*** gordc has quit IRC | 00:10 | |
*** david-lyle is now known as david-lyle_afk | 00:14 | |
*** mdbooth has quit IRC | 00:29 | |
openstackgerrit | Merged openstack/cliff: Hide prompt in batch/pipe mode https://review.openstack.org/157999 | 00:32 |
---|---|---|
*** mdbooth has joined #openstack-oslo | 00:35 | |
*** jaypipes has quit IRC | 00:40 | |
*** ViswaV has quit IRC | 00:51 | |
*** salv-orlando has quit IRC | 01:00 | |
*** mtanino has quit IRC | 01:01 | |
*** dims_ has joined #openstack-oslo | 01:11 | |
*** dims_ has quit IRC | 01:11 | |
*** dims_ has joined #openstack-oslo | 01:11 | |
*** dims has quit IRC | 01:12 | |
*** ViswaV has joined #openstack-oslo | 01:23 | |
*** dims has joined #openstack-oslo | 01:37 | |
*** dims_ has quit IRC | 01:40 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 01:44 | |
*** ViswaV has quit IRC | 01:48 | |
*** noelbk has quit IRC | 02:15 | |
*** stevemar has joined #openstack-oslo | 02:49 | |
*** takedakn has joined #openstack-oslo | 02:52 | |
*** yamahata has quit IRC | 02:52 | |
*** takedakn has quit IRC | 03:02 | |
*** takedakn has joined #openstack-oslo | 03:03 | |
*** dims has quit IRC | 03:24 | |
*** takedakn has quit IRC | 03:49 | |
*** wendar has quit IRC | 04:04 | |
*** wendar has joined #openstack-oslo | 04:04 | |
*** dims has joined #openstack-oslo | 04:25 | |
*** dims has quit IRC | 04:30 | |
*** jecarey has joined #openstack-oslo | 04:30 | |
*** yamahata has joined #openstack-oslo | 05:00 | |
*** jamespage has quit IRC | 05:05 | |
*** jamespage has joined #openstack-oslo | 05:06 | |
*** jecarey has quit IRC | 05:29 | |
*** achanda has quit IRC | 05:50 | |
*** noelbk has joined #openstack-oslo | 05:56 | |
*** ajo has joined #openstack-oslo | 06:00 | |
*** inc0 has joined #openstack-oslo | 06:06 | |
*** ViswaV has joined #openstack-oslo | 06:11 | |
*** ViswaV_ has joined #openstack-oslo | 06:15 | |
*** ViswaV has quit IRC | 06:16 | |
*** achanda has joined #openstack-oslo | 06:19 | |
*** takedakn has joined #openstack-oslo | 06:27 | |
*** takedakn has quit IRC | 07:04 | |
*** ajo has quit IRC | 07:08 | |
*** subscope has quit IRC | 07:14 | |
openstackgerrit | Sabari proposed openstack/oslo.vmware: Add get_datastore_by_ref method to oslo.vmware https://review.openstack.org/159357 | 07:42 |
*** dims has joined #openstack-oslo | 08:03 | |
*** e0ne has joined #openstack-oslo | 08:03 | |
*** dims has quit IRC | 08:07 | |
*** achanda has quit IRC | 08:09 | |
*** ihrachyshka has joined #openstack-oslo | 08:12 | |
*** ViswaV_ has quit IRC | 08:15 | |
*** dulek has joined #openstack-oslo | 08:22 | |
*** dtantsur|afk is now known as dtantsur | 08:26 | |
*** ihrachyshka has quit IRC | 08:36 | |
*** salv-orlando has joined #openstack-oslo | 08:39 | |
*** andreykurilin_ has joined #openstack-oslo | 08:49 | |
*** rushiagr_away is now known as rushiagr | 08:49 | |
*** yamahata has quit IRC | 08:53 | |
*** i159 has joined #openstack-oslo | 08:54 | |
*** stevemar has quit IRC | 08:57 | |
*** dulek has quit IRC | 08:57 | |
*** dulek has joined #openstack-oslo | 08:58 | |
*** dulek_ has joined #openstack-oslo | 09:04 | |
*** dulek has quit IRC | 09:04 | |
*** SpamapS has quit IRC | 09:04 | |
*** SpamapS has joined #openstack-oslo | 09:05 | |
*** takedakn has joined #openstack-oslo | 09:05 | |
*** MasterPiece has quit IRC | 09:10 | |
*** SpamapS has quit IRC | 09:10 | |
*** noelbk has quit IRC | 09:13 | |
*** jgrimm is now known as zz_jgrimm | 09:15 | |
*** dulek_ has quit IRC | 09:31 | |
*** SpamapS has joined #openstack-oslo | 09:32 | |
*** SpamapS has joined #openstack-oslo | 09:32 | |
*** andreykurilin_ has quit IRC | 09:32 | |
*** jaosorior has joined #openstack-oslo | 09:38 | |
*** salv-orlando has quit IRC | 09:49 | |
*** achanda has joined #openstack-oslo | 10:01 | |
*** ajo has joined #openstack-oslo | 10:09 | |
*** achanda has quit IRC | 10:17 | |
*** takedakn has quit IRC | 10:20 | |
*** ihrachyshka has joined #openstack-oslo | 10:41 | |
*** achanda has joined #openstack-oslo | 10:50 | |
*** dims has joined #openstack-oslo | 10:52 | |
*** himangi has joined #openstack-oslo | 10:54 | |
*** dims_ has joined #openstack-oslo | 10:55 | |
*** dims has quit IRC | 10:58 | |
*** amotoki has joined #openstack-oslo | 11:12 | |
*** dtantsur is now known as dtantsur|bbl | 11:14 | |
*** salv-orlando has joined #openstack-oslo | 11:15 | |
*** achanda has quit IRC | 11:17 | |
*** pcaruana has joined #openstack-oslo | 11:19 | |
*** rushiagr is now known as rushiagr_away | 11:33 | |
sdague | what's the path for oslo policy to not complain if this directory doesn't exist? - https://review.openstack.org/#/c/152082 | 11:41 |
*** salv-orlando has quit IRC | 11:44 | |
*** cdent has joined #openstack-oslo | 11:54 | |
openstackgerrit | Sergey Gotliv proposed openstack/oslo-incubator: Improve logging to debug invalid "extra_specs" entries https://review.openstack.org/159432 | 12:08 |
*** kgiusti has left #openstack-oslo | 12:11 | |
*** dims_ has quit IRC | 12:14 | |
*** dims has joined #openstack-oslo | 12:14 | |
*** rodrigods is now known as rodrigods_ | 12:16 | |
*** rodrigods_ is now known as rodrigods__ | 12:16 | |
*** rodrigods__ is now known as rodrigod` | 12:16 | |
*** rodrigod` is now known as rodrigods | 12:17 | |
*** _amrith_ is now known as amrith | 12:17 | |
*** himangi has quit IRC | 12:20 | |
*** dtantsur|bbl is now known as dtantsur | 12:31 | |
*** mtanino has joined #openstack-oslo | 12:33 | |
*** takedakn has joined #openstack-oslo | 12:42 | |
inc0 | dhellmann, dansmith, dims -> https://review.openstack.org/#/c/159380/ it turns out that I need global requirements after all... | 12:44 |
*** dulek has joined #openstack-oslo | 12:44 | |
*** dulek_ has joined #openstack-oslo | 12:46 | |
*** rodrigods is now known as rodrigod` | 12:51 | |
dims | inc0: of course you do :) | 12:53 |
inc0 | I was kinda hoping that local heat requirements.txt would be enough | 12:53 |
*** rodrigod` is now known as rodrigods | 12:53 | |
dims | inc0: there's a way to do, but won't recommend it | 12:53 |
*** salv-orlando has joined #openstack-oslo | 12:54 | |
inc0 | please share | 12:54 |
kragniz | dims: you tease | 12:54 |
inc0 | I just want reviews to start | 12:55 |
inc0 | k-3 is getting closer:( | 12:55 |
dims | inc0: can you please update projects.txt too? http://git.openstack.org/cgit/openstack/requirements/tree/projects.txt | 12:55 |
dims | kragniz: not teasing, it a path for projects in stackforge that want to depend as much as possible on listed requirements but need a couple more that are not in global (say nova-docker needs docker-py). getting you that switch in a sec :) | 12:56 |
dims | kragniz: inc0: REQUIREMENTS_MODE=soft from http://git.openstack.org/cgit/openstack-infra/devstack-gate/tree/devstack-vm-gate-wrap.sh#n334 | 12:59 |
inc0 | dims, added to projects | 13:01 |
inc0 | I don't think heat uses this;) I guess I'll wait for global-reqirements | 13:02 |
openstackgerrit | Merged openstack/cliff: Fix pep8 tests for lambda https://review.openstack.org/155778 | 13:05 |
*** kgiusti has joined #openstack-oslo | 13:11 | |
*** sputnik13 has quit IRC | 13:17 | |
*** rushiagr_away is now known as rushiagr | 13:22 | |
*** takedakn has quit IRC | 13:37 | |
*** mtanino has quit IRC | 13:44 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 13:59 | |
*** viktors|afk is now known as viktors | 14:04 | |
*** exploreshaifali has joined #openstack-oslo | 14:11 | |
*** salv-orlando has quit IRC | 14:15 | |
*** bknudson has quit IRC | 14:15 | |
*** gordc has joined #openstack-oslo | 14:17 | |
*** bknudson has joined #openstack-oslo | 14:36 | |
*** jungleboyj has quit IRC | 14:36 | |
*** zz_jgrimm is now known as jgrimm | 14:43 | |
dhellmann | inc0, dims : we aren't ready to add versionedobjects to the global requirements yet, are we? | 14:47 |
*** mriedem has joined #openstack-oslo | 14:49 | |
inc0 | dhellmann, I thought I could avoid that by adding that to local requirements.txt of heat | 14:49 |
