Tuesday, 2017-01-31

« Monday, 2017-01-30 Index Wednesday, 2017-02-01 »
openstackgerritHongbin Lu proposed openstack/kuryr-libnetwork: Add prefix to the specified name of subnetpool  https://review.openstack.org/42662300:05
openstackgerritHongbin Lu proposed openstack/kuryr-libnetwork: [WIP] Support creating from existing subnetpool  https://review.openstack.org/42659500:05
*** hongbin has quit IRC00:19
*** saneax is now known as saneax-_-|AFK00:34
*** mattmceuen has quit IRC00:45
*** tonanhngo has quit IRC01:31
*** dims has quit IRC01:35
*** dims_ has joined #openstack-kuryr01:35
*** dougbtv has quit IRC01:36
*** tonanhngo has joined #openstack-kuryr01:40
*** dims_ has quit IRC01:42
*** dims has joined #openstack-kuryr01:42
*** tonanhngo has quit IRC01:44
*** tonanhngo has joined #openstack-kuryr01:46
*** tonanhngo has quit IRC01:50
*** tonanhngo has joined #openstack-kuryr01:52
*** tonanhngo has quit IRC01:56
*** hongbin has joined #openstack-kuryr02:55
*** hongbin has quit IRC04:08
*** yedongcan1 has joined #openstack-kuryr05:52
*** saneax-_-|AFK is now known as saneax06:36
*** janki has joined #openstack-kuryr06:49
*** yedongcan1 has quit IRC07:02
*** gsagie has quit IRC07:02
*** yedongcan has joined #openstack-kuryr07:03
*** yamamoto has quit IRC07:34
*** jchhatbar has joined #openstack-kuryr07:54
*** janki has quit IRC07:57
*** yamamoto has joined #openstack-kuryr08:10
irenabapuimedo, hi08:14
*** yamamoto has quit IRC08:37
apuimedoirenab: hi08:45
*** yamamoto has joined #openstack-kuryr08:59
*** garyloug has joined #openstack-kuryr09:00
*** devvesa has joined #openstack-kuryr09:02
*** devvesa has quit IRC09:02
*** yamamoto has quit IRC09:06
*** yamamoto has joined #openstack-kuryr09:09
*** yamamoto has quit IRC09:27
*** yedongcan1 has joined #openstack-kuryr09:36
*** yedongcan has quit IRC09:38
*** yamamoto has joined #openstack-kuryr09:39
janonymousapuimedo:Hi, Please find the fullstack cases log, http://paste.openstack.org/show/596995/  it works on 16.04 xenial09:43
janonymousapuimdeo: Apart from i found one socket permission error which is trivial i guess: http://paste.openstack.org/show/596990/ Line:12709:44
*** yamamoto has quit IRC09:45
janonymousapuimedo: i will share the client python brief shortly,in a day or two09:45
irenabapuimedo, the df kueye k8s gate failure will be resolved once https://review.openstack.org/#/c/426796/ is merged and osc upper constraints are updated10:04
*** jchhatbar is now known as janki10:08
*** yedongcan1 has left #openstack-kuryr10:23
*** neiljerram has joined #openstack-kuryr10:31
*** yamamoto has joined #openstack-kuryr10:42
*** ltomasbo has quit IRC11:00
apuimedogood! Thanks irenab11:06
*** ltomasbo has joined #openstack-kuryr11:07
*** yamamoto has quit IRC11:08
*** gsagie has joined #openstack-kuryr11:43
*** yamamoto has joined #openstack-kuryr11:45
*** yamamoto has quit IRC11:45
*** yamamoto has joined #openstack-kuryr11:53
*** janki has quit IRC11:58
*** garyloug has quit IRC11:59
*** yamamoto has quit IRC12:00
*** garyloug has joined #openstack-kuryr12:08
*** yamamoto has joined #openstack-kuryr12:11
*** yamamoto has quit IRC12:15
*** yamamoto has joined #openstack-kuryr12:42
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr: Make segmentation driver testable  https://review.openstack.org/42719012:58
apuimedo\o/12:59
apuimedo:-)12:59
ltomasbo:)13:02
apuimedoltomasbo: pep8 doesn't like it13:04
ltomasboI will check, I just fix it to include the unit test for vlan seg_driver and forgot to check the pep813:04
apuimedoltomasbo: 2017-01-31 13:01:25.254845 | ./kuryr/lib/segmentation_type_drivers/__init__.py:29:9: F841 local variable 'driver' is assigned to but never used13:05
