Tuesday, 2016-09-06

*** hoangcx has joined #openstack-sdks00:46
*** markvoelker has joined #openstack-sdks00:53
*** aohuanxuan has joined #openstack-sdks01:20
*** yanyanhu has joined #openstack-sdks01:23
*** yuanying has joined #openstack-sdks01:24
*** yuanying has quit IRC01:24
*** zhurong has joined #openstack-sdks01:25
openstackgerritRui Chen proposed openstack/python-openstackclient: Add functional tests for unset multiple volume type props  https://review.openstack.org/33533801:27
*** hoangcx has quit IRC01:32
*** hoangcx has joined #openstack-sdks02:06
*** salv-orl_ has joined #openstack-sdks02:29
*** salv-orlando has quit IRC02:32
*** RuiChen has joined #openstack-sdks02:33
*** annp has joined #openstack-sdks02:35
*** amotoki has joined #openstack-sdks02:48
*** amotoki has quit IRC02:50
*** yuanying has joined #openstack-sdks02:52
*** markvoelker has quit IRC02:53
*** shu-mutou-AFK is now known as shu-mutou02:55
*** yuanying has quit IRC02:55
*** yuanying has joined #openstack-sdks02:57
aohuanxuandtroyer: Hi, Dean, Does OSC core team still hold new feature merges now?02:59
*** yuanying has quit IRC03:00
*** yuanying has joined #openstack-sdks03:00
*** singhj has joined #openstack-sdks03:01
*** hoangcx has quit IRC03:01
*** singhj has quit IRC03:03
*** amotoki has joined #openstack-sdks03:07
*** yuanying has quit IRC03:18
*** amotoki has quit IRC03:18
*** yuanying has joined #openstack-sdks03:20
*** hoangcx has joined #openstack-sdks03:23
*** gildub has joined #openstack-sdks03:33
*** gildub_ has joined #openstack-sdks03:33
openstackgerritMerged openstack/python-openstackclient: Add "volume service set" command  https://review.openstack.org/35619003:40
*** amotoki has joined #openstack-sdks03:54
*** yuanying has quit IRC04:00
*** yuanying has joined #openstack-sdks04:02
*** shu-mutou is now known as shu-mutou-AFK04:12
*** adriant has quit IRC04:59
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move unit tests to new "unit" test module  https://review.openstack.org/36586305:16
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move all functional tests to tests module  https://review.openstack.org/36586405:18
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move unit tests to new "unit" test module  https://review.openstack.org/36586305:21
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: better functional test collection  https://review.openstack.org/33013705:21
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move all functional tests to tests module  https://review.openstack.org/36586405:21
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: remove the old post_test_hook  https://review.openstack.org/36586705:24
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move all functional tests to tests module  https://review.openstack.org/36586405:24
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: remove the old post_test_hook  https://review.openstack.org/36586705:25
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move unit tests to new "unit" test module  https://review.openstack.org/36586305:33
*** gildub has quit IRC05:34
*** gildub_ has quit IRC05:35
*** gildub has joined #openstack-sdks05:35
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move all functional tests to tests module  https://review.openstack.org/36586405:37
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: remove the old post_test_hook  https://review.openstack.org/36586705:37
*** e0ne has joined #openstack-sdks05:46
*** e0ne has quit IRC05:47
*** reedip has quit IRC05:50
*** reedip has joined #openstack-sdks06:03
openstackgerritxu-haiwei proposed openstack/python-openstacksdk: Add 'dependents' property to Node  https://review.openstack.org/36587706:11
