Monday, 2014-02-17

*** jcooley_ has joined #savanna00:07
*** tmckay has quit IRC00:10
*** jcooley_ has quit IRC00:13
*** matsuhashi has joined #savanna00:23
*** jcooley_ has joined #savanna00:30
*** matsuhashi has quit IRC00:33
*** matsuhashi has joined #savanna00:34
*** jcooley_ has quit IRC00:37
*** matsuhashi has quit IRC00:38
*** matsuhashi has joined #savanna01:04
*** tmckay has joined #savanna01:04
*** nosnos has joined #savanna01:05
*** jcooley_ has joined #savanna01:30
*** jcooley_ has quit IRC01:37
*** matsuhashi has quit IRC02:18
*** jcooley_ has joined #savanna02:26
*** matsuhashi has joined #savanna02:26
*** jcooley_ has quit IRC02:31
*** tmckay has quit IRC02:32
*** witlessb has joined #savanna02:56
*** witlessb has quit IRC02:56
*** jcooley_ has joined #savanna03:01
*** matsuhashi has quit IRC03:48
*** jcooley_ has quit IRC03:53
*** _nadya_ has joined #savanna04:02
*** _nadya_ has quit IRC04:03
*** matsuhashi has joined #savanna04:32
openstackgerritJenkins proposed a change to openstack/savanna: Imported Translations from Transifex  https://review.openstack.org/7091806:09
*** _nadya_ has joined #savanna06:13
*** saju_m has joined #savanna06:15
*** _nadya_ has quit IRC06:21
*** akuznetsov has quit IRC06:26
*** _nadya_ has joined #savanna06:27
*** dmitryme has joined #savanna06:28
*** akuznetsov has joined #savanna06:29
openstackgerritLee Li proposed a change to openstack/savanna: Remove dependent module py3kcompat  https://review.openstack.org/7353506:39
*** nosnos has quit IRC06:42
*** NikitaKonovalov is now known as NikitaKonovalov_06:49
*** nosnos has joined #savanna06:52
openstackgerritYongli He proposed a change to openstack/savanna: Use six.moves cStringIO instead of cStringIO  https://review.openstack.org/7394206:53
*** dmitryme has quit IRC06:56
*** _nadya_ has quit IRC07:03
*** _nadya_ has joined #savanna07:05
*** _nadya_ has quit IRC07:09
*** NikitaKonovalov_ is now known as NikitaKonovalov07:16
openstackgerritYongli He proposed a change to openstack/savanna: Use six.moves cStringIO instead of cStringIO  https://review.openstack.org/7394207:18
*** mattf is now known as _mattf07:21
*** boris-42 has quit IRC07:40
*** akuznetsov has quit IRC07:48
*** akuznetsov has joined #savanna07:51
*** saju_m has quit IRC08:31
*** sreshetniak has quit IRC08:45
*** dmitryme has joined #savanna08:54
*** saju_m has joined #savanna08:58
openstackgerritDmitry Mescheryakov proposed a change to openstack/savanna: Do eventlet.monkey_patch at the very beginning  https://review.openstack.org/7299208:59
*** witlessb has joined #savanna09:08
*** DinaBelova_ is now known as DinaBelova09:50
*** boris-42_ has joined #savanna09:52
*** bogdando has joined #savanna09:53
*** matsuhashi has quit IRC09:54
*** matsuhashi has joined #savanna10:02
*** aignatov is now known as aignatov_10:02
*** akuznetsov has quit IRC10:08
openstackgerritDaniele Venzano proposed a change to openstack/savanna: Add Spark plugin to Savana  https://review.openstack.org/7398910:09
*** akuznetsov has joined #savanna10:09
*** matsuhashi has quit IRC10:17
*** matsuhashi has joined #savanna10:17
*** matsuhashi has quit IRC10:22
openstackgerritNikita Konovalov proposed a change to openstack/savanna: Moving rest to Pecan/WSME framework  https://review.openstack.org/6390810:29
*** aignatov_ is now known as aignatov10:37
*** dmitryme has quit IRC10:38
openstackgerritSergey Reshetnyak proposed a change to openstack/savanna-image-elements: Add hadoop v2 support to hadoop element  https://review.openstack.org/7265110:38
*** NikitaKonovalov is now known as NikitaKonovalov_10:39
*** matsuhashi has joined #savanna10:47
*** tosky has joined #savanna10:50
openstackgerritSergey Reshetnyak proposed a change to openstack/savanna: Add Hadoop 2 vanilla plugin  https://review.openstack.org/7364110:52
openstackgerritA change was merged to openstack/python-savannaclient: Remove dependent module py3kcompat  https://review.openstack.org/7375711:06
