| *** matsuhashi has joined #savanna | 00:06 | |
| *** matsuhashi has quit IRC | 00:21 | |
| *** matsuhashi has joined #savanna | 00:21 | |
| *** matsuhashi has quit IRC | 00:25 | |
| *** matsuhas_ has joined #savanna | 00:25 | |
| *** tmckay is now known as _tmckay | 00:53 | |
| *** jmaron has quit IRC | 00:55 | |
| *** qwerty_nor has quit IRC | 01:01 | |
| *** nosnos has joined #savanna | 01:01 | |
| *** ErikB has quit IRC | 01:39 | |
| *** ErikB has joined #savanna | 01:58 | |
| *** xuhanp has joined #savanna | 02:02 | |
| *** akuznets_ has joined #savanna | 02:08 | |
| *** akuznetsov has quit IRC | 02:09 | |
| *** ErikB has quit IRC | 02:10 | |
| *** IlyaE has quit IRC | 02:12 | |
| openstackgerrit | Andrew Lazarev proposed a change to openstack/savanna: [IDH] Fixed cluster start without jobtracker service https://review.openstack.org/72817 | 02:36 |
|---|---|---|
| *** ErikB has joined #savanna | 02:41 | |
| *** ErikB has quit IRC | 02:45 | |
| *** witlessb has joined #savanna | 02:56 | |
| *** witlessb has quit IRC | 02:56 | |
| *** ErikB has joined #savanna | 03:37 | |
| *** matsuhas_ has quit IRC | 03:39 | |
| *** ErikB has quit IRC | 03:41 | |
| *** crobertsrh has joined #savanna | 04:31 | |
| *** crobertsrh is now known as _crobertsrh | 04:33 | |
| *** ErikB has joined #savanna | 04:36 | |
| *** ErikB has quit IRC | 04:41 | |
| *** matsuhashi has joined #savanna | 04:52 | |
| *** Ch00k has joined #savanna | 05:22 | |
| *** Ch00k has quit IRC | 05:34 | |
| *** Ch00k has joined #savanna | 05:34 | |
| *** aignatov_ is now known as aignatov | 05:36 | |
| *** ErikB has joined #savanna | 05:37 | |
| *** Ch00k has quit IRC | 05:39 | |
| *** Ch00k has joined #savanna | 05:40 | |
| *** ErikB has quit IRC | 05:41 | |
| *** nosnos_ has joined #savanna | 05:43 | |
| *** nosnos_ has quit IRC | 05:44 | |
| *** nosnos_ has joined #savanna | 05:44 | |
| *** DinaBelova_ is now known as DinaBelova | 05:45 | |
| *** nosnos has quit IRC | 05:46 | |
| *** _nadya_ has joined #savanna | 06:01 | |
| *** saju_m has joined #savanna | 06:02 | |
| *** nosnos_ has quit IRC | 06:04 | |
| *** nosnos has joined #savanna | 06:04 | |
| openstackgerrit | Jenkins proposed a change to openstack/savanna: Imported Translations from Transifex https://review.openstack.org/70918 | 06:09 |
| *** Ch00k has quit IRC | 06:22 | |
| *** Ch00k has joined #savanna | 06:23 | |
| *** _nadya_ has quit IRC | 06:24 | |
| *** Ch00k has quit IRC | 06:35 | |
| openstackgerrit | Andrew Lazarev proposed a change to openstack/savanna: Fixed hadoop dir creation during hadoop-swift lib download https://review.openstack.org/72847 | 06:36 |
| *** ErikB has joined #savanna | 06:37 | |
| openstackgerrit | Andrew Lazarev proposed a change to openstack/savanna: Fixed hadoop dir creation during hadoop-swift lib download https://review.openstack.org/72847 | 06:37 |
| *** ErikB has quit IRC | 06:41 | |
| *** bogdando has quit IRC | 06:44 | |
| *** bogdando has joined #savanna | 06:45 | |
| *** aignatov is now known as aignatov_ | 07:00 | |
| *** aignatov_ is now known as aignatov | 07:02 | |
| *** tnovacik has joined #savanna | 07:09 | |
| *** aignatov is now known as aignatov_ | 07:19 | |
| openstackgerrit | A change was merged to openstack/savanna: Remove compatibility code allowing "args" as dict https://review.openstack.org/72689 | 07:20 |
| *** coolsvap has joined #savanna | 07:21 | |
| *** coolsvap1 has joined #savanna | 07:26 | |
| *** coolsvap has quit IRC | 07:28 | |
| *** Ch00k has joined #savanna | 07:34 | |
| *** coolsvap1 has quit IRC | 07:36 | |
| *** coolsvap has joined #savanna | 07:36 | |
| *** ErikB has joined #savanna | 07:37 | |
| *** ErikB has quit IRC | 07:41 | |
| *** IlyaE has joined #savanna | 08:04 | |
| *** mattf is now known as _mattf | 08:25 | |
| *** witlessb has joined #savanna | 08:28 | |
| *** boris-42_ has joined #savanna | 08:30 | |
| *** ErikB has joined #savanna | 08:37 | |
| *** IlyaE has quit IRC | 08:41 | |
| *** ErikB has quit IRC | 08:41 | |
| *** Ch00k has quit IRC | 08:41 | |
| *** Ch00k has joined #savanna | 08:43 | |
| *** dmitryme has joined #savanna | 08:58 | |
| *** coolsvap1 has joined #savanna | 08:59 | |
| *** coolsvap has quit IRC | 09:01 | |
| *** aignatov_ is now known as aignatov | 09:02 | |