*** gildub has quit IRC06:49
*** gildub has joined #openstack-sdks06:50
*** gildub has quit IRC06:51
*** gildub has joined #openstack-sdks06:52
*** markvoelker has joined #openstack-sdks06:54
*** markvoelker has quit IRC06:59
*** namnh has joined #openstack-sdks06:59
*** hoangcx has quit IRC07:01
*** gildub has quit IRC07:02
*** gildub has joined #openstack-sdks07:02
*** hoangcx has joined #openstack-sdks07:06
*** jpena|off is now known as jpena07:08
*** GrineBloude is now known as GreenBlood07:08
*** GreenBlood has quit IRC07:08
*** GreenBlood has joined #openstack-sdks07:08
*** fzdarsky has joined #openstack-sdks07:22
*** Nh3 has joined #openstack-sdks07:35
openstackgerritQiangTang proposed openstack/python-openstackclient: Refactor image project add / remove  https://review.openstack.org/36591007:35
*** gildub has quit IRC07:38
*** Nh3 has quit IRC08:12
*** lucas-dinner is now known as lucasagomes08:16
*** markvoelker has joined #openstack-sdks08:19
*** markvoelker has quit IRC08:24
*** Nh3 has joined #openstack-sdks08:26
*** salv-orlando has joined #openstack-sdks08:29
*** salv-orl_ has quit IRC08:32
*** hoangcx has quit IRC08:59
openstackgerritQiangTang proposed openstack/python-openstackclient: Refactor image project add / remove  https://review.openstack.org/36591009:23
openstackgerritQiangTang proposed openstack/python-openstackclient: Refactor image project add / remove  https://review.openstack.org/36591009:25
*** mfedosin has joined #openstack-sdks09:25
*** Nh3 has quit IRC09:38
*** cdent has joined #openstack-sdks09:39
*** Nh3 has joined #openstack-sdks09:40
*** reedip has quit IRC09:51
*** namnh has quit IRC09:54
*** annp has quit IRC09:57
*** zhurong has quit IRC09:58
*** yanyanhu has quit IRC10:01
*** reedip has joined #openstack-sdks10:04
openstackgerritRui Chen proposed openstack/python-openstackclient: Add functional tests for unset multiple volume type props  https://review.openstack.org/33533810:07
*** e0ne has joined #openstack-sdks10:15
*** Nh3 has quit IRC10:17
*** cdent has quit IRC10:42
*** HenryG_ is now known as HenryG10:46
*** cdent has joined #openstack-sdks11:05
*** clouddig has joined #openstack-sdks11:18
openstackgerritHuanxuan Ao proposed openstack/python-openstackclient: Multi REST API calls error handling of "volume unset" command  https://review.openstack.org/36605911:20
*** lucasagomes is now known as lucas-hungry11:22
*** rtheis has joined #openstack-sdks11:22
*** Nh3 has joined #openstack-sdks11:24
*** e0ne has quit IRC11:25
*** thrash|g0ne is now known as thrash11:28
*** amotoki_ has joined #openstack-sdks11:29
*** amotoki has quit IRC11:31
*** gildub has joined #openstack-sdks11:35
*** mfedosin has quit IRC11:37
*** sdague has joined #openstack-sdks11:39
*** jpena is now known as jpena|lunch11:48
*** mfedosin has joined #openstack-sdks11:51
*** salv-orl_ has joined #openstack-sdks11:52
*** salv-orlando has quit IRC11:52
*** reedip has quit IRC12:02
*** zhurong has joined #openstack-sdks12:10
*** Nh3 has quit IRC12:12
*** edmondsw has joined #openstack-sdks12:13
*** reedip has joined #openstack-sdks12:16
*** gildub has quit IRC12:18
*** aohuanxuan has quit IRC12:21
*** Nh3 has joined #openstack-sdks12:25
*** lucas-hungry is now known as lucasagomes12:27
*** figleaf is now known as edleafe12:45
*** jpena|lunch is now known as jpena12:48
openstackgerritMerged openstack/python-openstacksdk: Fix telemetry/metering service version  https://review.openstack.org/36547112:48
openstackgerritMerged openstack/python-openstacksdk: Rebase telemetry resources to resource2/proxy2  https://review.openstack.org/36551312:48
briancurtinQiming: fyi since i’ve seen you working on a bunch of rebases, i’m doing the object store one right now (just so we don’t step on each other)13:03
*** _elmiko is now known as elmiko13:04
Qimingokay, briancurtin13:04
*** e0ne has joined #openstack-sdks13:17
*** Nh3 has quit IRC13:18
*** jpena is now known as jpena|off13:19
*** dtantsur has joined #openstack-sdks13:19
dtantsurfolks, is it a know issue/feature in 3.2.0 that non-password (e.g. token) authentication no longer works?13:20
dtantsurin --debug I see "Deferring keystone exception: __init__() got an unexpected keyword argument 'token'", can it be some version mismatch on my side?13:21
reediprthies : ping13:22
*** shadower has joined #openstack-sdks13:23
dtantsurhmm, I guess it's because it still picks "password" plugin... the builtin help says it should be smarter when only --os-toien and --os-auth-url are provided..13:25