openstackgerritA change was merged to openstack/savanna: Sort modules in openstack-common.conf  https://review.openstack.org/7305011:29
*** DinaBelova is now known as DinaBelova_11:36
*** matsuhashi has quit IRC11:37
*** aignatov is now known as aignatov_11:40
*** aignatov_ is now known as aignatov11:50
*** aignatov is now known as aignatov_11:51
*** aignatov_ is now known as aignatov11:57
openstackgerritSergey Reshetnyak proposed a change to openstack/savanna: Spped up attaching cinder volumes  https://review.openstack.org/7401512:33
openstackgerritSergey Reshetnyak proposed a change to openstack/savanna: Attach volumes in parallel  https://review.openstack.org/7401512:39
*** _mattf is now known as mattf12:45
mattfanyone have thoughts as to what's causing https://review.openstack.org/#/c/73010/ jenkins to fail?12:46
openstackgerritWei Wang proposed a change to openstack/savanna: Use cStringIO from six.moves instead of cStringIO  https://review.openstack.org/7402612:55
openstackgerritSofiia Kostiuchenko proposed a change to openstack/savanna: Added SCALE_ADDITIONAL_NODES_NUMBER and SCALE_NEW_NODES_NUMBER parameters to HDP plugin config.  https://review.openstack.org/7230012:58
*** IvanBerezovskiy has joined #savanna13:07
*** nosnos has quit IRC13:08
mattfaignatov, pls take another look at https://review.openstack.org/#/c/72933/13:13
aignatovmattf, why do not create bug fix and unit tests in a single patch?13:16
*** DinaBelova_ is now known as DinaBelova13:24
*** NikitaKonovalov_ is now known as NikitaKonovalov13:25
aignatovok, I've approved but I think bug fix related code and unit tests for it should go in the single patch set.13:31
*** aignatov is now known as aignatov_13:43
mattfaignatov_, the bug fix and unit test for the bug are in the single patch/commit13:59
*** aignatov_ is now known as aignatov14:03
*** saju_m has quit IRC14:15
openstackgerritDmitry Mescheryakov proposed a change to openstack/savanna: Intial Agent remote implementation  https://review.openstack.org/7101514:24
mattfanyone have thoughts as to what's causing https://review.openstack.org/#/c/73010/ jenkins to fail?14:25
mattfit's py27 only. i can't reproduce locally. the code that's failing isn't obviously touched by the patch.14:26
aignatovjust try 'recheck no bug' ;) if you think that this some unrelated jenkins fails :)14:31
mattfi do, it's been failing consistently. patches that dep on it have been passing though.14:33
mattfi'm not sure how to debug it at this point.14:33
*** tmckay has joined #savanna14:42
*** jmaron has joined #savanna14:48
*** eanxgeek|log is now known as eanxgeek14:48
openstackgerritA change was merged to openstack/savanna: Fix swift data source credential validation  https://review.openstack.org/7293314:49
*** themistymay has joined #savanna14:53
*** tmckay has left #savanna14:56
*** tmckay has joined #savanna14:56
*** rp__ has joined #savanna14:58
*** jmaron has quit IRC15:01
*** rp__ is now known as zerstoeren15:01
*** _nadya_ has joined #savanna15:14
*** jmaron has joined #savanna15:15
*** _nadya_ has quit IRC15:19
*** eanxgeek is now known as eanxgeek|log15:19
*** coolsvap has joined #savanna15:26
*** themistymay has quit IRC15:27
*** jmaron has quit IRC15:32
*** themistymay has joined #savanna15:33
*** jmaron has joined #savanna15:47
*** IvanBerezovskiy has left #savanna15:47
tmckaySergeyLukjanov, ping, tempest question15:50
openstackgerritJonathan Maron proposed a change to openstack/savanna: Enable use of rootwrap for neutron private IP interactions  https://review.openstack.org/7339615:52
*** aignatov is now known as aignatov_15:54
*** NikitaKonovalov is now known as NikitaKonovalov_15:59
SergeyLukjanovtmckay, pong, gtg for 30-40 mins, short comments: I saw your email, cli patch will updated today16:01
*** boris-42_ has quit IRC16:01
tmckaySergeyLukjanov, okay.  When you're available, I have a question about syncing client/api changes in projects with tempest tests16:02