ltomasboumm, yep, I'll fix it in a sec13:05
*** garyloug has quit IRC13:07
apuimedo;-)13:08
openstackgerritMerged openstack/kuryr-kubernetes: Skip stale K8s events  https://review.openstack.org/42291013:11
*** yamamoto has quit IRC13:24
*** garyloug has joined #openstack-kuryr13:24
openstackgerritMerged openstack/kuryr-kubernetes: Remove link to modindex  https://review.openstack.org/42559713:26
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr: Make segmentation driver testable  https://review.openstack.org/42719013:29
*** saneax is now known as saneax-_-|AFK13:30
apuimedoltomasbo: posted comments on https://review.openstack.org/#/c/427190/213:35
ltomasboapuimedo, I used it in L38 and L4913:37
apuimedoltomasbo: can't see it13:38
apuimedoI only see in 36 and 47 you defining a method local variable of hte same name13:39
ltomasboohh, too much multiplexing, you are right13:42
ltomasbowill take another look after lunch13:43
leifmadseno/13:44
leifmadsenapuimedo: ltomasbo: do either of you have any documentation around getting a dev environment up and running for testing / playing with Kuryr?13:45
apuimedoleifmadsen: we mostly just use devstack :-)13:45
leifmadsenif not, and you had some raw notes, I'd be happy to work through it and write up some documentation for the docs site13:45
leifmadsenapuimedo: I'd like to see if I could then maybe get it working with oooq13:45
apuimedooooq?13:46
ltomasboif you want kuryr + kubernetes, I wrote a step-by-step post for devstack13:46
leifmadsenTripleO Q13:46
leifmadsenQuickstart*13:46
ltomasboleifmadsen, https://ltomasbo.wordpress.com/2017/01/29/side-by-side-and-nested-kubernetes-and-openstack-deployment-with-kuryr/13:46
leifmadsenltomasbo: ok great! that might give me enough info to be dangerous13:47
leifmadsenI've never used devstack heh13:47
leifmadsenso I'll probably try and integrate with oooq13:47
leifmadsenltomasbo: I think I'm just going to add your blog to my blog notification feed :)13:47
apuimedoleifmadsen: doesn't tripleo require packages and puppet integration (which we do not have atm)13:47
leifmadsensome good stuff here13:47
leifmadsenapuimedo: TripleO does yea, but I can build an ansible playbook to run in a post-deploy13:47
apuimedoltomasbo: and you go have lunch. Blood sugar levels are important for patch generation!13:48
ltomasboleifmadsen, :D thanks!13:48
ltomasbo:D13:48
*** dougbtv has joined #openstack-kuryr14:06
*** saneax-_-|AFK is now known as saneax14:19
openstackgerritMerged openstack/kuryr: Updated from global requirements  https://review.openstack.org/42366914:22
*** yamamoto has joined #openstack-kuryr14:25
*** yamamoto has quit IRC14:32
*** v1k0d3n has quit IRC14:49
*** hongbin has joined #openstack-kuryr14:49
*** v1k0d3n has joined #openstack-kuryr14:50
*** saneax is now known as saneax-_-|AFK14:52
*** dougbtv has quit IRC15:17
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr: Add randomness to the returned vlan_ids  https://review.openstack.org/42264115:41
openstackgerritLuis Tomas Bolivar proposed openstack/kuryr: Make segmentation driver testable  https://review.openstack.org/42719015:44
*** gsagie has quit IRC15:46
ltomasboapuimedo, the randomness depends on the second one15:55
ltomasbonot the code, but the unit test15:55
ltomasboand the second one know is happy with the pep815:56
ltomasbos/know/now15:56
apuimedookey dokey15:57
apuimedothat sounds better15:57
apuimedo:-)15:57
*** tonanhngo has joined #openstack-kuryr16:38
*** tonanhngo_ has joined #openstack-kuryr16:40
*** tonanhngo has quit IRC16:43
*** tonanhngo_ has quit IRC16:45
*** tonanhngo has joined #openstack-kuryr16:47
*** tonanhngo has quit IRC16:51
garylougHi all, for the 'try nested pods locally' section of kuryr-k8s readme file, can anyone recommend an OS and version of OpenStack that this has been tested with?16:58