rtheisreedip: hi13:26
*** zhurong has quit IRC13:26
reediprtheis : hi , does openstackclient allow bulk_create ????13:26
rtheisreedip: I don't think so13:26
reediprtheis : neutron support bulk create for network and other properties13:26
reedipin the API13:26
rtheisokay, that may be something osc could consider supporting13:27
reediprthies: I was working on  https://bugs.launchpad.net/neutron/+bug/160466213:28
openstackLaunchpad bug 1604662 in neutron "Bulk creation for security group returns 500 error." [Medium,In progress] - Assigned to Reedip (reedip-banerjee)13:28
reediprtheis: so thought if API does support it, why should the OSC as well ? Also , I confirmed from amotoki, python_bindings support this feature in NeutronClient13:29
amotoki_reedip: From CLI perspective, it is not easy to use bulk creation.13:30
amotoki_reedip: I am not sure we need to support bulk operation in OSC. SDK support makes sense.13:30
reedipamotoki_ , rtheis: I know OSC ties up with SDK ( OSC acts as an input to the SDK) . Is there any other input to the SDK as well in the form of a Client Request ?13:31
reedipamotoki_ , rthies: so SDK would act the same for API requests as well  ??13:32
amotoki_reedip: you can use SDK from usual python programs.13:32
*** Nh3 has joined #openstack-sdks13:32
reedipamotoki_ Hmm, so API and other scripting can be used with SDK13:32
amotoki_reedip: of course13:32
briancurtinopenstacksdk is a library. you can do anything you want with it13:32
reedipamotoki_ Ok, sounds good . rtheis : I will put up a BP for this ???13:32
amotoki_reedip: bp for which?13:33
reedipbriancurtin : okay :)13:33
reedipamotoki_ does the Openstack SDK support Bulk Create for network ?13:34
briancurtinor, probably not13:34
reedipamotoki_ I couldnt find it in the code13:34
reedipbriancurtin : I cannot find it in the network code.13:35
briancurtinit could if people think it’s reasonable to do. we haven’t implemented 100% of all available calls13:35
reedipbriancurtin : should this be put up as an agenda for the nest weekly meeting ?13:35
*** jpena|off is now known as jpena13:36
*** markvoelker has joined #openstack-sdks13:36
briancurtinwe haven’t been having the meetings, and if you want it you should just implement it. everyone already has more than enough work to do, so talking about something isn’t going to make it be completed13:36
amotoki_reedip: I don't know it. perhaps no13:36
*** john-davidge has joined #openstack-sdks13:37
amotoki_I think bulk support in SDK needs to be discussed carefully.13:37
amotoki_for example, what kind of interface should be provided?13:37
reedipbriancurtin, amotoki_ , rthies : okay, then I will start to work on this. it can be discussed further on launchpad13:38
*** cleong has joined #openstack-sdks13:45
openstackgerritMerged openstack/python-openstackclient: Use assertIn(A, B) instead of assertTrue(A in B )  https://review.openstack.org/36463513:53
*** annegentle has joined #openstack-sdks13:58
*** Nh3 has quit IRC13:59
openstackgerritMerged openstack/python-openstackclient: better functional test collection  https://review.openstack.org/33013714:01
openstackgerritMerged openstack/python-openstackclient: Clean up FakeClient classes in volume fakes  https://review.openstack.org/35815414:03
*** amotoki_ has quit IRC14:09
*** Nh3 has joined #openstack-sdks14:10
openstackgerritMerged openstack/js-openstack-lib: Use babel-plugin-transform-runtime instead of babel-polyfill  https://review.openstack.org/36572814:15
stevemarreedip: how would bulk create work?14:16
stevemarreedip: wouldn't you want different properties for each resource you are creating?14:16
*** amotoki has joined #openstack-sdks14:17
reedipstevemar : bulk create would require similar properties to be mentioned together14:17
reedipstevemar : what I mean is, if a resource has the following properties14:17
reedipstevemar:          Resource => A. Necessary properties => a1, a2   . Optional Properties => b1, b214:18