| openstackgerrit | Yongli He proposed a change to openstack/savanna: Use six.moves.urllib.parse instead of urlparse https://review.openstack.org/72883 | 09:07 |
| openstackgerrit | Sergey Reshetnyak proposed a change to openstack/savanna-image-elements: Add DIB work dir to gitignore https://review.openstack.org/72891 | 09:23 |
| *** ErikB has joined #savanna | 09:36 | |
| *** DinaBelova is now known as DinaBelova_ | 09:40 | |
| *** ErikB has quit IRC | 09:41 | |
| *** _mattf is now known as mattf | 10:24 | |
| *** xuhanp has quit IRC | 10:28 | |
| openstackgerrit | A change was merged to openstack/python-savannaclient: Removed copy_if_defined function from ClusterManager https://review.openstack.org/72493 | 10:30 |
| *** ErikB has joined #savanna | 10:37 | |
| *** ErikB has quit IRC | 10:41 | |
| *** matsuhashi has quit IRC | 10:41 | |
| *** aignatov is now known as aignatov_ | 10:41 | |
| openstackgerrit | Yaroslav Lobankov proposed a change to openstack/savanna: Cinder test to integration tests was added https://review.openstack.org/68932 | 10:43 |
| openstackgerrit | Yaroslav Lobankov proposed a change to openstack/savanna: Cinder test to integration tests was added https://review.openstack.org/68932 | 10:47 |
| *** saju_m has quit IRC | 10:57 | |
| *** aignatov_ is now known as aignatov | 11:21 | |
| *** dmitryme has quit IRC | 11:31 | |
| *** aignatov is now known as aignatov_ | 11:37 | |
| *** ErikB has joined #savanna | 11:37 | |
| *** saju_m has joined #savanna | 11:40 | |
| *** ErikB has quit IRC | 11:41 | |
| *** dmitryme has joined #savanna | 11:42 | |
| *** matsuhashi has joined #savanna | 11:44 | |
| *** saju_m has quit IRC | 11:49 | |
| *** DinaBelova_ is now known as DinaBelova | 12:04 | |
| *** Ch00k has quit IRC | 12:08 | |
| *** coolsvap_ has joined #savanna | 12:14 | |
| *** coolsvap1 has quit IRC | 12:17 | |
| *** Ch00k has joined #savanna | 12:25 | |
| *** aignatov_ is now known as aignatov | 12:26 | |
| openstackgerrit | Matthew Farrellee proposed a change to openstack/savanna: Fix swift data source credential validation https://review.openstack.org/72933 | 12:36 |
| *** ErikB has joined #savanna | 12:36 | |
| *** ErikB has quit IRC | 12:41 | |
| *** aignatov is now known as aignatov_ | 12:45 | |
| *** aignatov_ is now known as aignatov | 12:48 | |
| openstackgerrit | A change was merged to openstack/savanna-image-elements: Add options for Spark and Vanilla plugins https://review.openstack.org/72310 | 12:55 |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack/savanna: Fix thread local usage in conext https://review.openstack.org/72655 | 13:03 |
| *** saju_m has joined #savanna | 13:05 | |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack/savanna: Fix thread local usage in conext https://review.openstack.org/72655 | 13:06 |
| *** saju_m has quit IRC | 13:06 | |
| *** saju_m has joined #savanna | 13:07 | |
| *** qwerty_nor has joined #savanna | 13:12 | |
| * mattf mumbles about run_tests in savanna | 13:20 | |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack/savanna: Fix thread local usage in conext https://review.openstack.org/72655 | 13:27 |
| SergeyLukjanov | mattf, it's just not working? | 13:27 |
| openstackgerrit | Matthew Farrellee proposed a change to openstack/savanna: Fix swift data source credential validation https://review.openstack.org/72933 | 13:28 |
| SergeyLukjanov | mattf, note that discovery for tests should be executed under the corresponding (unit and integration) | 13:29 |
| mattf | SergeyLukjanov, the unit / integration split w/ an env var blocks me from adding run_tests.sh. run_tests.sh handles pep8. so i forgot to manually pep8 when i ran tools/run_tests | 13:29 |
| mattf | i don't know how to get something into the env for the discovery except via the tox setenv | 13:30 |
| * mattf longs for a single test entry point for savanna | 13:30 | |
| SergeyLukjanov | mattf, IMO run_tests.sh should only run unit tests :) | 13:32 |
| mattf | IMO too | 13:35 |
| mattf | but the decision to pick unit or integration is out of run_tests.sh's control | 13:36 |
| *** _tmckay is now known as tmckay | 13:36 | |
| *** ErikB has joined #savanna | 13:36 | |
| SergeyLukjanov | mattf, probably it should set env by 'export' for running unit tests? | 13:38 |