inc0 | I was wrong. | 14:50 |
inc0 | so for now we have unusable, released lib :( | 14:50 |
dhellmann | inc0: what's stopping you? | 14:50 |
inc0 | zuul | 14:50 |
dhellmann | inc0: "it doesn't work" doesn't help me debug -- do you have a log or something? | 14:50 |
inc0 | gate requirements | 14:50 |
inc0 | let me show you | 14:50 |
dims | dhellmann: so, heat wants to use o.vo in kilo itself | 14:51 |
*** jgrimm is now known as zz_jgrimm | 14:51 | |
inc0 | https://review.openstack.org/#/c/146841/ | 14:52 |
dhellmann | dims: and when dansmith is sure the API is solid, then that should be fine. In the mean time, I *thought* we could test integration without adding the lib to the global requirements list | 14:52 |
inc0 | and there is gate in zuul called requirements which throws that oslo.versionedobjects is not in global requirements, so go away | 14:53 |
dims | dhellmann: to do what inc0 needs we have to switch off some of the requirements jobs and throw in REQUIREMENTS_MODE=soft in the gate jobs | 14:53 |
dims | right | 14:53 |
dhellmann | dims: I didn't realize devstack was also testing the requirements, I thought we had a separate job for that | 14:53 |
dansmith | dhellmann: this is what I was saying yesterday, I don't think devstack will even start a stack | 14:54 |
dims | dhellmann: i've done this for nova-docker (set things up so docker-py which is not in global reqs) can be used | 14:54 |
dhellmann | bah | 14:54 |
dims | but don't recommend it! | 14:54 |
dansmith | dhellmann: it doesn't normally, but thought maybe you were indicating that it was overridden in the gate for testing new things | 14:54 |
dhellmann | sdague: can we separate the testing of valid requirements sets from the installing and running of tempest so we can test early releases of libraries? | 14:54 |
dansmith | dhellmann: isn't everyone that starts using this in a well-understood buyer-beware sort of situation? | 14:55 |
dhellmann | dansmith: everyone always says they understand, and then gets pissy when it turns out that we do change something. So not in this Real Life. :-( | 14:55 |
dhellmann | dims: isn't this how we did oslo.log? I thought we avoided adding it to the requirements list for a while. | 14:56 |
dansmith | sigh | 14:57 |
dansmith | dhellmann: I think that if heat isn't able to merge things nowish that use it, then we're pretty much sunk for kilo | 14:57 |
sdague | dhellmann: the g-r sync is actually really important in devstack, it's the reason it was created | 14:57 |
inc0 | maybe we should enable some sort of backdoor for selected projects for that? | 14:57 |
*** tedross_ has joined #openstack-oslo | 14:57 | |
inc0 | we need gunea pigs... | 14:57 |
dansmith | dhellmann: even if devstack would let us do testing, we still have to do the sync before they can merge, right? | 14:58 |
dims | dhellmann: digging up a review | 14:58 |
dhellmann | sdague: I understand, but now I'm in a situation that I have to release an untested library with an unproven API (no offense dansmith) just to *try* it in a project. Is there some middleground? Split that into a separate job, maybe? | 14:58 |
sdague | because without it you get the old case, where projects install conflicting versions of the same library, then pkg_resources explodes on a service start | 14:58 |
dims | dhellmann: inc0: https://review.openstack.org/#/c/147641/ | 14:59 |
dhellmann | sdague: right, I'm saying run the sync in "light" mode inside devstack and run the full thing separately in another job to test the actual requirements match. Didn't we used to do that? | 14:59 |
*** exploreshaifali has quit IRC | 14:59 | |
dims | dhellmann: inc0: you can use that technique to test | 14:59 |
sdague | dhellmann: no, g-r was created for the devstack case | 14:59 |
sdague | that's actually the bug that we were fixing | 14:59 |
dhellmann | sdague: at some point we had a job that *just* looked at the requirements list, though, I thought | 14:59 |
sdague | we also have those jobs | 15:00 |
dhellmann | right now the only case I need to support is adding a new lib to a project | 15:00 |
sdague | but devstack has always worked this way | 15:00 |
sdague | since we created g-r | 15:00 |
dhellmann | ok. I want to change it. | 15:00 |
sdague | it's really the point | 15:00 |
*** prad has joined #openstack-oslo | 15:00 | |
sdague | so, I think that's a bad idea | 15:00 |
dims | sdague: let's see how far inc0 gets with the technique i used above for oslo.log | 15:01 |
dhellmann | would it not work to sync the requirements in the same mode as projects that don't enforce g-r? that only allows new things, right? | 15:01 |
sdague | dhellmann: and it opens up this whole other wedge problem | 15:01 |
sdague | which is not theoretical, it was blocking us a couple times a month | 15:01 |
inc0 | uhh...should I do that in heat? | 15:01 |
dhellmann | sdague: but if we're still also requiring that requirements actually match in a separate job, nothing that failed that would merge, right? | 15:01 |
sdague | dhellmann: I think we're talking past each other, maybe | 15:02 |
dhellmann | sdague: what I'm saying is that we're combining the 2 tests now, "do all of these things work together" and "is this project using valid versions of the requirements" | 15:02 |
dhellmann | and I want those tests to be distinct, so that I can try things that fail the second to see if they pass the first before I fix the second | 15:03 |
sdague | dhellmann: so, the problem with that | 15:03 |
sdague | is we install 20 projects | 15:03 |
sdague | or more | 15:03 |
dhellmann | I don't understand why we have to do the requirements check inside devstack itself, I guess. | 15:03 |
dhellmann | I understand why we need to sync there, but not check. | 15:04 |
*** daniel3_ has joined #openstack-oslo | 15:04 | |
sdague | hmm.... | 15:04 |
inc0 | dhellmann, but since we run gate tests, we should run these on same setup it would before merge | 15:04 |
inc0 | maybe just fail this test and allow installation of rest? | 15:05 |
dhellmann | inc0: it would be. You would add a new requirement and the gate check would run with it. The *separate* requirements job would fail, until the global-requirements were updated and then you would have to re-run the gate check. | 15:05 |
inc0 | so I'll have one merge-blocking failed test, but I could work on it and rest of gates would produce meaningful output? | 15:05 |
dansmith | inc0: that's what I was hoping for | 15:06 |
dhellmann | inc0: right | 15:06 |
inc0 | well...it didn't.. | 15:06 |
dhellmann | inc0: yes, I misremembered how we did this for oslo.log and now I'm trying to see if we can change that aspect of the gate | 15:07 |
sdague | dhellmann: so, that's going to make an assumption that the requirements job covers all the inbound possible changes for this for all projects | 15:07 |
sdague | which... I'm not sure we have been careful about | 15:07 |
* dhellmann parses that sentence | 15:07 | |
dhellmann | does the requirement job not verify that the requirements listed in requirements.txt match exactly what is in global-requirements.txt? | 15:08 |
dhellmann | interestingly, I don't see the requirements job running for heat here: https://review.openstack.org/#/c/146841/ | 15:08 |
sdague | it does, but I'm not sure what the guaruntees are for everything that can be specified in devstack | 15:08 |
sdague | and if something manages to bump that | 15:09 |
sdague | and slip in, then we end up wedged in devstack | 15:09 |
*** e0ne is now known as e0ne_ | 15:09 | |
dhellmann | ok, fair | 15:09 |
sdague | we could create a job in the experimental queues that did the soft update | 15:09 |
sdague | so it wouldn't impact the main tests, but could be run on demand | 15:10 |
dhellmann | could we change the existing job to do the check at the end? | 15:10 |
dhellmann | that would let it actually try to run the tests, and then only fail if the requirements don't match properly (assuming the code works) | 15:10 |