ltomasbohi garyloug17:08
ltomasboI've tried that with a fedora24 and master branch OpenStack17:09
ltomasboalso, for the VM to run the nested pods17:09
ltomasboI was also using fedora2417:09
ltomasborecomendation for the VM is to have at least 4GB and 2 vcpus17:10
garylougthanks very much ltomasbo17:43
ltomasboyour welcome garyloug!17:43
*** tonanhngo has joined #openstack-kuryr18:03
*** dougbtv has joined #openstack-kuryr18:41
*** pcaruana has quit IRC19:31
*** garyloug has quit IRC19:38
apuimedohongbin: ping19:42
hongbinapuimedo: hey20:06
openstackgerritIlya Chukhnakov proposed openstack/kuryr-kubernetes: K8s Services support: LBaaSSpecHandler  https://review.openstack.org/42744020:46
apuimedo:-)21:20
apuimedoivc_: http://logs.openstack.org/40/427440/1/check/gate-kuryr-kubernetes-python35/cc8c52d/console.html21:23
*** yamamoto has joined #openstack-kuryr21:23
apuimedothe mismatch looks identical to me21:23
apuimedoand it only happens on py3521:23
apuimedois it unicode vs bytes?21:23
ivc_apuimedo don't think its unicode/byte related21:26
* apuimedo installing py35 (arch recently moved to 36) to try out21:29
apuimedoivc_: (py35) celebdor@s21 ~/code/github.com/openstack/kuryr-kubernetes (23e1860*) $ python -m testtools.run kuryr_kubernetes.tests.unit.controller.handlers.test_lbaas.TestLBaaSSpecHandler.test_generate_lbaas_port_specs21:30
apuimedoTests running...21:30
apuimedoRan 1 test in 0.004s21:30
apuimedoOK21:30
apuimedoIt passes on my machine ;/21:30
ivc_now thats funny21:30
ivc_it should work in fact21:31
*** yamamoto has quit IRC21:33
ivc_apuimedo that py35 env was a completely fresh one, right? i.e. all requirements installed from scratch?21:36
* ivc_ still waiting on tox -e py3521:36
apuimedoivc_: I did cheat by using a fresh one with "pip install ." instead of using the global reqs21:36
apuimedoso that's the only thing21:37
apuimedolet me check21:37
apuimedoivc_: just clean installed it with 'tox -e py35' and then aborted before running all the tests21:41
apuimedoand run only the failing one21:41
apuimedocelebdor@s21 ~/code/github.com/openstack/kuryr-kubernetes (23e1860*) $ source .tox/py35/bin/activate21:41
apuimedo(py35) celebdor@s21 ~/code/github.com/openstack/kuryr-kubernetes (23e1860*) $ python -m testtools.run kuryr_kubernetes.tests.unit.controller.handlers.test_lbaas.TestLBaaSSpecHandler.test_generate_lbaas_port_specs21:41
apuimedoTests running...21:41
apuimedoRan 1 test in 0.004s21:41
apuimedoOK21:41
apuimedo(py35) celebdor@s21 ~/code/github.com/openstack/kuryr-kubernetes (23e1860*) $21:41
apuimedoI hate it when I cna't reproduce21:41
ivc_can u run tox -e py35 as is? mine is still building :/21:43
apuimedook21:43
*** tonanhngo has quit IRC21:44
apuimedoivc_: are you running it on http://i.imgur.com/sjUBumK.png ? :P21:44
ivc_nop, but i did not have py3.5-dev packages and had to rerun it21:45
apuimedoaha21:46
apuimedoI had to build from source py3521:46
apuimedoxD21:46
apuimedoso I got the dev part for sure21:46
apuimedoxD21:46
*** tonanhngo has joined #openstack-kuryr21:46
apuimedoivc_: I can confirm it fails21:47
ivc_interesting. so tox fails while testtools.run succeedes21:47
apuimedohttp://paste.openstack.org/show/597093/21:47
apuimedoivc_: can it be a problem due to another test that ran before it?21:48
* apuimedo didn't look at the code yet21:48
ivc_it shouldn't21:48
apuimedoI'm gonna take a quick look21:49
ivc_tox -e py35 '.*test_generate_lbaas_port_specs'21:50
ivc_fails too21:50
ivc_now i need to figure out how that one is different from 'python -m testtools.run'21:50