| mattf | somewhere in the chain the env gets wiped out, so export DISVER_DIR (or whatever it is) doesn't make it to the testr discovery call | 13:39 |
| SergeyLukjanov | mattf, myabe just 'BLA_BLA=/path command'? | 13:40 |
| dmitryme | matff, as for me, I just run fox -e py27,pep8. I never use tools/run_tests.sh | 13:40 |
| dmitryme | fox -> tox :-) | 13:41 |
| *** ErikB has quit IRC | 13:41 | |
| mattf | dmitryme, yeah, call me lazy. i like to just run ./run_test.sh (tab completes) in each repo | 13:42 |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack/savanna: Fix thread local usage in context https://review.openstack.org/72655 | 13:42 |
| tmckay | dmitryme, me too. And occasionally I use python -m testtools so that I can set break points in the tests | 13:42 |
| mattf | $ grep DIS run_tests.sh | 13:43 |
| mattf | env DISCOVER_DIRECTORY=savanna/tests/unit ${wrapper} $TESTRTESTS | 13:43 |
| mattf | SergeyLukjanov, ^^ my latest attempt from a week or so ago | 13:43 |
| * mattf stash pops and tries again | 13:43 | |
| mattf | i got a new laptop, so i have all new installed versions of things | 13:44 |
| dmitryme | mattf: I actually use alias for that. Just discovered that in my bash it autocompletes as well | 13:44 |
| mattf | dmitryme, point dmitryme | 13:44 |
| mattf | fyi, i had to 'pip install -U tox==1.6.1 virtualenv==1.10.1' this morning to get tox to work. known bug in 1.7.0. | 13:45 |
| SergeyLukjanov | mattf, correct, we've faced this problem in the gate | 13:46 |
| mattf | whoa, promising | 13:47 |
| mattf | $ grep DIS run_tests.sh | 13:47 |
| mattf | export DISCOVER_DIRECTORY=savanna/tests/unit | 13:47 |
| mattf | ^^ before wrapper call and it worked this time | 13:47 |
| SergeyLukjanov | mattf, yay! | 13:47 |
| mattf | woo, and changing that to integration -> connection refused | 13:48 |
| mattf | i've no idea what dep was causing the problem tho, hmmm | 13:48 |
| mattf | btw, i'm seeing a bunch of deprecation warnings in the test runs now (new since i upgraded to f20) | 13:49 |
| openstackgerrit | Matthew Farrellee proposed a change to openstack/savanna: Add run_test.sh for running tests https://review.openstack.org/72948 | 13:50 |
| mattf | here goes, someone else can try it too ^^ | 13:50 |
| * mattf does a little jig | 13:50 | |
| mattf | glad to have that off my git stash queue | 13:51 |
| *** ErikB has joined #savanna | 13:56 | |
| mattf | SergeyLukjanov, is this pecan/wsme stuff going to somehow auto generate validations? | 13:58 |
| ruhe | mattf: pecan/wsme can perform basic validations based on wsme models | 13:59 |
| *** jmaron has joined #savanna | 14:00 | |
| mattf | i'm looking at the calendar and at the pecan/wsme patch and getting concerned about hitting i3 | 14:00 |
| mattf | i'm really starting to thing that swapping frameworks is too risky for icehouse | 14:01 |
| SergeyLukjanov | mattf, we're not switching the framework | 14:02 |
| *** coolsvap_ has quit IRC | 14:02 | |
| ruhe | mattf: i thought the idea is to bring pecan into savanna in icehouse, but keep old API ad default | 14:02 |
| SergeyLukjanov | only adding it for v2 api, that'll not be stable in I awywau | 14:03 |
| SergeyLukjanov | anyway* | 14:03 |
| mattf | v2 could absolutely be stable in I | 14:03 |
| SergeyLukjanov | and we could continue work on v2 api after feature freeze if it'll be self-containing | 14:03 |
| mattf | my mind is headed to: have pecan/wsme w/ partial v2 api or have v2 api & tooling for pecan/wsme conversion | 14:04 |
| mattf | my mind is headed to: have (pecan/wsme w/ partial v2 api) or (have v2 api & tooling for pecan/wsme conversion) | 14:04 |
| SergeyLukjanov | mattf, v2 pi will not be released w/o pecan/wsme to avoid migration problems and regressions in future, so, IMO it's useful to do it under the current API framework | 14:05 |
| mattf | "will not"? | 14:06 |
| *** IvanBerezovskiy has joined #savanna | 14:06 | |
| SergeyLukjanov | it's a requirement to have new apis working on top of standard framework | 14:06 |
| mattf | we can always pecan/wsme for v3, where v3 has the enhanced functionality that's been desired for sometime | 14:06 |
| SergeyLukjanov | I'd like to see the v2 api impl completed early in J | 14:07 |
