/v2 limits_absolute resource https://review.openstack.org/116932 | 20:28 |
| *** rmcall has quit IRC | 20:55 | |
| *** HenryG has quit IRC | 20:55 | |
| *** stevemar has quit IRC | 21:05 | |
| *** mfer has quit IRC | 21:26 | |
| *** terrylhowe has quit IRC | 21:38 | |
| *** terrylhowe has joined #openstack-sdks | 21:39 | |
| openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: Make user_id a higher priority for v2 auth https://review.openstack.org/125799 | 21:42 |
| *** briancurtin has quit IRC | 21:53 | |
| *** briancurtin has joined #openstack-sdks | 21:54 | |
| *** briancurtin has quit IRC | 21:58 | |
| *** briancurtin has joined #openstack-sdks | 22:54 | |
| *** briancurtin has joined #openstack-sdks | 22:54 | |
| *** briancurtin has quit IRC | 22:58 | |
| *** victsou has quit IRC | 23:06 | |
| *** victsou has joined #openstack-sdks | 23:07 | |
| *** victsou is now known as vsilva | 23:09 | |
| *** HenryG has joined #openstack-sdks | 23:10 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!