apuimedoivc_: it passes for you with the testtools.run as well?21:52
ivc_yup21:52
ivc_ok i have no idea wtf is wrong with it. gonna dig more21:54
apuimedoxD21:54
apuimedoIt's pretty wtf21:54
apuimedohey...22:02
apuimedoivc_: won't this be related to the map getting a different random ordering seed?22:02
apuimedoah no, it gets converted to spec22:03
ivc_it then get converted back to dict in __eq__22:03
ivc_but shouldn't map equality not rely on ordering?22:04
ivc_s/map/dict/22:04
apuimedoit's a comparableversionedobject, and yeah, in __eq__ it gets cast back to map22:06
apuimedoivc_: it should not, but it's the only thing that comes to my mind22:06
apuimedoexcept..22:06
ivc_unless there's some fundamental bug in test system22:08
ivc_thing is when you run tox, it uses '-m subunit.run' which shows errors while '-m testtools.run' works fine22:09
apuimedoivc_: well, at least there's nothing fucked with the test library assertion22:12
apuimedosince if you print(ret ==expected_ports) it prints False22:13
ivc_http://paste.openstack.org/show/597102/22:14
ivc_it prints True22:14
apuimedowell, yeah, outside of testr22:15
apuimedoI'm thinking __eq__ is not getting called22:15
apuimedoinside testr22:15
*** yamamoto has joined #openstack-kuryr22:15
apuimedofor god knows which reason22:15
apuimedoI'll test my assumption22:15
ivc_i'd understand if 'self.assertEqual' would not call __eq__, but if you say 'a == b' is also not calling '__eq__' than its ... something new22:17
ivc_apuimedo anyway, as long as http://paste.openstack.org/show/597102/ works as expected, i think we found a bug in testr22:19
apuimedoivc_: I can't live with this22:20
apuimedo:P22:20
ivc_i'm gonna have nightmares today for sure22:20
ivc_that level of uncertainty is disturbing22:20
apuimedoivc_: I found it!!!22:21
ivc_cool, whats that?22:21
apuimedothis is the code I put to find it22:21
apuimedoand here's the funny result22:22
apuimedohttp://paste.openstack.org/show/597104/22:23
apuimedolook at the versioned_object.changes value22:23
apuimedo:O22:23
ivc_oh22:23
ivc_WTF22:23
ivc_how is that 'comparable object' then...22:24
apuimedoxD22:24
*** dougbtv has quit IRC22:24
ivc_but thats good22:24
apuimedowell, it is comparable as long as you load it only once I suppose22:24
apuimedoxD22:24
ivc_its better that way than if it was an issue with testr/py3.522:25
apuimedoindeed22:25
ivc_now i wonder why it consistently works on py27 and testtools.run22:25
apuimedoivc_: I bet they just use 'values()' on a dict22:25
apuimedowhich is not randomized22:25
apuimedoand testtools.run sets the modifier not to randomize22:26
apuimedo(was added finally to py3)22:26
apuimedolet me dig it out22:26
ivc_well its not that interesting anymore. but that ComparableVersionedObject...22:27
ivc_i can easily fix that by doing 'reset_changes' before comparison22:27
ivc_or do we remove that mixin for good to avoid further confusion?22:28
apuimedoivc_: I think in testtools.run they must be setting https://docs.python.org/3/using/cmdline.html#envvar-PYTHONHASHSEED22:28
apuimedolet me try it out22:28
apuimedodafuq22:29
apuimedoivc_: in tox.ini we already had code to set the hash seed to 022:30
apuimedoivc_: ok, grab your seats22:31
apuimedoivc_: I removed the hashseed22:31
apuimedoand now the tests pass22:31
openstackgerritIlya Chukhnakov proposed openstack/kuryr-kubernetes: K8s Services support: LBaaSSpecHandler  https://review.openstack.org/42744022:31
ivc_i've added 'obj_reset_changes' to base __init__22:32
ivc_it makes sense by itself to not have 'changes' after __init__ imo22:32
ivc_and it solves the problem22:32
ivc_sort of22:32
ivc_does not mean ComparableVersionedObject is doing good job on __eq__22:32