| mattf | where's the requirement coming from? | 14:07 |
| SergeyLukjanov | polish it during the J and make it default in J release | 14:07 |
| SergeyLukjanov | we've agreed on it at summit and pecan/wsme was approved as a standard solution | 14:07 |
| mattf | it's good sw practice to have the deps stable before you move to them | 14:07 |
| SergeyLukjanov | IMO we should have at least several dev milestones between implemented api and it's 'stable' release | 14:08 |
| mattf | agreed on it is kinda loose interpretation. imho it wasn't an issue to use pecan/wsme to long as it didn't threaten inclusion of v2 | 14:08 |
| mattf | we're just a few weeks from i3 and don't have a base for an api impl w/ pecan/wsme | 14:09 |
| SergeyLukjanov | mattf, we need to land a base before the i3 and then you can work on v2 impl inside the api/v2 package | 14:09 |
| SergeyLukjanov | mattf, because it'll not affect anything except the v2 api, that'll be available as a 'draft' api | 14:10 |
| SergeyLukjanov | w/o any guarantees that it works | 14:10 |
| mattf | what's driving the pecan/wsme desire? | 14:10 |
| ruhe | mattf: imho, it's strange OpenStack culture to use the newest and coolest frameworks :) | 14:11 |
| * mattf lols | 14:11 | |
| SergeyLukjanov | what does you mean - why OpenStack folks like pecan/wsme or why I'd like to leverage it? | 14:12 |
| mattf | ruhe, newest and coolest is fun, so long as you also add "functional" | 14:12 |
| SergeyLukjanov | if #1 - I don't know, IMO it sucks, but it's a standard now | 14:12 |
| mattf | who is currently using pecan/wsme for their service? | 14:13 |
| SergeyLukjanov | if #2 I'm strongly sure that we shouldn't release new API versions each release cycle and so don't want to postpone pecan/wsme leverage to the next releases | 14:13 |
| ruhe | mattf: Ceilometer, Tuskar, Ironic | 14:13 |
| SergeyLukjanov | mattf, ceilo, ironic, tuskar, storyboard wip, climate wip | 14:13 |
| ruhe | mattf: also most of the major projects have blueprints (some of them are in progress) to migrate to pecan/wsme | 14:14 |
| mattf | of those ceilometer is the most relevant | 14:14 |
| SergeyLukjanov | mattf, AFAIK nova and glance are moving their currently draft new (v3) APIs to pecan/wsme to standardize frameworks | 14:14 |
| mattf | what about the other cores? | 14:14 |
| mattf | when are nova & glance doing it? | 14:14 |
| *** _crobertsrh is now known as crobertsrh | 14:14 | |
| ruhe | mattf: J i think | 14:14 |
| SergeyLukjanov | mattf, they are polishing their APIs and AFAIK there was no plans to make them default in I | 14:15 |
| ruhe | speaking seriously - it's good. all OpenStack projects should have the same standartized and unified infrastructure code. it helps new developers to focus in domain logic instead of battling with various frameworks and libraries | 14:15 |
| SergeyLukjanov | ruhe, exactly | 14:16 |
| mattf | i'm not objecting to the goal, i'm questioning our execution | 14:16 |
| SergeyLukjanov | mattf, what's you concern now? I'm afraid that I misunderstand you atm | 14:17 |
| mattf | afaict, we can either have a v2 api or the groundwork for pecan/wsme in I, not both | 14:17 |
| ruhe | mattf: this is a valid question. my concerns is - do we have enough tests to make sure that new api based on pecan has the same functionality and exposes the same api | 14:17 |
| ruhe | mattf: based on my experience with pecan i agree with you | 14:18 |
| mattf | ruhe, that is what is sparking my concern right now. switching from flask to pecan is more than getting the model right. there are a lot of tests to write and i don't see examples of unit/func tests for pecan | 14:18 |
| SergeyLukjanov | mattf, we have an example in storyboard and it'll be moved to savanna too | 14:19 |
| mattf | the ability to do validations and unit tests w/ pecan are unknown to me, so i'm over estimating the effort to do them | 14:19 |
| ruhe | mattf: functional tests for savanna/pecan are in my backlog. but it's a lot of work. first step (imho, others may correct me)- is to have enough functional tests for existing api | 14:20 |
| * mattf assumes he is overestimating | 14:20 | |
| mattf | ruhe, existing being v1* ? | 14:21 |
| ruhe | mattf: yes | 14:21 |
| mattf | i'm going to spend my time w/ tests for v2 | 14:21 |