*** DinaBelova is now known as DinaBelova_16:08
*** zerstoeren has quit IRC16:13
*** jmaron has quit IRC16:19
*** crobertsrh2 has joined #savanna16:48
*** DinaBelova_ is now known as DinaBelova17:02
*** tosky has quit IRC17:13
*** zerstoeren has joined #savanna17:19
*** akuznetsov has quit IRC17:26
openstackgerritA change was merged to openstack/savanna: Add support retrying rest call in IDH plugin  https://review.openstack.org/6965617:32
*** zerstoeren has quit IRC17:33
*** _nadya_ has joined #savanna17:34
openstackgerritA change was merged to openstack/savanna: Fix cluster scaling in IDH plugin  https://review.openstack.org/7045217:37
*** boris-42_ has joined #savanna17:40
*** akuznetsov has joined #savanna17:42
*** akuznetsov has quit IRC17:42
*** mattf is now known as _mattf17:44
*** akuznetsov has joined #savanna17:47
SergeyLukjanovtmckay, I'm partially here17:48
SergeyLukjanovtmckay, what's up?17:48
tmckaySergeyLukjanov, thanks.  A couple of questions17:49
tmckaySergeyLukjanov, first question, is it acceptable to use the python-savannaclient cli as the starting point for a scenario test, rather than the client itself?  Since the "cli" test dir contains readonly ops, this is the only way to do real end-to-end testing of the cli in tempest17:50
tmckayI haven't seen an example of this, but no explicit objection either...17:50
SergeyLukjanovtmckay, in tempest scenarios tests are client-based17:51
SergeyLukjanovtmckay, anyway, we should test client itself, cli isn't so important as python bindigs17:51
SergeyLukjanovwe're using python bindings in tests, dashboard and cli17:51
SergeyLukjanovso, it should be the high prio tested17:52
SergeyLukjanovand I'm sure that it's enough to have transformation tests for cli17:52
SergeyLukjanovunit tests that'll test how cli args converted to python bindings calls17:53
tmckayBut there is still a gap.  How do we know that the cli actually works, if there are not tests in savanna or tempest that start with the cli and do more than readonly operations?17:53
tmckayhmm, transformation tests are an interesting idea17:53
SergeyLukjanovunit tests will test correctness of cli args -> python bindings transforms17:53
SergeyLukjanovbtw I've seen tons of projects with such approach17:54
tmckayso those unit tests would be based in savanna? (as opposed to tempest)17:54
SergeyLukjanovtmckay, it'll be based in client17:55
SergeyLukjanovtmckay, you can find examples in nova client17:55
tmckaySergeyLukjanov, right, okay, the python-savannaclient unit tests.17:55
SergeyLukjanovtmckay, exactly17:55
tmckaySergeyLukjanov, the second question is about syncing changes in tempest and api/client for a project17:55
*** _mattf is now known as mattf17:56
tmckayFor instance, your patch currently is using plurals for plugins-list and plugins-show, and the requirements call out 0.4.117:56
SergeyLukjanovtmckay, for example, here is my change to python-novaclient where I'm adding new API feature support to client and CLI - https://review.openstack.org/#/c/3144017:56
SergeyLukjanovtmckay, tempest runs tests for the client muster, that's why savanna job is failing17:56
tmckaybut, running tempest from /opt/stack with the latest master branches for everything, those tests are out of date.  they are now plugin-list and plugin-show17:57
tmckayyes17:57
SergeyLukjanovCLI should be stable after introducing tempest tests17:57
SergeyLukjanovit's usually implemented for stable API17:57
*** akuznetsov has quit IRC17:57
SergeyLukjanovso, CLI is stable enough due to the API stability17:57
tmckayah.  What if a change needs to be made?  Do we strive for backward compatibility>17:58
tmckay?17:58
SergeyLukjanovtmckay, I17:58
SergeyLukjanovI17:58
SergeyLukjanovoops :)17:58
SergeyLukjanovI'm not sure how it should be done17:58
SergeyLukjanovbut that's how CLI tested in OpenStack17:59
SergeyLukjanovin tempest only bunch of very simple read-only tests17:59
SergeyLukjanovand everything else in unit tests of python-XXXclient17:59