sdague | no, I think you miss the fail condition | 15:10 |
sdague | we do the sync | 15:10 |
sdague | we install the requirements | 15:10 |
dansmith | TBH, I don't see what the point of doing that is | 15:10 |
*** yamahata has joined #openstack-oslo | 15:11 | |
dansmith | inc0: is going to put up code that he know works, and if the gate verifies it, he's going to want to merge it | 15:11 |
sdague | the failure condition is that project A installs requirements, project B installs different versions of same requirements | 15:11 |
dansmith | that doesn't tell us anything about the API or the fitness of the library other than that it does some basic stuff | 15:11 |
sdague | project A can't start because the wrong versions are now available | 15:11 |
*** jecarey has joined #openstack-oslo | 15:11 | |
dansmith | it's weeks down the road where we might hit some snag, and that's going to be after he'll have wanted to be merging code | 15:11 |
dhellmann | dansmith: because if it *doesn't* work when you do the same thing in nova, you're going to want to change the API and may break what was working for inc0 in another project | 15:11 |
sdague | it fails in an extremely cryptic tempest error | 15:11 |
dhellmann | I'm trying to avoid having to honor our first draft of the API for all time | 15:11 |
dansmith | dhellmann: but what about what we're discussing doing right now is going to fix that? | 15:12 |
dhellmann | dansmith: I am prioritizing API stability over early adoption. | 15:12 |
*** sputnik13 has joined #openstack-oslo | 15:12 | |
sdague | way after the real failure is, so it's really hard to sort out what's going on | 15:12 |
dhellmann | it will let us test the library in several projects before committing to the API | 15:12 |
sdague | dhellmann: so I think in that case, what you *actually* want is a different job | 15:13 |
dansmith | dhellmann: but I'll not be ready to test the library in nova for months, are we going to hold up inc0 that long? | 15:13 |
dhellmann | sdague: ok, yeah. So how about the experimental job. How would that work? Can we set it up so all projects have it? | 15:13 |
dhellmann | dansmith: if need be. | 15:13 |
dansmith | hmm, that is not at all what I thought we were doing here | 15:13 |
dhellmann | dansmith: we can have that conversation, but if you're ready to commit to the API that works for heat also working in nova, then for this lib we can go ahead. I am also trying to solve this problem for everyone else that comes later, though. | 15:13 |
*** salv-orlando has joined #openstack-oslo | 15:14 | |
dansmith | well, it impacts me very little, compared to inc0 | 15:14 |
dhellmann | My goal for this lib was a release by the end of the cycle. I do not want to jeopardize long-term stability for an impatient early adopter. | 15:14 |
sdague | right, the issue being that if we bump the API for nova, heat can't stay on the old version, because we force them off with g-r. Which really is about the problems of having a system level install of libraries. | 15:15 |
inc0 | but while I agree nova is top priority for this, I think caucious early adoption would be very beneficial | 15:15 |
sdague | so... there is another possible option | 15:15 |
*** SridharGaddam has quit IRC | 15:15 | |
inc0 | we'll catch out any nova-specific things in the process | 15:15 |
dhellmann | inc0: the key word there is cautious, which is what I'm trying to be | 15:15 |
dhellmann | sdague: what's that? | 15:16 |
sdague | hmmm.... actually, never mind, other ideas have problems | 15:16 |
dansmith | so, I thought the point here was that this is our own library, we can evolve the API and the requirement as needed, | 15:16 |
inc0 | dhellmann, and I totally agree, but we won't find bugs unless we actually start using it | 15:16 |
dhellmann | dansmith: we can, but we are constrained by the fact that all projects must always be using the same version of the library. | 15:16 |
dansmith | if we find something that is a no-go for nova, we can iterate heat from 0.1.0 to 0.1.7 before we merge 0.1.8 which is what nova actually needs | 15:16 |
dansmith | sure | 15:16 |
inc0 | can't we somehow restrict adoption of this for a time being? | 15:17 |
dhellmann | inc0: no, there is no way to restrict adoption once we release the first version | 15:17 |
sdague | inc0: it's released on pypi, you can't | 15:17 |
dhellmann | the *only* thing stopping anyone from using it now is that it's not in global-requirements | 15:17 |
inc0 | also, I'll be here, if we find something hard for nova, I can work with dansmith to solve it in a way that will be ok for heat too | 15:17 |
sdague | I suppose you could put a big warnings.warn in it that this library interface is currently unstable | 15:17 |
sdague | until you are convinced it's not | 15:18 |
sdague | at least new adopters would see that when they ran tests | 15:18 |
dansmith | sdague: nobody is going to use this but us.. we've been saying that | 15:18 |
inc0 | we could remove all docs about that and make code criptic enough for anyone to understand;) | 15:18 |
dhellmann | dansmith: "us" includes all of stackforge | 15:18 |
dansmith | sdague: so I honestly don't think it's even a concern | 15:18 |
dansmith | dhellmann: stackwho? :D | 15:18 |
* dansmith kids | 15:18 | |
*** e0ne_ has quit IRC | 15:19 | |
sdague | yeh, so this feels like the standard snake eating it's tale of testing / releasing. You don't want to release until it's tested, but no one really uses it until it's released, which is a bunch of your testing. | 15:20 |
sdague | honestly, I'd just go with release, and tell people the interface will evolve, use at your own risk | 15:20 |
sdague | and if people yell later, just point them at that part of the readme | 15:21 |
dhellmann | sdague: right. our unique gate structure lets us release and test without pushing to production, though, and I'd like to take advantage of that | 15:21 |
dansmith | lemme ask this: | 15:21 |
dhellmann | dansmith: you're the lead for the library, so I'm going to defer to you. I would prefer to wait, and figure out a way to have test jobs going in multiple projects before we commit to an API, but if you want to go ahead we can. I'll direct questions and complaints to you, though. ;-) | 15:21 |
dansmith | actually nevermind | 15:22 |
dansmith | dhellmann: no, don't do that :) | 15:22 |
dansmith | (defer I mean) | 15:22 |
dhellmann | well, you're going to be doing most of the work either way, so if this was what you planned and intended, I don't want to stand in the way. I can work with sdague to set up the testing I want for future libraries in parallel. | 15:24 |
dansmith | this is my first rodeo, so what I intended isn't entirely relevant here I don't think | 15:25 |
dansmith | but what I definitely didn't get, | 15:25 |
dansmith | was that you wanted to see happy runs of changes from nova, heat, and maybe cinder before we started using it at all | 15:25 |
dhellmann | I'd be happy with 2, but yeah. | 15:25 |
dansmith | and I feel like that's a tooon of queued up code for something like this | 15:26 |
dansmith | I mean nova probably landed 50 patches to get this stuff in place before we actually started sending stuff over the wire | 15:26 |
dhellmann | For example, in oslo.policy we found when we integrated with keystone that they needed a bunch of data that hadn't been exposed publicly so they were pulling from config options. | 15:26 |
dhellmann | so you're not planning to have a single patch to swap what's in nova to use the new lib? | 15:26 |
dansmith | I will have to do some surgery on nova first | 15:27 |
dansmith | it won't be one patch | 15:27 |
sdague | out of curiosity, where's the heat patch in question | 15:28 |
inc0 | hold on | 15:28 |
inc0 | https://review.openstack.org/#/c/146841/ | 15:28 |
inc0 | and following patches | 15:28 |