| *** saju_m has quit IRC | 14:21 | |
| mattf | i'll noodle on this some more, but if we cannot do v2 w/o a pecan base and pecan isn't going to be ready until the end of I, then we cannot have v2 as part of I | 14:22 |
| mattf | a concern i have there is our first release where we're considered integrated (hope), we'll have an api that is being deprecated. | 14:23 |
| ruhe | mattf: what's your estimate to build v2 on top of current code (without pecan)? | 14:25 |
| SergeyLukjanov | mattf, we can't have a released v2 as part of I, two weeks left | 14:25 |
| SergeyLukjanov | mattf, it could be only as draft api version | 14:25 |
| *** aignatov is now known as aignatov_ | 14:25 | |
| SergeyLukjanov | so, for me the diff is to have (draft v2 api as part of I based on current code and then migrate it to pecan in J w/ probably some changes, that's ok because it was not release yet) vs. (include pecan base in I with draft v2 api probably not fully completed and than finalize it in J) | 14:27 |
| *** aignatov_ is now known as aignatov | 14:28 | |
| SergeyLukjanov | additionally, I'm concerned on releasing new API with only enpoints cleanup, probably we need models cleanup too | 14:28 |
| SergeyLukjanov | like hadoop_version prop. | 14:28 |
| SergeyLukjanov | I'm not really against drafting v2 on flask, but it looks like a big overcommitment | 14:30 |
| ruhe | SergeyLukjanov: so, what's the plan? bring pecan/wsme into savanna in I. what else? | 14:31 |
| SergeyLukjanov | ruhe, I'd like to bring it into savanna when it'll be ready, preferably in I | 14:32 |
| ruhe | SergeyLukjanov: but keep existing v1flask api as default, right? | 14:32 |
| SergeyLukjanov | ruhe, if mattf strongly sure that he'd like to spent time drafting API on flask - I'm not against it | 14:33 |
| SergeyLukjanov | ruhe, anyway, we keep v1 default till the end of J | 14:33 |
| SergeyLukjanov | ruhe, it's impossible to make one API default in two weeks or two month too | 14:33 |
| ruhe | SergeyLukjanov: sure, that's i'm asking :) | 14:33 |
| SergeyLukjanov | we should migrate all integration tests, client, tempest test too | 14:33 |
| ruhe | *that's why i'm asking | 14:34 |
| SergeyLukjanov | and I'm afraid that we should cleanup model too and give up the v1 api in J | 14:38 |
| *** saju_m has joined #savanna | 14:38 | |
| *** Ch00k has quit IRC | 14:44 | |
| *** Ch00k has joined #savanna | 14:48 | |
| *** IlyaE has joined #savanna | 14:56 | |
| *** dmitryme has quit IRC | 15:00 | |
| mattf | ruhe, estimate - given the reuse that can be leveraged from v1, i expect by i3 | 15:05 |
| mattf | SergeyLukjanov, v2 should be default during J. we should be targeting pecan for when the bulk of the integrated projects are targetting it. i've a feeling this communities drive will reduce compat issues to near 0, or provide everyone with a standard upgrade path. | 15:06 |
| mattf | v2 can't be default until J is released. we need a cycle where we have a base v2 that users can migrate to before we declare it default | 15:07 |
| mattf | tho, i'm not entirely sure we agree on what "default" means, because we don't really have a plan for removing old apis | 15:08 |
| SergeyLukjanov | what's your thoughts about the plan with some time stamps? | 15:14 |
| SergeyLukjanov | mattf, when I’ve proposed the v2 api for I at design summit, the idea was to polish it, migrate to Pecan/WSME and release as part of I as the main/stable API that would be used by our users | 15:23 |
| SergeyLukjanov | mattf, and let me try to explain my concerns: | 15:24 |
| SergeyLukjanov | 1) we could name new API stable only after it'll be covered by tests (integration at least, functional preferred); | 15:24 |
| SergeyLukjanov | 2) we need about 2 dev milestones to check that API really fits our needs (if any non-style changes like endpoints cleanup); | 15:24 |
| SergeyLukjanov | 3) I think that we should have either model cleanup or significant changes to endpoints to release new *major* API version; | 15:24 |
| SergeyLukjanov | 4) I'd really like to have a Pecan/WSME leveraged for new API to avoid situation when we'll need to migrate working/stable API to new framework; | 15:24 |
| SergeyLukjanov | 5) probably something else | 15:24 |