*** akuznetsov has joined #savanna18:00
tmckayokay.  So stability of the cli and the api is expected.  If something *does* change, I imagine writing some kind of compat code would be the way to go.  Or, everything has to be synced around a release milestone.18:01
*** zerstoeren has joined #savanna18:01
SergeyLukjanovtmckay, anyway, we'll have a small bunch of tests in tempest, so, I don't think that we'll have issues often18:02
SergeyLukjanovand if we'll have, we could add some conditional checks18:02
SergeyLukjanovand that's why I'd like to hold my CLI change right now - to stabilize CLI18:03
SergeyLukjanovfor example, it looks like we should rename table headers for list/get ops18:03
mattfaignatov_, can i get a +2 on https://review.openstack.org/#/c/73111/ ?18:03
SergeyLukjanovto be consistent with other OpenStack CLIs18:03
tmckaySergeyLukjanov, when the tests run in the real tempest gate, does it pull from master or does it look for a released version?  If it pulls from master, your work-in-progress CR should be failing too (but I haven't found the error message yet)18:06
SergeyLukjanovtmckay, it's working on master + CR18:06
SergeyLukjanovtmckay, and it's failing :)18:06
tmckay:) okay, thanks18:07
mattfSergeyLukjanov, lend a +A to https://review.openstack.org/#/c/73261/ ?18:07
SergeyLukjanovtmckay, check-tempest-dsvm-savanna-full-nv FAILURE in 48m 07s (non-voting)18:07
SergeyLukjanovtmckay, savanna is enabled only in jobs with name *savanna*18:07
SergeyLukjanovmattf, it depends on not-approved patch anyway18:07
SergeyLukjanovmattf, but I'll +A it :)18:08
mattfyeah, i'm still stuck on https://review.openstack.org/#/c/73010/418:09
SergeyLukjanovmattf, I'm advise to avoid such chains of absolutely independent changes - they could stuck on review for a while18:09
tmckaySergeyLukjanov, okay, I found the right log and the error. ;-)18:09
mattfi can't get it to break locally18:09
mattfSergeyLukjanov, yeah, i'm leaning my lesson this time around18:09
SergeyLukjanovmattf, re py27 fail, it looks very surprising18:11
SergeyLukjanovmattf, what's your feeling about CLI stability?18:11
SergeyLukjanovtmckay, ^^18:11
mattfi've not been following, what's the question?18:15
mattfSergeyLukjanov, i'm coming up empty tr py27 fail...18:15
SergeyLukjanovmattf, the question is that I have a patch with CLI base/tests for tempest, but we should be sure that we'll break it in the nearest future to avoid pain with updating tempest on the go18:16
SergeyLukjanovmattf, it'll contain only read-only very simple checks18:17
mattfissue is how do we update the cli & tempest at the same time if they are in separate repos?18:17
SergeyLukjanovmattf, I'll try to take a look on your py27 fail today night/tomorrow18:17
SergeyLukjanovmattf, yup, but it's quite synthetic issue because we should add tests when we feel API and CLI stable18:18
SergeyLukjanovmattf, tmckay, I'll try to research how other projects dealing with it18:19
mattfthe cli should be at a the point of a v1.018:19
mattfthere's a proposed change to guess --name or --id, but that could be done in a backward compat way, so v1.118:20
SergeyLukjanovmattf, yup18:20
SergeyLukjanovmattf, and headers should be update18:20
SergeyLukjanovmattf, like Name vs. name to be consistent with other CLIs18:21
mattfName vs name ?18:21
mattfoh, table headers?18:22
SergeyLukjanovyup18:22
mattfimho table headers are case-insensitive18:22
SergeyLukjanovi've missed the 'table' word :)18:22
SergeyLukjanovbut it's inconsistent with all other clients, so, should be polished18:22
mattfi immediately thought #include <error> and was sad18:22
mattfwe can certainly poliish it.18:23
mattfas a general rule, we should not be facilitating screen-scraping w/ the cli18:23
openstackgerritA change was merged to openstack/savanna: Enable HDP 2 deployment leveraging HDP plugin  https://review.openstack.org/7196718:24
SergeyLukjanovmattf, I'll finalize tempest patch today, to just have a base and then when we'll be ready ask folks to review18:24