sdague | concrete examples might let me come up with other ideas that won't break other things we depend on | 15:28 |
dhellmann | dansmith: ok, I also didn't anticipate that level of change being needed, I guess. In the past we've done adoption in one patch but those were smaller libs or matched more closely to the incubated API | 15:28 |
dansmith | dhellmann: well, there were a lot of novaisms to remove, and we changed things about the registry that were very magical, and/or modifying application state | 15:29 |
dhellmann | sdague: if it makes things simpler, the only case I care about this for is a new requirement. We get testing of requirement updates by modifying global-requirements itself. | 15:29 |
dhellmann | dansmith: right | 15:29 |
inc0 | well, heat won't be frozen project eighter. If API would require a change, I think we can coordinate | 15:29 |
dansmith | dhellmann: nothing was fundamental, but I'll have work to do to make it fit | 15:29 |
dhellmann | we have had to do that in the past, but we typically did it in one patch so there was a before and after rather than gradual transition | 15:30 |
inc0 | I think only problem we'll have would be rogue implementers who would ignore warnings of possible API changes | 15:30 |
dansmith | so dhellmann, let me put up another straw man to shoot down: | 15:30 |
dhellmann | maybe the patch would be too big in nova | 15:30 |
sdague | so, yeh, heat is a good instance where if we backed that off, they'd be able to break the world here, as they don't have requirements enforcement job | 15:31 |
dhellmann | inc0: There is no amount of education or publicity that will prevent that. And then when you change the API, they will say "Oslo broke my project!" | 15:31 |
sdague | however, inc0 what about making the heat functional job go to soft requirements | 15:31 |
dhellmann | why doesn't heat have a requirements job? | 15:31 |
sdague | because at least for heat, most of your real testing is there anyway | 15:31 |
sdague | dhellmann: because no one ever added it? | 15:32 |
dhellmann | sdague: ok, I thought maybe there was some reason not to have it | 15:32 |
sdague | the job management is all manual, stuff gets missed all the time | 15:32 |
sdague | yeh, I don't know, I'm just looking at the job output | 15:32 |
dhellmann | yeah, that's cool, I thought maybe there was some technical reason | 15:32 |
inc0 | sdague, how could I do that? I guess that would require messing up with heat CI, and that something I'd have to discuss with Angus (PTL) | 15:33 |
dhellmann | sdague, inc0 : https://review.openstack.org/159508 | 15:33 |
sdague | inc0: you would set REQUIREMENTS_MODE=soft in the definition of that job | 15:35 |
*** zz_jgrimm is now known as jgrimm | 15:35 | |
dansmith | sdague: and that lets him start up a devstack with a local requirement out of sync? | 15:35 |
dims | dansmith: yes | 15:36 |
dansmith | okay, so: | 15:36 |
dansmith | if I can do that locally, I can try to sprint to a patch(set) for nova that uses the library for some amount of workyness | 15:36 |
dansmith | I have no idea how long it will take | 15:36 |
dansmith | but I can try and get something up so we have an idea, or see if I hit any dragons | 15:37 |
dansmith | I dunno that we want to define any new jobs for nova, maybe experimental | 15:37 |
dansmith | but that would maybe let us get some test runs on both projects for dhellmann ? | 15:37 |
*** jungleboyj has joined #openstack-oslo | 15:37 | |
dims | dansmith: you can run it through gate too using this technique - https://review.openstack.org/#/c/147641/ | 15:37 |
dansmith | dims: with a cross-project linkage so zuul uses it? | 15:38 |
dims | dansmith: this predates Depends-On | 15:38 |
dims | but yes, that can be tried too | 15:38 |
inc0 | or I could link Depends-On to global requirements patch | 15:38 |
dims | inc0: worth trying | 15:39 |
dansmith | dims: I don't see how that works for testing an uncommitted nova change otherwise | 15:39 |
inc0 | I guess that would solve zuul problem as well | 15:39 |
dims | dansmith: you file a review against nova, and not merge it | 15:39 |
sdague | yeh, Depends-On actually solves this without any new jobs | 15:39 |
dims | dansmith: then you file a review similar to above and pull in the changeset from nova | 15:39 |
*** e0ne has joined #openstack-oslo | 15:39 | |
dims | sdague: hope so | 15:39 |
dansmith | dims: oh right I didn't even finish reading it | 15:40 |
sdague | that's actually probably the right way to do it, just Depends-On the requirements change | 15:40 |
dims | sdague: +1 | 15:40 |
dansmith | dhellmann: please send coffee to portland | 15:40 |
dims | sdague: i just don't know if requirements repo can get pulled | 15:40 |
sdague | dims: yeh, it is | 15:40 |
sdague | that should work | 15:40 |
dims | cool | 15:40 |
dims | there you go guys, try that first and then we have a fallback inc0 and dansmith | 15:41 |
dansmith | just so we're clear, | 15:41 |
dansmith | we're talking about inc0 depending his, and me writing something for nova that depends, | 15:41 |
dansmith | so we can see that both work, yes? | 15:41 |
dims | yep | 15:41 |
inc0 | cool, I'll do that then, but forgive me if that will be tomoroow;) | 15:41 |
inc0 | dcan't we jyust both write depend on to requirement patch? | 15:42 |
dims | inc0: yes that's what you should try first | 15:42 |
inc0 | great | 15:42 |
dansmith | yes | 15:42 |
inc0 | and if all will be well, we'll simply merge g-r | 15:43 |
inc0 | thanks guys | 15:43 |
dhellmann | I like the depends on thing, dims! We should write that down. | 15:43 |
dhellmann | dansmith: if portland needs more coffee, we're all doomed. | 15:43 |
dansmith | haha | 15:43 |
dansmith | well, one specific place in portland does.. | 15:43 |
dims | dhellmann: y, let's try it with o.vo and i'll write it down | 15:43 |
dhellmann | dims: yep | 15:44 |
dhellmann | I have to run an errand, but it sounds like we have a way forward. Thank you all for working it out! | 15:44 |
ihrachyshka | dhellmann, who's behind oslo.policy? | 15:48 |
dims | ihrachyshka: Ian and Steve - http://stackalytics.com/?module=oslo.policy&metric=commits | 15:57 |
ihrachyshka | dims, do they have irc nicks? | 15:58 |
ihrachyshka | oh, I think I can get that from gerrit names | 15:58 |
*** inc0 has quit IRC | 15:58 | |
ihrachyshka | sigmavirus24, hey | 15:58 |
sigmavirus24 | yes? | 15:59 |
sigmavirus24 | steve seems to not be on at the moment | 15:59 |
sigmavirus24 | Also I'm not core | 15:59 |
sigmavirus24 | So morganfainberg and bknudson and others should be asked this question as well | 15:59 |
sigmavirus24 | ihrachyshka: what's up? | 15:59 |
morganfainberg | Hi. | 15:59 |
morganfainberg | What's up? | 16:00 |
ihrachyshka | sigmavirus24, so I try to apply the lib to neutron, and I get errors due to register decorator not avail | 16:00 |
*** yamahata has quit IRC | 16:00 | |
ihrachyshka | we apply @policy.register('tenant_id') to a Check class | 16:00 |
ihrachyshka | and it's missing in oslo.policy | 16:00 |
ihrachyshka | though docs still refer to iut | 16:00 |
ihrachyshka | *it | 16:01 |
ihrachyshka | http://docs.openstack.org/developer/oslo.policy/api.html#registering-new-special-checks | 16:01 |
morganfainberg | Ok hold on. I need to switch contexts here. | 16:01 |
morganfainberg | Wow. That's some descriptive docs. :P | 16:02 |
*** himangi has joined #openstack-oslo | 16:03 | |
morganfainberg | Ok I found the code looking to see why you don't get it on import ihrachyshka | 16:04 |
morganfainberg | ihrachyshka: https://github.com/openstack/oslo.policy/blob/master/oslo_policy/_checks.py#L206 it's there. | 16:04 |
ihrachyshka | morganfainberg, but it's _checks | 16:05 |
ihrachyshka | probably private | 16:05 |
ihrachyshka | I would expect it in policy.py? | 16:05 |
morganfainberg | Yes. Looks like it is missing from policy.py | 16:05 |
ihrachyshka | morganfainberg, what's the plan to release something official btw? | 16:06 |