| SergeyLukjanov | mattf, so, please, explain how you see the v2 api implementation and graduation from draft and we could discuss it much more productive | 15:25 |
| SergeyLukjanov | btw initially, I thought that we should have working impl of v2 api before i1 to name it stable in I | 15:25 |
| SergeyLukjanov | oh, one more thing, personally, I think that it's quite strange and useless to release *major* API version with only endpoints clenup | 15:27 |
| SergeyLukjanov | if we'd like to do it, v1.2 could work too | 15:27 |
| SergeyLukjanov | on the current code sure | 15:27 |
| SergeyLukjanov | nope, that's crazy idea, it'll not be compatible with v1, so, don't read it ;) | 15:28 |
| *** jmaron has quit IRC | 15:29 | |
| *** akuznets_ has quit IRC | 15:32 | |
| *** aignatov is now known as aignatov_ | 15:33 | |
| *** sballe has joined #savanna | 15:33 | |
| ruhe | so, there is no need to release v2 api. just polish v1 and let it be v1 | 15:33 |
| ruhe | and migrate to pecan without any rush | 15:34 |
| ruhe | and even when migration to pecan is complete it is not v2 unless there are major changes | 15:36 |
| *** akuznetsov has joined #savanna | 15:36 | |
| *** akuznetsov has quit IRC | 15:36 | |
| *** coolsvap has joined #savanna | 15:37 | |
| *** akuznetsov has joined #savanna | 15:39 | |
| *** akuznetsov has quit IRC | 15:40 | |
| *** jmaron has joined #savanna | 15:40 | |
| SergeyLukjanov | anyway, v1.2 still could be discussed | 15:44 |
| *** witlessb_ has joined #savanna | 15:46 | |
| *** aignatov_ is now known as aignatov | 15:47 | |
| *** venza_ has joined #savanna | 15:52 | |
| *** witlessb has quit IRC | 15:56 | |
| *** venza has quit IRC | 15:56 | |
| *** witlessb_ is now known as witlessb | 15:56 | |
| *** coolsvap1 has joined #savanna | 16:01 | |
| *** coolsvap has quit IRC | 16:03 | |
| *** nosnos has quit IRC | 16:12 | |
| *** tnovacik has quit IRC | 16:13 | |
| *** IvanBerezovskiy has left #savanna | 16:20 | |
| *** dmitryme has joined #savanna | 16:25 | |
| *** Ch00k has quit IRC | 16:28 | |
| *** tmckay is now known as _tmckay | 16:35 | |
| *** aignatov is now known as aignatov_ | 16:44 | |
| *** coolsvap1 has quit IRC | 16:44 | |
| mattf | back, SergeyLukjanov still there? | 16:53 |
| SergeyLukjanov | mattf, gtg in 5 mins | 16:53 |
| mattf | back later tonight? | 16:54 |
| SergeyLukjanov | mattf, yup, but could be quite overloaded for realtime chat | 16:54 |
| mattf | ok, high latency irc mode chat... | 16:54 |
| SergeyLukjanov | mattf, anyway, you can write here, I'll read the scrollback | 16:54 |
| mattf | i agree cleanup can't be 1.2 (some would argue 1.2 is indicative of a problem already) | 16:55 |
| mattf | (1) makes sense - label stable after covered by tests | 16:55 |
| openstackgerrit | Dmitry Mescheryakov proposed a change to openstack/savanna: Do eventlet.monkey_patch it the very beginning https://review.openstack.org/72992 | 16:56 |
| mattf | i think i agree w/ (2) - i read it as we need to migrate dashboard & cli to new api before we can declare it "default" aka declare v1 deprecated | 16:56 |
| SergeyLukjanov | yup, I think that we'd like to have one api declared as stable | 16:57 |
| mattf | re (3) - adding model cleanup other than things that maintain compatibility, such as relaxing name+id for node groups, is a major effort and shouldn't block v2 | 16:57 |
| mattf | as for (4) - i'd like to land pecan/wsme when other integrated projects do. my gut feeling (yeah, just my exp here) says that when nova + glance + cinder + neutron get to use pecan some of the issues we're anticipating will get resolved | 16:58 |
| mattf | there's risk in us going one wya w/ pecan and then others going another. we may have 2 times when our wire encoding compat changes | 16:59 |
| *** coolsvap has joined #savanna | 17:00 | |
| mattf | it's safer in my mind to have v2 then convert flask->pecan and if that break compat we can go to v3. it'll be breaking compat for others too, so we can stick w/ the group. | 17:00 |
| * mattf hopes no one takes him as an example of proper English | 17:00 | |
| mattf | having v2 for I gives us all of J to migrate and further stabilize before making v2 "preferred" in J and deprecating v1 | 17:03 |
| mattf | how long v1 lives on after that is a mystery to me atm | 17:03 |