apuimedoivc_: http://paste.openstack.org/show/597106/22:33
apuimedocheck this22:33
apuimedoI prefer this change to your patch22:33
apuimedo(since osloversionedobjects tox.ini does not have the hashseed set either)22:34
ivc_uhm, what 'this change' ?22:34
apuimedoivc_: sed -i /PYTHONHASHSEED/d tox.ini22:35
ivc_well i don't mind that change, but it does not fix the problem with __eq__22:36
apuimedowell, __eq__ works under normal conditions22:37
apuimedoor you mean that they should sort the changes?22:37
ivc_no, i mean i did not think that __eq__ accounts for changes and we should do 'obj_reset'22:38
apuimedooh, that's a bit of a funny behavior22:38
ivc_if you don't like adding it to __init__, i'll add it to the test itself22:38
apuimedoone sec22:38
ivc_that funny behaviour is the reason for test failure22:38
apuimedoyeah22:39
ivc_its just that adding it to a test would be much more invasive and imo it makes sense as part of __init__22:39
apuimedoivc_: I don't know enough about the obj_reset22:40
apuimedoI see it is used quite a bit in neutron22:40
apuimedobut I have to read the code22:40
apuimedoone sec22:40
ivc_its quite simple. we dont use it :)22:40
ivc_its purpose is to track changes to the object22:41
ivc_https://github.com/openstack/neutron/blob/master/neutron/objects/base.py#L32322:42
ivc_in neutron they reset right after they load it from DB22:42
apuimedodoes that mean that after further value changes (not the same attributes) you'd have to reset before __eq__?22:42
ivc_you just need to be aware of how __eq__ works (i.e. it accounts for updates too)22:43
apuimedoivc_: that's a bit silly22:43
apuimedothey should not have overloaded the meaning like that22:44
apuimedothey should have added an extra comparison method22:44
ivc_well they probably should22:44
apuimedowell, anyways, my fault for accepting OVO without reading it all22:44
apuimedoxD22:44
apuimedoI'm almost tempted to tell you to put it in an __eq__ method in our base objects22:44
ivc_we don't have a better alternative unfortunatelly22:45
apuimedobut then we'd have a mess between our objects and the ovos from os-vif22:45
apuimedoxd22:45
apuimedodarn...22:45
apuimedoanyway22:45
apuimedo__init__ is fine22:45
apuimedoalso, please drop the useless python hash seed22:45
apuimedotox sets it anyway22:45
apuimedoby defualt22:45
ivc_thats unrelated to this patch22:45
apuimedotrue22:45
apuimedoI'll send it22:45
ivc_also i've reverted some assertRaisesRegex changes by irenab like in https://review.openstack.org/#/c/427440/2/kuryr_kubernetes/tests/unit/controller/drivers/test_default_project.py@3822:47
apuimedoIt's a bit evil to do it in the same patch22:47
ivc_it just does not make sense to add a new test with assertRaises and keep that assertRaisesRegex22:48
* ivc_ is open for discussion why he thinks assertRaisesRegex is evil22:50
openstackgerritAntoni Segura Puimedon proposed openstack/kuryr-kubernetes: testing: drop zero hashseed  https://review.openstack.org/42749522:50
apuimedoivc_: I have to say I never saw it before irenab used it22:51
ivc_it would be funny if your hashseed patch fails jenkins py35 xD22:51
apuimedoivc_: I'll take it as a sign of God or Guido that I should stop looking at code today22:52
ivc_XD22:52
apuimedo(if it were to happen)22:52
*** limao has joined #openstack-kuryr22:56
ivc_apuimedo https://review.openstack.org/#/c/427440/ passed the gate now :)22:57
*** limao_ has joined #openstack-kuryr23:00
*** limao has quit IRC23:01
*** saneax-_-|AFK is now known as saneax23:01
*** yuanying has joined #openstack-kuryr23:04
« Monday, 2017-01-30 Index Wednesday, 2017-02-01 »

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