morganfainberg | If it is meant to be public (docs indicate it should be) | 16:06 |
ihrachyshka | morganfainberg, neutron would be glad to switch prior to kilo | 16:06 |
morganfainberg | ihrachyshka: soon â„¢. | 16:06 |
morganfainberg | Keystone as well | 16:06 |
ihrachyshka | morganfainberg, should I just send a patch moving the function to public place? | 16:07 |
morganfainberg | Well don't "move" it. It probably just needs a reference in policy.py. | 16:07 |
*** sigmavirus24 is now known as Slackwarebot | 16:07 | |
morganfainberg | I'm looking now to see where it goes. | 16:07 |
*** Slackwarebot is now known as sigmavirus24 | 16:07 | |
ihrachyshka | ok right, smth like generate = _checks.generate | 16:07 |
*** himangi has quit IRC | 16:08 | |
morganfainberg | Yeah. | 16:08 |
*** himangi has joined #openstack-oslo | 16:08 | |
morganfainberg | This was probably just missed. And is a valid bug. Since we are close to releasing officially you can fix it and we can get that added. | 16:09 |
ihrachyshka | morganfainberg, ack, on my way | 16:10 |
morganfainberg | We should be releasing soon. So changing over in kilo should be possible. Keystone really would like this as well. | 16:10 |
morganfainberg | :). Good catch. | 16:10 |
ihrachyshka | morganfainberg, yeah, was tired of waiting for release and decided to try it out before | 16:12 |
morganfainberg | ^_^ | 16:12 |
*** himangi has quit IRC | 16:12 | |
openstackgerrit | Ihar Hrachyshka proposed openstack/oslo.policy: Expose register decorator as part of public API https://review.openstack.org/159525 | 16:12 |
ihrachyshka | morganfainberg, ^^ | 16:12 |
*** himangi has joined #openstack-oslo | 16:13 | |
morganfainberg | Yep. We also see oslo.policy announcement in -keystone since most of the cores are keystone folks. | 16:13 |
morganfainberg | :) | 16:13 |
ihrachyshka | morganfainberg, hm, other stuff is missing too | 16:16 |
ihrachyshka | like Check class | 16:16 |
ihrachyshka | it's also in _checks. | 16:16 |
morganfainberg | We might have gotten a bit aggressive in internal api-ing things. | 16:17 |
morganfainberg | It is always easier to make something a public api from private than the inverse. | 16:17 |
ihrachyshka | morganfainberg, yeah, but that thing is mentioned in public docs | 16:18 |
ihrachyshka | it's not like neutron got consuming something hidden | 16:18 |
morganfainberg | Again. Too aggressive. :). | 16:18 |
bnemec | sdague: Not sure if you ever got an answer, but the missing policy.d thing should be fixed by https://github.com/openstack/oslo.policy/commit/d8129456855a63e124cef8435dc12b5bd4a82d03 | 16:18 |
sdague | bnemec: cool, is that released? | 16:19 |
ihrachyshka | morganfainberg, ok, I'll add Check in the patch too | 16:19 |
bnemec | sdague: I believe so, but I'm not sure if anyone has adopted the lib yet. | 16:19 |
bnemec | Seems like most projects are still on the incubator version. | 16:20 |
morganfainberg | sdague: it was tagged. Going to be officially released soon (needs a few bug fixes before it's ready for full consumption) | 16:20 |
sdague | bnemec: ok, could we fix the incubator as well then? | 16:20 |
sdague | because, that's a really noisy state to be in for the release | 16:20 |
sdague | if most projects aren't going to switch over | 16:20 |
morganfainberg | Keystone has a patch to use it. Next tag + requirements update will likely go out soon (we were hoping this week) | 16:20 |
openstackgerrit | Ihar Hrachyshka proposed openstack/oslo.policy: Expose register decorator as part of public API https://review.openstack.org/159525 | 16:20 |
bnemec | sdague: Yeah, that would be fine with me. I think dhellmann wanted to focus on getting the lib adopted instead, but this late in the cycle I don't know if that's still going to happen. | 16:21 |
*** exploreshaifali has joined #openstack-oslo | 16:22 | |
morganfainberg | bnemec: I think a couple projects will adopt. But not all. | 16:22 |
morganfainberg | Keystone and neutron likely. I doubt nova will. | 16:22 |
sdague | yeh, and all of them are currently generating warnings on tihs | 16:22 |
bnemec | Yeah | 16:22 |
morganfainberg | So we should fix incubator. | 16:22 |
morganfainberg | (Unfortunately) | 16:22 |
ihrachyshka | morganfainberg, re option registration, where is the best option to do it? when importing neutron.policy, or when calling neutron.policy.init()? | 16:24 |
viktors | dhellmann: around? | 16:24 |
morganfainberg | ihrachyshka: hm. Not sure how neutron sets all that up so I don't have a good answer :(. Keystone we register at fixed times. | 16:25 |
morganfainberg | ihrachyshka: I think it just depends on how neutron generally does that. | 16:26 |
ihrachyshka | I need to check, but the answer will probably be 'randomly' | 16:26 |
ihrachyshka | that's the neutron culture! | 16:26 |
ekarlso | is oslo.policy released or ? | 16:28 |
*** tsekiyama has joined #openstack-oslo | 16:28 | |
morganfainberg | ekarlso: tagged but needs another patch or so and another tag + requirements update before being used. | 16:29 |
morganfainberg | ekarlso: very very soon â„¢ it will be released imo. | 16:29 |
morganfainberg | ihrachyshka: not sure if I should laugh, cry, or something else ;) | 16:30 |
ihrachyshka | #sadpanda is the right tag | 16:30 |
ihrachyshka | oh, it also adopted a new option group... | 16:33 |
*** viktors is now known as viktors|afk | 16:34 | |
*** david-lyle_afk is now known as david-lyle | 16:36 | |
*** rushiagr is now known as rushiagr_away | 16:40 | |
*** himangi has quit IRC | 16:42 | |
*** amotoki has quit IRC | 16:43 | |
*** stevemar has joined #openstack-oslo | 16:44 | |
openstackgerrit | Greg Hill proposed openstack/taskflow: add jobboard trash method https://review.openstack.org/158459 | 16:44 |
*** himangi has joined #openstack-oslo | 16:46 | |
*** rushiagr_away is now known as rushiagr | 16:47 | |
*** pcaruana has quit IRC | 16:50 | |
*** salv-orlando has quit IRC | 16:53 | |
*** takedakn has joined #openstack-oslo | 16:53 | |
*** amotoki has joined #openstack-oslo | 16:54 | |
dhellmann | viktors|afk: sorry I missed you. Congrats on the oslo.db release! I see a few things in https://launchpad.net/oslo.db/+milestone/1.5.0 that aren't "done" so I'll move those to next-kilo. | 16:54 |
*** ihrachyshka has quit IRC | 17:00 | |
*** amotoki has quit IRC | 17:02 | |
*** dulek has quit IRC | 17:10 | |
*** dulek_ has quit IRC | 17:10 | |
*** i159 has quit IRC | 17:11 | |
*** dtantsur is now known as dtantsur|afk | 17:13 | |
*** noelbk has joined #openstack-oslo | 17:17 | |
*** ViswaV has joined #openstack-oslo | 17:20 | |
openstackgerrit | Merged openstack/cliff: Allow to call initialize_app when running --help https://review.openstack.org/158343 | 17:21 |
*** ZZelle has joined #openstack-oslo | 17:22 | |
ZZelle | Hi everyone! | 17:22 |
ZZelle | Would it make sense to move https://github.com/openstack/heat/blob/master/heat/api/middleware/ssl.py in oslo.middleware? | 17:22 |
ZZelle | <sigmavirus24> ZZelle: that would be a better question for #openstack-oslo | 17:22 |
openstackgerrit | Ihar Hrachyshka proposed openstack/oslo.policy: Expose stuff used in Neutron as part of public API https://review.openstack.org/159525 | 17:23 |
*** ihrachyshka has joined #openstack-oslo | 17:24 | |
ihrachyshka | morganfainberg, so I've updated the patch with all we need in neutron: https://review.openstack.org/#/c/159525/ | 17:24 |
ihrachyshka | morganfainberg, but I suspect someone should really go thru the list of other stuff and decide whether it also belongs to public | 17:25 |
sigmavirus24 | ZZelle: I also qualified that with "If it's being used outside of heat" | 17:25 |
sigmavirus24 | I also think we should avoid providing ssl middleware when far better options exist for deployments (and we should be pointing deployers towards them) | 17:26 |
ZZelle | sigmavirus24, it's being used outside of heat :), even outside of OpenStack | 17:26 |