reedipstevemar: user can create multiple resources of type A, specifying the common attributes a1 and a2 for each resource14:19
reedipstevemar: user can update the optional properties b1 and b2 for the individual resource copy14:19
stevemarreedip: so like `openstack project create --description projectAdescription, projectBdescription projectA projectB`14:20
reedipstevemar: it would be difficult to implement it on the CLI layer, as  mentioned by amotoki earlier. But it would be easier to implement the same on the SDK layer14:21
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move unit tests to new "unit" test module  https://review.openstack.org/36586314:21
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move all functional tests to tests module  https://review.openstack.org/36586414:21
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: remove the old post_test_hook  https://review.openstack.org/36586714:21
stevemarreedip: ah okay14:22
reedipstevemar : there is also a point which I wanted to discuss about duplicated parameters14:22
reedipstevemar: https://review.openstack.org/#/c/230321/   This was something which I proposed in NeutronClient but I feel it would be better to implement it on OSC. However rtheis suggested cliff would be a better option.14:24
reedipstevemar: what is your opinion?14:25
*** zigo has quit IRC14:25
*** salv-orlando has joined #openstack-sdks14:29
*** Nh3 has quit IRC14:32
*** zigo has joined #openstack-sdks14:32
*** salv-orl_ has quit IRC14:32
*** zigo is now known as Guest3914714:33
*** reedip is now known as reedip_outofoffi14:34
*** reedip_outofoffi is now known as reedip14:34
*** cdent has quit IRC14:37
*** amotoki has quit IRC14:38
*** fguillot has joined #openstack-sdks14:38
*** Guest39147 has quit IRC14:39
*** Nh3 has joined #openstack-sdks14:40
*** zigo_ has joined #openstack-sdks14:42
*** john-davidge has quit IRC14:45
*** zigo_ has quit IRC14:47
*** zigo_ has joined #openstack-sdks14:48
stevemarreedip: oh no question :)14:49
stevemarreedip: was just confused about how we were going to do that from the CLI14:49
*** annegentle has quit IRC14:53
openstackgerritTom Jose Kalapura proposed openstack/python-openstackclient: Add sort support to project list  https://review.openstack.org/33520414:54
stevemardtroyer: o/14:57
dtroyerhey steve-o14:57
stevemardtroyer: i've got an interesting patch for ya https://review.openstack.org/#/c/365863/14:58
dtroyerso I see why integ moved into unit, is there a path to moving them back out at some point?  or is that distinction not terribly useful?15:01
stevemardtroyer: unfortunately, i couldn't figure out a way to split the tests into 3: unit/integration/functional15:01
stevemardtroyer: i'm working within the limitations of testr :(15:01
stevemarno plan to pull them back out15:01
stevemarthe hiccup is that the functional tests are run by setting OS_TEST_PATH to osc.tests.functional15:02
dtroyeris there any reason to maintain the integ directory then?15:02
stevemarall tests run with py27 and py35 are run with testr.conf, and that sets OS_TEST_PATH to just osc.tests.unit, i didn't see an option for omitting a directory intead15:03
stevemardtroyer: i like the distinction *shrugs*15:03
*** jpena is now known as jpena|away15:03
dtroyerok, I have no issues with keeping it, just wasn't sure if it still served a purpose other than mental organization15:03
dtroyerwhich I suppose is enough actually15:04
stevemardtroyer: yeah :)15:04
stevemardtroyer: also, are we opening the flood gates for features and bug fixes?15:05
stevemarjust gonna backport things to stable/newton15:05
dtroyerI didn't think so, but a feature merged anyway15:05
dtroyerwhat backport?15:05
stevemaryeah, i saw that, wasn't sure if i should propose a revert of that one15:05
stevemardtroyer: i assumed https://review.openstack.org/#/c/364518/ would be backported to stable/newton15:06
stevemardtroyer: i can also wait to talk about this all on thursday if you want your morning back :)15:07
dtroyermeetings start soon…15:07
dtroyerI didn't see your comments on 36451815:07
stevemari'm hoping i did something wrong15:08