| mattf | migrate == clients using v2 | 17:03 |
| SergeyLukjanov | agreed w/ the v1 life time mystery | 17:03 |
| mattf | not important now, but two reasonable paths are 0) leave it until it becomes a burden (basically lives forever), 1) give it 2-3 releases and delete | 17:05 |
| openstackgerrit | Robert Nettleton proposed a change to openstack/savanna-image-elements: Adds Hortonworks Data Platform element to savanna-image-elements https://review.openstack.org/72994 | 17:05 |
| SergeyLukjanov | I'll return back later and write something else | 17:08 |
| openstackgerrit | Dmitry Mescheryakov proposed a change to openstack/savanna: Do eventlet.monkey_patch at the very beginning https://review.openstack.org/72992 | 17:09 |
| openstackgerrit | A change was merged to openstack/savanna: Fixed hadoop dir creation during hadoop-swift lib download https://review.openstack.org/72847 | 17:11 |
| *** mst89 has joined #savanna | 17:20 | |
| *** dmitryme has quit IRC | 17:24 | |
| *** akuznetsov has joined #savanna | 17:27 | |
| openstackgerrit | Matthew Farrellee proposed a change to openstack/savanna: Expand swift data source credential tests https://review.openstack.org/73010 | 17:40 |
| *** matsuhashi has quit IRC | 17:48 | |
| *** _nadya_ has joined #savanna | 17:48 | |
| *** saju_m has quit IRC | 17:50 | |
| *** saju_m has joined #savanna | 17:51 | |
| *** matsuhashi has joined #savanna | 17:51 | |
| *** matsuhashi has quit IRC | 17:55 | |
| *** saju_m has quit IRC | 17:58 | |
| *** saju_m has joined #savanna | 17:59 | |
| *** Ch00k has joined #savanna | 18:00 | |
| *** aignatov_ is now known as aignatov | 18:27 | |
| SergeyLukjanov | _tmckay, awesome work on job types, thank you! | 18:35 |
| *** aignatov is now known as aignatov_ | 18:36 | |
| *** _nadya_ has quit IRC | 18:37 | |
| *** saju_m has quit IRC | 18:51 | |
| *** _nadya_ has joined #savanna | 18:52 | |
| *** _tmckay is now known as tmckay | 18:54 | |
| tmckay | SergeyLukjanov, you're welcome, I'm glad you like it :) | 18:55 |
| tmckay | I made the mistake of going to the store in a snow storm.... | 18:55 |
| tmckay | SergeyLukjanov, I was thinking that I need to make another pass and define constants for the job type strings | 18:56 |
| tmckay | too much possibility for typos | 18:56 |
| SergeyLukjanov | tmckay, yup, good idea | 18:57 |
| *** Ch00k has quit IRC | 19:00 | |
| *** saju_m has joined #savanna | 19:00 | |
| *** Ch00k has joined #savanna | 19:01 | |
| *** Ch00k has quit IRC | 19:08 | |
| *** ErikB has quit IRC | 19:08 | |
| *** Ch00k has joined #savanna | 19:12 | |
| *** tosky has joined #savanna | 19:14 | |
| *** ErikB has joined #savanna | 19:18 | |
| *** IlyaE has quit IRC | 19:21 | |
| *** aignatov_ is now known as aignatov | 19:43 | |
| *** _nadya_ has quit IRC | 19:45 | |
| *** ErikB has quit IRC | 19:46 | |
| *** ErikB has joined #savanna | 19:46 | |
| *** tosky has quit IRC | 19:47 | |
| *** saju_m has quit IRC | 19:48 | |
| openstackgerrit | Chad Roberts proposed a change to openstack/savanna-dashboard: Treating blanked out config values as defaults for cluster templates https://review.openstack.org/73028 | 19:50 |
| openstackgerrit | A change was merged to openstack/savanna: Add support for dotted job types https://review.openstack.org/72686 | 19:56 |
| openstackgerrit | A change was merged to openstack/savanna: Remove all support for "Jar" as a job type (alias for "MapReduce") https://review.openstack.org/72733 | 19:59 |
| *** ErikB has quit IRC | 19:59 | |
| *** ErikB has joined #savanna | 20:11 | |
| *** mst89 has quit IRC | 20:24 | |
| *** Ch00k has quit IRC | 20:25 | |
| *** IlyaE has joined #savanna | 20:34 | |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack/savanna: Sort modules in openstack-common.conf https://review.openstack.org/73050 | 20:40 |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack/savanna: Fix how migration's cli register db connection opt https://review.openstack.org/73051 | 20:40 |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack/savanna: Standardize config sample locations https://review.openstack.org/73052 | 20:40 |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack/savanna: Auto generate and check config sample https://review.openstack.org/73053 | 20:40 |