sigmavirus24 | ihrachyshka: I would argue that those specific checks shouldn't be used by people using oslo.policy | 17:27 |
ZZelle | the ssl middleware is wrongly named as it does not provide ssl termination but add support for X-Forwarded defacto standard header | 17:27 |
sigmavirus24 | Just because they're using them now does not mean they should have been using them in the first place, but I haven't seen the examples of people using these checks directly | 17:27 |
sigmavirus24 | ZZelle: which one? there are several (since there is a de facto set of 2 or 3, there is no single de facto header) | 17:28 |
*** ViswaV_ has joined #openstack-oslo | 17:28 | |
sigmavirus24 | I've seen Proto and Protocol at least | 17:28 |
sigmavirus24 | I think I've seen a third | 17:28 |
*** ViswaV has quit IRC | 17:28 | |
*** yamahata has joined #openstack-oslo | 17:29 | |
ihrachyshka | sigmavirus24, http://git.openstack.org/cgit/openstack/neutron/tree/neutron/policy.py#n158 | 17:29 |
sigmavirus24 | ihrachyshka: was that rejected while oslo.policy was in incubation? | 17:31 |
ZZelle | sigmavirus24, an internal project :) | 17:32 |
ihrachyshka | sigmavirus24, define that | 17:32 |
ZZelle | sigmavirus24, and neutron is about to add a public endpoint url when a correct middleware would better feet | 17:32 |
sigmavirus24 | ihrachyshka: so I don't understand the /need/ for subattributes but was optional typing as a subattr ever proposed to be included in oslo.policy while it was in incubation? | 17:32 |
ihrachyshka | sigmavirus24, I don't think so. haven't tracked that. | 17:33 |
dhellmann | sileht: https://bugs.launchpad.net/oslo.messaging/+bug/1426046 | 17:34 |
openstack | Launchpad bug 1426046 in oslo.messaging "there is no public base class for notifier drivers" [Undecided,New] | 17:34 |
*** mriedem is now known as mriedem_away | 17:34 | |
sigmavirus24 | ihrachyshka: it seems like something that should be proposed for inclusion in oslo.policy. I want to see if any other project needs those symbols your making public though | 17:34 |
ihrachyshka | sigmavirus24, some of them should obviously be public (Check, register). are we speaking about specific check types? | 17:35 |
sigmavirus24 | ihrachyshka: those two I agree make sense | 17:36 |
sigmavirus24 | the specific check types seem wrong though | 17:36 |
sigmavirus24 | I'm checking other projects now to see what they use | 17:36 |
ihrachyshka | sigmavirus24, so full disclosure, I don't know much about neutron case, so I will be able to discuss it tomorrow only | 17:37 |
sigmavirus24 | ihrachyshka: so nova would agree that register and Check should be public symobls | 17:37 |
sigmavirus24 | (which I was already +1 for) | 17:37 |
ihrachyshka | sigmavirus24, let's discuss the matter tomorrow or the next week. will there be a release before that time? | 17:40 |
sigmavirus24 | ihrachyshka: probably | 17:40 |
sigmavirus24 | keystone needs it for the changes stevemar is making in keystone | 17:41 |
* stevemar scrolls up | 17:44 | |
stevemar | ihrachyshka, sigmavirus24 i'd be cool with the decorators being public... we could postpone releasing a new version until this is in, they keystone stuff is not *that* necessary to get in | 17:46 |
sigmavirus24 | stevemar: right but this patch isn't totally in the clear yet | 17:48 |
sigmavirus24 | mostly because I've yet to find that neutron is not alone | 17:48 |
ihrachyshka | stevemar, I just need one day to participate in proper discussion on why the neutron usage | 17:49 |
*** salv-orlando has joined #openstack-oslo | 17:49 | |
stevemar | ah, consider why it's being used? | 17:49 |
ihrachyshka | right, I don't know it yet for sure | 17:54 |
*** tedross_ has quit IRC | 18:03 | |
*** kbyrne has quit IRC | 18:03 | |
*** SridharG has joined #openstack-oslo | 18:03 | |
*** kbyrne has joined #openstack-oslo | 18:05 | |
sigmavirus24 | ihrachyshka: they're adding/allowing for additional pieces to be added to policy rules, extensions of sorts | 18:06 |
sigmavirus24 | There's no built-in way for that so they're using those classes directly, which is probably a bad idea | 18:06 |
sigmavirus24 | stevemar: ^ | 18:06 |
stevemar | sigmavirus24, eek | 18:07 |
sigmavirus24 | stevemar: yep | 18:07 |
sigmavirus24 | that's how I read the code at least | 18:07 |
sigmavirus24 | stevemar: mostly type annotations it seems | 18:07 |
openstackgerrit | Ben Nemec proposed openstack/oslo.policy: Add missing space to help message https://review.openstack.org/159558 | 18:09 |
*** ChuckC has quit IRC | 18:17 | |
openstackgerrit | Ben Nemec proposed openstack/oslo-incubator: Do not log on missing or empty policy_dirs https://review.openstack.org/159562 | 18:18 |
*** ajo has quit IRC | 18:18 | |
*** ajo has joined #openstack-oslo | 18:19 | |
*** tedross_ has joined #openstack-oslo | 18:19 | |
*** inc0 has joined #openstack-oslo | 18:19 | |
bnemec | sdague: ^ | 18:21 |
*** ajo has quit IRC | 18:29 | |
*** saikrishna has joined #openstack-oslo | 18:33 | |
*** cdent has quit IRC | 18:34 | |
*** ihrachyshka has quit IRC | 18:37 | |
*** mtanino has joined #openstack-oslo | 18:40 | |
*** ChuckC has joined #openstack-oslo | 18:40 | |
*** ChuckC has quit IRC | 18:47 | |
*** ChuckC has joined #openstack-oslo | 18:49 | |
*** ihrachyshka has joined #openstack-oslo | 18:50 | |
*** dims_ has joined #openstack-oslo | 18:51 | |
*** dims_ has quit IRC | 18:53 | |
*** dims_ has joined #openstack-oslo | 18:54 | |
*** exploreshaifali has quit IRC | 18:55 | |
*** dims has quit IRC | 18:55 | |
*** e0ne is now known as e0ne_ | 19:00 | |
*** prad has quit IRC | 19:01 | |
ihrachyshka | sigmavirus24, stevemar, ok, I discussed the matter with guys. so the plan is to move the sub-attr matching feature under oslo.policy hood, and maybe some more custom checks we maintain (I'll look closer tomorrow). and it means that we won't be able to consume the initial releases anyway. | 19:02 |
sigmavirus24 | ihrachyshka: yeah, y'all will probably want to create some blueprints/specs for that | 19:05 |
sigmavirus24 | stevemar: dhellmann those would go to oslo-specs, right? | 19:06 |
ihrachyshka | sigmavirus24, right. I've reported a neutron bug for now, but yes, seems like a valid case for specs | 19:07 |
ihrachyshka | sigh | 19:07 |
*** SridharG has quit IRC | 19:07 | |
ihrachyshka | I thought it will be as trivial as usual. | 19:07 |
ihrachyshka | :) | 19:07 |
*** jungleboyj has quit IRC | 19:07 | |
sigmavirus24 | ihrachyshka: sorry | 19:07 |
ihrachyshka | :) | 19:07 |
sigmavirus24 | the folks at neutron seem to have had other plans for your time | 19:07 |
sigmavirus24 | No good deed and what not ;) | 19:07 |
*** jungleboyj has joined #openstack-oslo | 19:08 | |
sigmavirus24 | ihrachyshka: would it be okay if we reverted your change back to PS1 to get those symbols released so projects like nova can start using oslo.policy when it wants? | 19:10 |
sigmavirus24 | We'll then discuss adding those symbols or allowing policy syntax extensions on the blueprint(s)/spec(s) | 19:10 |
sigmavirus24 | Is that okay? | 19:10 |
ihrachyshka | yeah, I was going to leave home though. if you can, please do, otherwise I'll do it tomorrow | 19:10 |
ihrachyshka | ok, now leaving | 19:11 |
*** inc0 has quit IRC | 19:11 | |
sigmavirus24 | bye ihrachyshka | 19:15 |
dansmith | dhellmann: https://review.openstack.org/159577 | 19:16 |
*** ihrachyshka has quit IRC | 19:16 | |
dansmith | dhellmann: passes our unit tests, putting it up while I build devstack to see what else needs to change | 19:16 |
dansmith | dhellmann: there are major hacks in there to save some typing and avoid having to refactor some things the right way | 19:16 |
dansmith | dhellmann: and I found a couple of gotchas that I need to fix in the library | 19:16 |
dansmith | dhellmann: so far, nothing fundamentally unusable yet | 19:17 |
dansmith | dhellmann: it also points out a couple places where we are (a) breaking (cheating really) our own model or (b) still need to do some cleanup we thought was done | 19:18 |