*** annegentle has joined #openstack-sdks15:08
dtroyerI'm not going to be able to put that all back into my head today, I hope so too ;)15:08
dtroyeryeah, we'll probably backport that one once we fix it, its big enough15:09
*** gouthamr has joined #openstack-sdks15:10
*** amotoki has joined #openstack-sdks15:11
*** gouthamr_ has joined #openstack-sdks15:12
*** gouthamr has quit IRC15:15
stevemardtroyer: time to re-evaluate our stance on stable releases me thinks15:15
*** thrash is now known as thrash|biab15:18
*** zigo_ is now known as zigo15:20
*** cdent has joined #openstack-sdks15:26
*** dtantsur has left #openstack-sdks15:29
*** sbelous has joined #openstack-sdks15:30
*** tangchen has quit IRC15:33
openstackgerritRodolfo Alonso Hernandez proposed openstack/python-openstacksdk: Add QoS minimum bandwidth rule object and CRUD commands.  https://review.openstack.org/35156515:37
openstackgerritRichard Theis proposed openstack/python-openstacksdk: Add network segment create, delete and update support  https://review.openstack.org/30222315:44
*** thrash|biab is now known as thrash15:44
openstackgerritCorentin Ardeois proposed openstack/js-openstack-lib: Add more service abstraction  https://review.openstack.org/36622215:49
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move unit tests to new "unit" test module  https://review.openstack.org/36586315:50
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: move all functional tests to tests module  https://review.openstack.org/36586415:50
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: remove the old post_test_hook  https://review.openstack.org/36586715:50
openstackgerritRodolfo Alonso Hernandez proposed openstack/python-openstacksdk: Add QoS DSCP marking rule object and CRUD commands.  https://review.openstack.org/35278915:56
*** Nh3 has quit IRC15:57
*** e0ne has quit IRC16:20
*** gouthamr_ has quit IRC16:20
*** stevemar has quit IRC16:31
*** gouthamr has joined #openstack-sdks16:33
openstackgerritRichard Theis proposed openstack/python-openstackclient: Add network segment create, delete and set support  https://review.openstack.org/30239516:42
*** lucasagomes is now known as lucas-dinner16:44
*** hogepodge has quit IRC16:58
openstackgerritRodolfo Alonso Hernandez proposed openstack/python-openstacksdk: Add QoS minimum bandwidth rule object and CRUD commands.  https://review.openstack.org/35156517:00
*** fzdarsky is now known as fzdarsky|afk17:01
rtheis"sudo pip install -e ." no longer installs my OSC changes.  Did something change recently to prevent this from working?17:02
rtheis"sudo easy_install -U ." seems to work now17:03
*** hogepodge has joined #openstack-sdks17:03
*** bradjones has quit IRC17:06
*** bradjones has joined #openstack-sdks17:11
*** bradjones has joined #openstack-sdks17:11
openstackgerritCorentin Ardeois proposed openstack/js-openstack-lib: Add more service abstraction  https://review.openstack.org/36622217:22
*** kevo has joined #openstack-sdks17:23
*** stevemar has joined #openstack-sdks17:26
*** e0ne has joined #openstack-sdks17:28
stevemarrtheis: o/17:30
stevemarrtheis: did this work for you? https://review.openstack.org/#/c/364518/117:30
*** e0ne has quit IRC17:36
*** clouddig has quit IRC17:37
*** amotoki has quit IRC17:39
*** cdent has quit IRC17:42
*** tomjosekal has joined #openstack-sdks17:49
rtheisstevemar: hi, I didn't pull it down for a test, but will try now17:50
*** fangxu has joined #openstack-sdks17:53
*** fangxu has quit IRC17:54
rtheisstevemar: I get "Missing value auth-url required for auth plugin password" when running "openstack complete"17:56
stevemarrtheis: okay cool, thats what i was getting cc dtroyer17:56
*** clouddig has joined #openstack-sdks17:56
*** e0ne has joined #openstack-sdks17:58
rtheisstevemar: I had not auth info set so based on dtroyer's commit message, I think my result is as expected17:59
rtheis*no auth info18:00
*** bknudson has joined #openstack-sdks18:00
*** crinkle_ is now known as crinkle18:07
*** Anticimex has joined #openstack-sdks18:28
Anticimexquestion regarding openstackclient18:28