mattfi.e. if you write savanna plugin-list | grep name and get nothing in the future because we changed it to Name, the response should be "don't do that, use the client"18:24
SergeyLukjanovmattf, cli isn't release yet18:24
SergeyLukjanovand use grep -i always :)18:24
mattfcli != api18:25
mattfthough i'm guilty of using cli's as apis18:25
mattfexpectations just need to be set appropriately =18:25
SergeyLukjanovmattf, due to the fact that savanna is part of openstack expectation is to be consistent with other projects18:26
mattfdo other project support the cli as an api?18:27
SergeyLukjanovmattf, what 'support' does you mean?18:27
*** zerstoeren has quit IRC18:27
mattfcommunity support where you'd provide a deprecation path w/ backward compatibility for "grep name" when "grep Name" would work18:28
mattfthere shouldn't be a deprecation path for something like name vs Name in a table header, or '  id  |' vs '       id            |' for that matter18:29
mattfif you're using the cli as an api you're either doing it wrong or you're assuming the responsibility if things break. the community isn't "supporting" it.18:29
SergeyLukjanovmattf, I don't think that CLI is something so important to worry about such stuff18:30
SergeyLukjanovand I think that python bindings should be always used for any automationg18:30
SergeyLukjanovanyway we could agree on one of the approaches and follow it18:31
mattfso we're in agreement about the "and i think"18:31
SergeyLukjanovyup18:31
mattfno matter how important we think the cli is, people will try to use it as an api.  we need to set expectations properly there, or we'll end up with upset users18:31
mattfor worse, we'll be stuck not being able to change the cli because it would break users18:31
mattfhow is this related to your topic of the header appearing upper or lower case for human eyes to consume? i agree we should be visually consistent, but we should avoid locking ourselves into a particular expected output for the cli18:33
SergeyLukjanovoh, I found that keystone and neutron are using lower case table headers18:33
SergeyLukjanovheh, so, there's already no consistency in CLIs :(18:34
SergeyLukjanovso, I'm ok with both approaches18:34
mattfcli tests could be non-voting. maybe we want to just delete 'plugin-list' in the future?18:34
* mattf shakes fist at py2718:34
SergeyLukjanovmattf, I think that when we'd like to move our CLI to the next API version, we could just add the second bunch of API tests18:35
mattfi need a better workflow for managing one-off changes18:35
SergeyLukjanovmattf, I'm using branches for all change18:35
mattfat the end of the day i want our cli to work, i want it to be free to change, but i don't want anyone to ever open a bug that says "hey, you changed the cli output"18:36
SergeyLukjanovgit fetch gerrit && git checkout -b super_mega_hack gerrit/master18:36
SergeyLukjanovmattf, sounds very nice :)18:36
* mattf lives in a dreamworld sometimes18:37
mattfimho, if you want the cli to be an api, then you provide functions that let the user format the output18:37
mattfi think keystone (or was it neutron) are starting to do that18:38
mattfe.g. savanna plugin-list --format "%s,%s" name,version18:38
mattfthen i can parse vanilla,1.2.1 in a repeatable fashion18:38
mattfthat's all cli v2 or v3 imho18:39
*** _nadya_ has quit IRC18:39
SergeyLukjanovsounds nice18:41
SergeyLukjanovcurrently, I just would like to be sure that our current read-only CLI will be quite stable and add simple tests for them to tempest18:42
*** themistymay has quit IRC18:43
*** aignatov_ is now known as aignatov18:48
*** openstackgerrit has quit IRC18:53
*** openstackgerrit has joined #savanna18:54
openstackgerritMatthew Farrellee proposed a change to openstack/savanna: Add run_test.sh for running tests  https://review.openstack.org/7294818:58
mattfok, restructured all my reviews18:59
mattfaignatov, SergeyLukjanov, if you can +A https://review.openstack.org/#/c/72948/ it'll get me unstuck18:59
aignatovok, +219:00
aignatov:-)19:00