| tmckay | mattf, re the NOSE_* environment variables in python-savannaclient/tox.ini Are those used by testr? I can't find any evidence of it. Do you know why they're there? | 20:46 |
| *** DinaBelova is now known as DinaBelova_ | 20:47 | |
| * tmckay continues to dive deep on this test stuff | 20:47 | |
| mattf | tmckay, i think they were removed recently | 20:48 |
| tmckay | k, maybe not merged (or I thought I pulled and didn't) | 20:49 |
| tmckay | thanks | 20:49 |
| mattf | on sunday in fact | 20:49 |
| mattf | lightning fast moving repo | 20:49 |
| tmckay | heh, helps to be on master | 20:49 |
| * mattf chuckles | 20:49 | |
| mattf | wouldn't hurt | 20:49 |
| *** ErikB has quit IRC | 20:53 | |
| *** boris-42_ has quit IRC | 20:55 | |
| *** boris-42_ has joined #savanna | 20:59 | |
| jmaron | mattf: hi. working on rootwrap. may not be able to specify rootwrap as default mechanism since that would force configuration of rootwrap facility for all users (including root): sudoers mods, creation of rootwrap filters dir, creation of filters file etc. Testing incomplete config, even for 'root', shows command execution will fail. So it appears that it'd be best to make it configurable… | 21:04 |
| *** mattf is now known as _mattf | 21:13 | |
| *** ErikB has joined #savanna | 21:17 | |
| *** _mattf is now known as mattf | 21:43 | |
| tmckay | SergeyLukjanov, still there? | 21:57 |
| SergeyLukjanov | tmckay, partially | 21:58 |
| SergeyLukjanov | tmckay, what's up? | 21:58 |
| tmckay | hi. I was looking back through irc posts on cli integration tests | 21:58 |
| tmckay | I have a stub to set up savannaclient for unit/integration. integration is empty... | 21:58 |
| tmckay | Can't find a blueprint for it, though. But I do see a reference to "tempest" | 21:59 |
| SergeyLukjanov | tmckay, I'm pretty sure that we don't need integration tests for cli now | 21:59 |
| tmckay | Is it okay to go ahead with a unit/integration stub for cli, and add a blueprint? | 21:59 |
| tmckay | okay, I know it's something that mattf had been meaning to look at | 22:00 |
| SergeyLukjanov | read-only integration tests for cli could be contributed to the tempest | 22:00 |
| tmckay | I thought I'd help out | 22:00 |
| SergeyLukjanov | aignatov said that he'd like to write some unit tests for client | 22:01 |
| SergeyLukjanov | so, you need to communicate with him, it was on last IRC meeting IIRC | 22:01 |
| tmckay | k. | 22:01 |
| SergeyLukjanov | tmckay, are you working on tests for python bindings or for cli? | 22:02 |
| tmckay | I was setting up python-savannaclient to have unit tests and integration tests, so that tox -e py27 and tox -e integration for example would operate like savanna-api. I thought that's what we wanted, but mabye I was mistaken | 22:04 |
| tmckay | integration tests is empty at the moment | 22:04 |
| SergeyLukjanov | we need py27 but there is no need for integration | 22:04 |
| SergeyLukjanov | we already have some integration tests for client in savanna/integration tests | 22:04 |
| SergeyLukjanov | because we're testing savanna using client | 22:04 |
| tmckay | I suppose that's true | 22:05 |
| SergeyLukjanov | IMO we need some unit tests to cover client's auth magic and etc/ | 22:05 |
| tmckay | okay, I can talk to aignatov about that | 22:05 |
| SergeyLukjanov | and unit tests to check that cli transforms cli args to python bindings calls correctly | 22:05 |
| SergeyLukjanov | tmckay, thanks | 22:06 |
| *** ErikB has quit IRC | 22:10 | |
| *** crobertsrh is now known as _crobertsrh | 22:31 | |
| openstackgerrit | Andrew Lazarev proposed a change to openstack/savanna: [IDH] Added ability to support several versions https://review.openstack.org/73087 | 22:33 |
| openstackgerrit | Andrew Lazarev proposed a change to openstack/savanna: [IDH] Added ability to support several versions https://review.openstack.org/73087 | 22:36 |
| openstackgerrit | Andrew Lazarev proposed a change to openstack/savanna: [IDH] Added ability to support several versions https://review.openstack.org/73087 | 22:44 |
| *** witlessb has quit IRC | 22:47 | |
| *** aignatov is now known as aignatov_ | 23:03 | |
| *** jmaron has quit IRC | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!