Anticimexis it within scope of the "openstack server create" to add the ability for it to auto-request-and-assign a floating ip while creating an instance?18:29
Anticimexin our openstack deployment, private networks are private, and users need floating ips to reach general internet (ie identical behaviour as AWS VPC)18:30
stevemarAnticimex: that would be awesome!18:39
*** mfedosin has quit IRC18:40
Anticimexstevemar: glad you agree :)18:43
Anticimexbased on this tremendous feedback sample size of 1 i may just go and create a bug for it then18:43
Anticimexor well, what's the proper procedure of adding a feature/function like this?18:44
AnticimexBP always required?18:44
stevemarAnticimex: oh a bug is fine18:48
stevemarAnticimex: so like... `os server create --flavor m1.small --image something --get-me-a-damn-ip-and-use-it` :)18:50
Anticimexi suppose, i'm not sure of other deployments variants18:50
Anticimexi'm happy to poc the one we use :)18:50
Anticimexwhere virtual networks by default are private18:50
Anticimexand floating ips are how you reach internet18:50
Anticimexit is literally identical to aws vpc18:50
Anticimex(i just tested 30 min ago :P )18:50
dtroyerThere are a good number of deployment considerations and edge cases to consider in regard to doing this… we will have to gracefully handle them...18:52
Anticimexi recall monty talked about "getmeanetwork" uh, vancouver perhaps18:52
Anticimexor tokyo18:52
Anticimexor makemeanetwork?18:52
Anticimexdtroyer: yes18:52
Anticimexwhich is why i'm a bit wary of the AI-like function18:52
rtheisI think we want to leverage that support18:53
Anticimexthe closest thing to my ask is "--auto-assign-public-address"18:53
Anticimexor somesuch18:53
dtroyerpart of the issue is the difficulty in discovering all of the conditions present in any given cloud's network18:53
Anticimexerr, "--auto-assign-floating-ip", for more correct terminology18:53
Anticimexdtroyer: indeed18:54
Anticimex(what I called AI)18:54
Anticimexrtheis: did/has that entered openstackclient?18:54
Anticimexi'm not on top of the commitlog :)18:55
rtheisI don't think it has18:55
rtheisdtroyer is right which is why I think the CLI should let neutron and nova figure it out via get-me-a-network support18:55
Anticimexfair enough, but you have to start somewhere18:58
Anticimexrtheis: do you know where one can see state of "get-me-a-network"?18:59
Anticimexa sort of cop-out way of doing it is for deployers to actually fill in the blanks for such a function18:59
Anticimexthough i guess one of montys driving forces were that it should work everywhere?19:00
rtheisanticimex: you could ask henryg and mriedem19:00
Anticimexok, thanks19:00
Anticimexlove complex cross project feature adds ;)19:01
* stevemar thinks there should be a silent alias option for "getmeanetwork"19:01
stevemarif we have a real command line option for it19:01
stevemarlike an easter egg for folks reading the source code19:01
stevemarrtheis: so you're syaing that in https://review.openstack.org/#/c/364518/ i'm getting the expected output?19:04
Anticimexi'm wondering if it's appropriate to go to mailing list about the auto-assign-floating-ip thing, in relation to getmeanetwork19:05
rtheisstevemar: I think so, but will defer to dtroyer19:05
stevemarrtheis: but `complete` used to work without any credentials19:06
rtheisah, I wasn't sure about the before case19:07
*** fzdarsky|afk is now known as fzdarsky19:08
*** e0ne has quit IRC19:08
stevemarrtheis: https://github.com/openstack-dev/devstack/commit/6930ba312f787e0459f7455ac6ba2a70b3c7c37d19:09
stevemarrtheis: https://review.openstack.org/#/c/366296/19:10
stevemaradded a depends-on, curious to see if it'll take the osc proposed version19:10
*** fangxu has joined #openstack-sdks19:12
*** cdent has joined #openstack-sdks19:14
*** cleong has quit IRC19:15
*** cdent has quit IRC19:17
*** fangxu has quit IRC19:20
*** mliima has quit IRC19:38
*** sindhude__ has joined #openstack-sdks19:59
*** fangxu has joined #openstack-sdks20:02
*** edmondsw has quit IRC20:13
*** jpena|away is now known as jpena|off20:15
stevemarrtheis: well damn, devstack worked20:22