dansmith | oh, dangit I forgot the depends-on | 19:19 |
*** ajo has joined #openstack-oslo | 19:22 | |
*** tedross_ has quit IRC | 19:22 | |
*** himangi has quit IRC | 19:26 | |
*** exploreshaifali has joined #openstack-oslo | 19:30 | |
*** saikrishna has quit IRC | 19:38 | |
*** tedross_ has joined #openstack-oslo | 19:39 | |
*** jecarey has quit IRC | 19:43 | |
*** achanda has joined #openstack-oslo | 19:45 | |
*** e0ne_ is now known as e0ne | 19:52 | |
*** exploreshaifali has quit IRC | 19:56 | |
*** jungleboyj_ has joined #openstack-oslo | 20:01 | |
*** jungleboyj has quit IRC | 20:03 | |
*** jecarey has joined #openstack-oslo | 20:03 | |
*** himangi has joined #openstack-oslo | 20:03 | |
*** ajo has quit IRC | 20:03 | |
*** ajo has joined #openstack-oslo | 20:07 | |
*** himangi has quit IRC | 20:15 | |
dhellmann | dansmith: ok, so do you anticipate breaking API changes to fix those issues? | 20:18 |
dhellmann | that's ultimately what we're looking for | 20:18 |
dansmith | dhellmann: I was trying to say: nothing so far needs a breaking change that I foresee | 20:19 |
dansmith | dhellmann: I'm not done yet of course, but just trying to give some incremental progress feedback | 20:19 |
openstackgerrit | Sabari proposed openstack/oslo.vmware: Add get_datastore_by_ref method to oslo.vmware https://review.openstack.org/159357 | 20:19 |
dhellmann | dansmith: ok, I wasn't sure how bit a "gotcha" was :-) | 20:19 |
openstackgerrit | Sabari proposed openstack/oslo.vmware: Add get_datastore_by_ref method to oslo.vmware https://review.openstack.org/159357 | 20:19 |
dansmith | dhellmann: on the level of forgetting to call super() a couple times :) | 20:19 |
dhellmann | dansmith: I'm a bit more comfortable moving ahead with a formal release now, then | 20:19 |
dhellmann | heh, sure | 20:19 |
dhellmann | so bugs but not design issues | 20:20 |
dansmith | dhellmann: yeah, there is one thing that doesn't work exactly like I intended, but it's only a problem because nova needs to override something for compatibility, | 20:20 |
dansmith | dhellmann: shouldn't impact anyone else, and even if I couldn't do another release I could work around it in a not horrible way | 20:21 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/oslo-incubator: Handle non-json http exceptions better https://review.openstack.org/144839 | 20:21 |
dansmith | anyway, let me continue down the devstack and tempest path before "a bit more comfortable" becomes "comfortable enough" | 20:21 |
*** tedross_ has left #openstack-oslo | 20:23 | |
dhellmann | dansmith: sounds good | 20:24 |
*** himangi has joined #openstack-oslo | 20:27 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 20:32 | |
*** himangi has quit IRC | 20:33 | |
*** ajo has quit IRC | 20:34 | |
jogo | dhellmann: https://launchpad.net/openstack | 20:36 |
jogo | is missing all of the oslo libs | 20:37 |
jogo | such as oslo.db | 20:37 |
jogo | the admin for the oslo.db group etc can add it to openstack | 20:37 |
dhellmann | jogo: https://launchpad.net/oslo | 20:37 |
dhellmann | ttx and I moved them to a different project group so we could see unified view of the next-kilo milestones | 20:37 |
dhellmann | it also lets us do some other things in LP that I can't think of off the top of my head | 20:38 |
dhellmann | not ideal, but necessary | 20:38 |
openstackgerrit | Merged openstack/oslo-incubator: Do not log on missing or empty policy_dirs https://review.openstack.org/159562 | 20:38 |
jogo | dhellmann: any way to add that group to openstack? | 20:38 |
dhellmann | jogo: this is also why blueprint links to oslo blueprints from gerrit are broken | 20:38 |
dhellmann | jogo: no, you can't nest project groups in lp | 20:38 |
dhellmann | add this to the list of things that gets better with storyboard | 20:39 |
* jogo blames lifeless for funky launchpad issues | 20:39 | |
*** himangi has joined #openstack-oslo | 20:40 | |
jogo | dhellmann: one big downside is you can't just go to launchpad openstack and click the low hanging fruit bug tag | 20:47 |
jogo | and see oslo stuff | 20:47 |
*** mriedem_away is now known as mriedem | 20:47 | |
dhellmann | jogo: yep, that is another downside | 20:47 |
mtreinish | jogo: why not just have the e-r test look for membership in both groups | 20:47 |
mtreinish | not the best, but at least it'll get you around the issue | 20:48 |
jogo | mtreinish: doing that | 20:48 |
jogo | next up is https://launchpad.net/openstack-api-site | 20:48 |
openstackgerrit | John Stanford proposed openstack/oslo.log: Make use_syslog=True log to syslog via /dev/log https://review.openstack.org/159606 | 20:49 |
openstackgerrit | Erno Kuvaja proposed openstack-dev/hacking: Add own block in imports for six overwrites https://review.openstack.org/159196 | 20:49 |
mtreinish | jogo: we have e-r queries for bugs in that group? | 20:49 |
jogo | mtreinish: yup https://bugs.launchpad.net/openstack-api-site/+bug/1335731 | 20:49 |
openstack | Launchpad bug 1335731 in openstack-api-site "dox-publish build fails to build for identity-api" [Critical,Fix released] - Assigned to Anne Gentle (annegentle) | 20:49 |
mtreinish | morganfainberg: ^^^ fix it... | 20:50 |
mtreinish | jogo: heh, that's weird. but ok | 20:50 |
morganfainberg | uh. | 20:51 |
morganfainberg | what? | 20:51 |
mtreinish | jogo: wait it's marked as fixed :) | 20:51 |
jogo | mtreinish: still has hits | 20:51 |
jogo | fixed recently | 20:51 |
morganfainberg | identity-api shouldn't be built! | 20:51 |
morganfainberg | attic ;) | 20:51 |
mtreinish | morganfainberg: heh, sorry I didn't realize it was fixed released either :) | 20:51 |
mtreinish | jogo: yeah that is weird, where was it hitting, because that was moved to attic more than 10 days ago | 20:52 |
jogo | mtreinish: didn't dig in too much | 20:52 |
morganfainberg | hehe | 20:53 |
jogo | mtreinish: good news is the docs thing is the last one | 20:57 |
*** morganfainberg is now known as needscoffee | 20:57 | |
*** kgiusti has left #openstack-oslo | 21:06 | |
*** rushiagr is now known as rushiagr_away | 21:06 | |
*** jecarey_ has joined #openstack-oslo | 21:14 | |
*** jecarey has quit IRC | 21:17 | |
openstackgerrit | Ihar Hrachyshka proposed openstack/oslo.policy: Expose stuff used in Neutron as part of public API https://review.openstack.org/159525 | 21:20 |
openstackgerrit | Ihar Hrachyshka proposed openstack/oslo.policy: Expose register and Check as part of public API https://review.openstack.org/159525 | 21:21 |
*** ihrachyshka has joined #openstack-oslo | 21:21 | |
mriedem | dhellmann: is there a special wiki for how to release oslo libaries and/or clients to get launchpad updated with a milestone? | 21:23 |
*** tsekiyama has quit IRC | 21:26 | |
*** tsekiyama has joined #openstack-oslo | 21:26 | |
openstackgerrit | Merged openstack/oslo.policy: Add missing space to help message https://review.openstack.org/159558 | 21:26 |
*** jungleboyj_ has quit IRC | 21:27 | |
*** achanda has quit IRC | 21:34 | |
dhellmann | mriedem: yes, there is! let me find that link... | 21:34 |
dhellmann | mriedem: https://wiki.openstack.org/wiki/Oslo/ReleaseProcess | 21:34 |
mriedem | dhellmann: ah, thanks | 21:34 |
*** achanda has joined #openstack-oslo | 21:35 | |
mriedem | dhellmann: also, btw, just sanity check on version - we have python-novaclient 2.21 today, we have like 1 or 2 changes on master, no requirements updates, just bug fixes for what busted 2.21, so i'm assuming the version is then 2.21.1? | 21:35 |
dhellmann | mriedem: let me look at the git logs, but that sounds right | 21:35 |
mriedem | http://paste.openstack.org/show/182694/ | 21:36 |
mriedem | nothing for: edem@ubuntu:~/git/python-novaclient$ git log --oneline --no-merges 2.21.0.. -- requirements.txt | 21:37 |
dhellmann | that revert looks like it might require a minor update, what's going on there? | 21:37 |
dhellmann | I guess that's a bug fix | 21:38 |