SergeyLukjanovmattf, you should rebase all dependent patches too19:00
SergeyLukjanovmattf, because they have outdated commit id in parants19:00
* mattf now has master + master-[0-4]19:01
mattfthat's in progress19:01
SergeyLukjanovit could be done in UI19:01
SergeyLukjanovby the rebase button19:01
SergeyLukjanovor you'd like to split the chain completely?19:01
SergeyLukjanovbtw in gerrit 2.8 you can rebase changes in UI to the specified refs19:02
mattfthey all depend on that run_tests.sh, because i use run_tests.sh to verify them19:02
* mattf clings to his cli19:02
openstackgerritMatthew Farrellee proposed a change to openstack/savanna: Expand swift data source credential tests  https://review.openstack.org/7301019:02
mattfi split the chain "complete" (sans the run_tests.sh)19:02
openstackgerritMatthew Farrellee proposed a change to openstack/savanna: Shorten swift-internal:// to swift://  https://review.openstack.org/7326119:03
openstackgerritMatthew Farrellee proposed a change to openstack/savanna: Delete 'links' only if it is present  https://review.openstack.org/7383619:03
openstackgerritMatthew Farrellee proposed a change to openstack/savanna: Expand node-group-template usage validation msg  https://review.openstack.org/7388819:04
SergeyLukjanovmattf, I'll approve the patches with 2x+2 and 2x+1V19:05
*** aignatov is now known as aignatov_19:07
*** jcooley_ has joined #savanna19:15
openstackgerritSergey Lukjanov proposed a change to openstack/savanna: Auto generate and check config sample  https://review.openstack.org/7305319:37
openstackgerritSergey Lukjanov proposed a change to openstack/savanna: Standardize config sample locations  https://review.openstack.org/7305219:37
openstackgerritSergey Lukjanov proposed a change to openstack/savanna: Fix how migration's cli register db connection opt  https://review.openstack.org/7305119:37
*** aignatov_ is now known as aignatov19:38
*** _nadya_ has joined #savanna19:38
SergeyLukjanovmattf, please, take a look on https://review.openstack.org/#/c/73051/ and https://review.openstack.org/#/c/73052/19:38
SergeyLukjanovaignatov, please, take a look at https://review.openstack.org/#/c/73052/19:38
*** aignatov is now known as aignatov_19:42
*** _nadya_ has quit IRC19:52
openstackgerritSergey Reshetnyak proposed a change to openstack/savanna: Add ability to support several versions vanilla plugin  https://review.openstack.org/7414520:11
*** witlessb has quit IRC20:13
*** witlessb has joined #savanna20:13
*** jcooley_ has quit IRC20:25
*** akuznetsov has quit IRC20:29
*** witlessb has quit IRC20:29
*** mattf is now known as _mattf20:31
*** aignatov_ is now known as aignatov20:48
*** DinaBelova is now known as DinaBelova_20:54
openstackgerritA change was merged to openstack/savanna: Add run_test.sh for running tests  https://review.openstack.org/7294820:56
*** _mattf is now known as mattf21:04
*** coolsvap has quit IRC21:44
mattfaignatov, SergeyLukjanov, quick before ci decides otherwise +A https://review.openstack.org/#/c/73010/21:48
mattftho, i guess we have to get lucky twice in a row for it to actually go through21:48
* mattf shakes fist at ghost in the py2721:48
SergeyLukjanovmattf, it's already +2'ed by me21:48
mattfdanke21:50
* mattf learns spasibo21:52
*** jmaron has joined #savanna21:59
SergeyLukjanovmattf, :)22:03
*** aignatov is now known as aignatov_22:10
openstackgerritSergey Lukjanov proposed a change to openstack/python-savannaclient: Make the single entrypoint for client  https://review.openstack.org/7418422:34
openstackgerritA change was merged to openstack/savanna: Shorten swift-internal:// to swift://  https://review.openstack.org/7326122:34
*** jmaron has quit IRC22:49
*** tmckay is now known as _tmckay23:14
openstackgerritA change was merged to openstack/savanna: Delete 'links' only if it is present  https://review.openstack.org/7383623:27
*** jcooley_ has joined #savanna23:37
*** mattf is now known as _mattf23:55

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