rtheisthat is interesting20:23
*** salv-orl_ has joined #openstack-sdks20:30
*** salv-orlando has quit IRC20:33
stevemardtroyer: done your meetings? :)20:38
*** fzdarsky has quit IRC20:53
*** craigs has joined #openstack-sdks20:54
dtroyerstevemar: am now, need a short break21:01
stevemardtroyer: ++21:01
*** sindhude__ has quit IRC21:06
*** sindhude__ has joined #openstack-sdks21:09
*** clouddig has quit IRC21:13
stevemarsindhude__: hey there21:19
sindhude__stevemar: hey21:19
stevemarsindhude__: so rtheis can probably help you out 100x faster than i can when it comes to networking commands in OSC21:19
stevemarrtheis: what is the deal with https://bugs.launchpad.net/python-openstackclient/+bug/1560297 ?21:19
openstackLaunchpad bug 1560297 in python-openstackclient "Add "floating ip associate/disassociate"" [Wishlist,Confirmed] - Assigned to Sindhu Devale (sindhu-devale-3)21:19
stevemarrtheis: sindhude__ wants to tackle the bug, what should be do?21:19
stevemarwhat should the command look like?21:20
sindhude__stevemar: I spoke to him21:20
stevemarsindhude__: damnnn21:21
stevemarsindhude__: still unsure of where to start?21:21
stevemaryou've pushed patches to gerrit before, i assume?21:21
sindhude__stevemar: yes I have21:21
stevemaryes, several to neutron ;)21:22
stevemarsindhude__: i suggest adding comments to the bug, ask tangchen and rtheis what they are looking for21:22
stevemarsindhude__: you probably just need to know what the command should look like, then  you can start implementing it21:23
stevemarsindhude__: whats the neutron command look like?21:23
sindhude__stevemar: ok... will do. rtheis said to propose the spec of the command before implementing it21:24
stevemarsindhude__: i think he just wants to see what the command will look like21:24
sindhude__stevemar, : all I am looking at is this etherpad: https://etherpad.openstack.org/p/osc-neutron-support21:25
rtheissindhude__: right, then we can save some time on implementation21:25
rtheisI'm not quite sure the right direction for this since I haven't looked into it much21:25
rtheistangchen is more familiar and may know what to do21:26
rtheisso you can go right to the implementation21:26
rtheisyou could check with him21:26
sindhude__rtheis, : ok ... will contact him. Thanks21:26
sindhude__thanks stevemar :)21:27
dtroyerstevemar: hey21:28
stevemarsindhude__: i'd look at https://github.com/openstack/python-neutronclient/blob/98b443adac9128d2836ef72719344a60fde7d3e9/neutronclient/neutron/v2_0/floatingip.py for staters21:32
stevemarsindhude__: but leave comments in the bug report if you want better feedback :)21:32
*** rtheis has quit IRC21:33
sindhude__stevemar: great.. that link helps. Will do. Thank you :)21:33
*** annegentle has quit IRC21:39
mordredjamielennox|away: https://review.openstack.org/#/c/352893/ btw - I fixed your TODO in the comment ...21:42
mordredAnticimex: --auto-assign-public-address is actually more correct21:43
mordredAnticimex: the thing it needs to do is 'whatever this cloud needs to have this vm be able to talk to things external to this cloud'21:43
*** adriant has joined #openstack-sdks21:44
mordredAnticimex: on some clouds, that's nothing. on some it's a floating ip21:44
mordredAnticimex: and then the actual task of getting a floating ip consistently is exceptionally hard21:44
mordredAnticimex: we have this implemented in shade, but it's a MASSIVE pile of code and very few people understand it end to end21:44
mordredAnticimex: that said, I do continue to think that it would be great if such a massive pile of code was not required21:45
mordredAnticimex: if you're curious - start here: https://github.com/openstack-infra/shade/blob/master/shade/openstackcloud.py#L4290-L4294 and follow the rabbit hole all the way down21:48
mordredAnticimex: but in any case - I can suggest a few things that can be implemented in neutron that would vastly simplify the whole endeavor21:48