mriedem | that bash completion change busted volume-attach and was in 2.21 | 21:38 |
mriedem | yeah | 21:38 |
mriedem | we didn't release since last sept | 21:38 |
mriedem | so it lurked forever | 21:38 |
dhellmann | e6883d2 looks like a new feature | 21:38 |
mriedem | says partial bug | 21:38 |
openstackgerrit | Steve Martinelli proposed openstack/oslo-incubator: Remove policy from oslo-incubator https://review.openstack.org/152812 | 21:39 |
mriedem | dhellmann: that just makes novaclient more consistent with other clients | 21:39 |
openstackgerrit | Steve Martinelli proposed openstack/oslo-incubator: Prevent update.py from updating policy https://review.openstack.org/152813 | 21:39 |
dhellmann | mriedem: when in doubt you want to bump the minor version. The exceptional cases will lead you to bumping the patch or major version. | 21:39 |
dhellmann | The "bug" here was that nova wasn't being consistent with the other projects. | 21:40 |
dhellmann | To make it consistent, it looks like a new feature was added. | 21:40 |
mriedem | ok, sure, i could see that as a wishlist bug | 21:41 |
mriedem | i.e. feature | 21:41 |
dhellmann | a new behavior was added that modifies the input provided by the user so they can use an abbreviation instead of the full value | 21:41 |
dhellmann | yeah, just because it's a bug in launchpad doesn't mean it's a semver bug :-) | 21:41 |
dhellmann | those other things do look patch-level worthy, but this one makes me think it'd be better to call it 2.22.0 | 21:42 |
mriedem | dhellmann: ok, will do, thanks for looking | 21:44 |
mriedem | will do monday anyway :) | 21:44 |
dhellmann | mriedem: glad to help, any time! | 21:45 |
*** takedakn has quit IRC | 21:45 | |
dansmith | dhellmann: ohai | 21:51 |
dhellmann | dansmith: greetings | 21:51 |
dansmith | dhellmann: https://review.openstack.org/#/c/159577/ | 21:51 |
dansmith | pep8 failed because I didn't care, requirements because, well, you know, and grenade for some glance reason | 21:51 |
dansmith | dhellmann: please feel free to be amazed :) | 21:51 |
dansmith | dhellmann: but, the interesting bit is that the partial-ncpu job worked, which is juno nova using in-tree objects, talking over the wire to this change using o.vo objects | 21:52 |
dansmith | dhellmann: which means they're interoperable over the wire as well, hence we should be good rolling to these with nova in a live upgrade situation | 21:53 |
*** e0ne has quit IRC | 21:54 | |
*** jecarey_ is now known as jecarey | 21:56 | |
dhellmann | dansmith: you've boosted my confidence over the threshold. | 21:58 |
dansmith | \o/ | 21:58 |
dhellmann | dansmith: does that mean we're ready to add 0.1.0 to the global requirements list, or do we need another release first | 21:58 |
dansmith | dhellmann: I didn't have to change anything in o.vo, this is running against the pypi version | 21:59 |
dansmith | dhellmann: now, I could clean up the few things I found if it would make you feel better | 21:59 |
dansmith | and/or we can wait for inc0 tomorrow and I can discuss some of the things I'm working around | 22:00 |
dhellmann | it's up to you, you're closer to the details than me | 22:00 |
dansmith | let's wait for tomorrow morning | 22:00 |
dansmith | I'll send him an email tonight and start working on some patches to what I need | 22:00 |
dhellmann | dansmith: sounds good, and thanks for pushing so hard on that today | 22:01 |
dansmith | dhellmann: one question I had was, are stackforge projects limited by g-r? | 22:01 |
dhellmann | they can do a "soft sync" to keep up to date while still using new things of their own | 22:01 |
dhellmann | we don't let them add new requirements to the global list, though | 22:02 |
dansmith | so it's not like g-r is blocking them from using this, and thus it's not like we can magically break our api between 0.1.0 and 0.1.1 because it never made it into g-r | 22:02 |
*** takedakn has joined #openstack-oslo | 22:02 | |
*** gordc has quit IRC | 22:04 | |
dhellmann | dansmith: true. This is all about managing expectations. If it's in g-r, then the expectation is we have deemed it safe to use. If not, they're aware they are on their own. | 22:04 |
dansmith | dhellmann: okay | 22:04 |
dhellmann | the integrated projects would be blocked from adopting it, and keeping them from breaking is more important as far as the gate | 22:05 |
dansmith | I didn't realize g-r had any such implication I guess | 22:05 |
dansmith | yeah | 22:05 |
*** e0ne has joined #openstack-oslo | 22:05 | |
dhellmann | I'm probably excessively conservative about first releases, but that seems better than any of the alternatives I've seen so far. :-) | 22:06 |
*** takedakn has quit IRC | 22:11 | |
*** needscoffee is now known as needsmostcoffee | 22:12 | |
*** needsmostcoffee is now known as morganfainberg | 22:13 | |
*** stevemar is now known as thinksmorganneed | 22:13 | |
*** thinksmorganneed is now known as stevemar | 22:13 | |
*** jungleboyj_ has joined #openstack-oslo | 22:16 | |
*** takedakn has joined #openstack-oslo | 22:19 | |
*** jgrimm is now known as zz_jgrimm | 22:20 | |
*** takedakn has quit IRC | 22:36 | |
*** takedakn has joined #openstack-oslo | 22:36 | |
*** jaosorior has quit IRC | 22:42 | |
*** rodrigods is now known as rodrigod` | 22:42 | |
*** rodrigod` is now known as rodrigods | 22:43 | |
*** takedakn has quit IRC | 22:43 | |
*** rodrigods is now known as rodrigod` | 22:51 | |
*** rodrigod` is now known as rodrigods | 22:51 | |
*** achanda has quit IRC | 22:51 | |
*** andreykurilin_ has joined #openstack-oslo | 22:53 | |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/oslo-incubator: Expose opts entry point for version_utils https://review.openstack.org/159646 | 22:53 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/oslo-incubator: Expose opts entry point for version_utils https://review.openstack.org/159646 | 22:54 |
stevemar | dhellmann, fwiw, i think most keystoners are in favor of https://review.openstack.org/#/c/148624/ in it's current state, esp. since there is a follow up patch to clean up tests. | 22:59 |
stevemar | There was a few patch proposed in oslo.policy too, to surface a few other functions | 22:59 |
*** achanda has joined #openstack-oslo | 23:02 | |
openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects: Make serializer use the provided base class for the indirection api https://review.openstack.org/159656 | 23:08 |
openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects: Allow passing serializer and indirection API objects to Fixture https://review.openstack.org/159657 | 23:08 |
openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects: Properly serialize/deserialize arguments in fake indirection api https://review.openstack.org/159658 | 23:08 |
*** bknudson has quit IRC | 23:09 | |
*** takedakn has joined #openstack-oslo | 23:10 | |
*** andreykurilin_ has quit IRC | 23:15 | |
*** andreykurilin_ has joined #openstack-oslo | 23:15 | |
*** stevemar has quit IRC | 23:19 | |
*** ihrachyshka has quit IRC | 23:19 | |
openstackgerrit | Dan Smith proposed openstack/oslo.versionedobjects: Properly serialize/deserialize arguments in fake indirection api https://review.openstack.org/159658 | 23:21 |
*** jecarey has quit IRC | 23:25 | |
*** andreykurilin_ has quit IRC | 23:26 | |
*** andreykurilin__ has joined #openstack-oslo | 23:26 | |
*** trown is now known as trown|outttypeww | 23:27 | |
*** mriedem has quit IRC | 23:36 | |
*** noelbk has quit IRC | 23:37 | |
dhellmann | stevemar: ack, I'll take a look in the morning | 23:41 |
*** amotoki has joined #openstack-oslo | 23:43 | |
*** amotoki has quit IRC | 23:44 | |
*** jungleboyj_ has quit IRC | 23:47 | |
*** daniel3_ has quit IRC | 23:48 | |
*** jungleboyj_ has joined #openstack-oslo | 23:53 | |
*** himangi has quit IRC | 23:54 | |
*** dims_ has quit IRC | 23:55 | |
*** salv-orlando has quit IRC | 23:57 | |
*** andreykurilin__ has quit IRC | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!