mordredAnticimex: a) api call to figure out if the deployers of teh cloud have decided to force their users to use floating ips b) an attribute somewhere that communicates "this network routes packets externally" c) add neutron port id to the nova addresses dict21:50
mordredAnticimex: b is more difficult because the answer to that question is actually a quality that an individual subnet in a network has (each subnet should have that flag) but servers get attached to Networks and routers also associate with networks - so the data model mismatch would need to be addressed as well21:51
*** stevemar has quit IRC21:52
mordred(this gets really fun if the cloud has a dual-stack network with a globally routable IPv6 subnet and an RFC1918 IPv4 subnet)21:52
*** annegentle has joined #openstack-sdks22:02
*** csterret_ has joined #openstack-sdks22:03
*** craigs has quit IRC22:04
*** csterret_ has quit IRC22:06
*** sindhude__ has quit IRC22:17
*** stevemar has joined #openstack-sdks22:24
*** stevemar has quit IRC22:26
*** stevemar has joined #openstack-sdks22:26
*** stevemar has quit IRC22:31
*** sindhude__ has joined #openstack-sdks22:41
*** annegentle has quit IRC22:46
*** sindhude__ has quit IRC22:48
Anticimexmordred: awesome feedback, thanks23:05
*** sindhude__ has joined #openstack-sdks23:07
*** stevemar has joined #openstack-sdks23:07
Anticimexmordred: b is becomes tricky. telco etc would probably want to add on routing policy and what not :s23:08
Anticimexmordred: a) i realized myself, or well, a more ocmplex version of it (leaving the operator to configure the puzzle themselves, but this is probably too complicated)23:09
Anticimexmordred: but full agreement on the vast amount of simplification if the code via API can query the operators intention23:09
Anticimexmordred: c) i'm not familiar with really, didn't even know nova knew of addresses :]23:10
Anticimexmordred: globally routable ipv6 + rfc1918 ipv4 is one of our deployment models, due to vast amount of images having versions of cloud-infra (i.e. all version in existance, i guess) that only speak ipv423:10
Anticimexso you *need* to have rfc1918 to support cloud-infra23:11
Anticimexerr, cloud-init23:11
Anticimexmordred: w.r.t to your b) , as a networking guy background i'd say that it is precisely routers that may or may not have routes towards "external" networks23:12
Anticimexmordred: so there is already an external flag, on networks, which basically states if the network is on the "outside" or "inside" of some sort of wall23:12
Anticimexmordred: i guess many use it to distinguish rfc1918 vs globally unique addresses23:13
Anticimexmordred: but when it comes to forwarding information, that's routers23:13
Anticimexbut routing policy may be arbitrary23:13
Anticimexso operator should decide the flag i guess23:15
Anticimexbut i don't like b) :P23:15
mordredAnticimex: ah - but you see - the routing flag on network only says whether or not you can attach a neutron router23:18
mordredAnticimex: and that is a construct you can use to get floating ips23:18
mordredbut - for clouds that provide externally routable ipv4 via directly routed (not NAT) ips, router:external is not relevant or something you can cout on23:19
*** markvoelker has quit IRC23:19
mordredthere currently exists no cloud that says "the ip you get assigned for fixed addresses from this network will route northbound, but it's not a NATted network"23:19
mordredAnticimex: also, for dual-stack, we just landed support for it to shade last week because the OSIC cloud that infra nodepool uses is deployed in that model23:20
*** fangxu has quit IRC23:20
*** gildub has joined #openstack-sdks23:24
*** bapalm has quit IRC23:35
*** sindhude__ has quit IRC23:38
*** fangxu has joined #openstack-sdks23:38
*** bapalm has joined #openstack-sdks23:41
fangxuhi guys, wondering if https://github.com/openstack/golang-client is still under maintenance ?23:41
fangxuseems like no code reviewer……23:41
*** sindhude__ has joined #openstack-sdks23:53
*** markvoelker has joined #openstack-sdks23:57
dtroyerfangxu: that would be pretty much just me…sorry, Tuesdays are generally messy for me23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!