Tuesday, 2015-08-25

dave-mccowanis it possible the database was non-empty when you started?00:00
rm_workall three are seeing 10 things and expecting 0, 1, 200:00
rm_workhmm00:00
rm_workMAYBE00:00
rm_worki'll have to restart00:00
rm_workbut it is prolly fine00:00
rm_worki'll test later tonight00:00
* rm_work leaves00:00
*** rm_work is now known as rm_work|away00:01
*** lisaclark1 has joined #openstack-barbican00:47
*** lisaclark1 has quit IRC00:47
*** SheenaG has joined #openstack-barbican00:48
*** lisaclark1 has joined #openstack-barbican00:48
*** everjeje has quit IRC01:02
*** lisaclark1 has quit IRC01:07
openstackgerritDave McCowan proposed openstack/barbican: Cleanup database after functional test runs  https://review.openstack.org/21625501:23
*** zz_dimtruck is now known as dimtruck01:25
*** gyee has quit IRC01:29
*** jhfeng has joined #openstack-barbican01:34
*** insequent has joined #openstack-barbican01:47
*** dave-mccowan has quit IRC01:56
*** dave-mccowan_ is now known as dave-mccowan01:56
*** lisaclark1 has joined #openstack-barbican02:00
*** lisaclark1 has quit IRC02:01
*** lisaclark1 has joined #openstack-barbican02:01
*** insequent has quit IRC02:08
*** silos has joined #openstack-barbican02:24
*** insequent has joined #openstack-barbican02:28
*** lisaclark1 has quit IRC02:32
*** tkelsey has joined #openstack-barbican02:36
*** SheenaG has quit IRC02:39
*** tkelsey has quit IRC02:40
*** silos has left #openstack-barbican02:42
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Remove test behaviors abstraction for secrets  https://review.openstack.org/21341402:50
*** vivek-ebay has quit IRC03:08
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Remove test behaviors abstraction for secrets  https://review.openstack.org/21341403:17
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Allow Barbican Client Secret Update Functionality  https://review.openstack.org/19687603:28
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Add Unit Tests for Store and Update Payload when Payload is zero  https://review.openstack.org/21213003:29
*** vivek-ebay has joined #openstack-barbican03:31
*** jhfeng has quit IRC03:34
*** dimtruck is now known as zz_dimtruck03:50
*** xaeth is now known as xaeth_afk04:15
*** xaeth_afk is now known as xaeth04:22
*** insequent has quit IRC04:29
*** insequent has joined #openstack-barbican04:30
*** vivek-ebay has quit IRC04:41
openstackgerritPradeep Kumar Singh proposed openstack/barbican: Make barbican/tests/plugin/crypto/test_p11_crypto py3 compatible  https://review.openstack.org/21652704:48
*** woodster_ has quit IRC04:49
*** xaeth is now known as xaeth_afk04:52
*** insequent has quit IRC04:57
*** rm_work|away is now known as rm_work05:03
openstackgerritAdam Harwell proposed openstack/barbican: Use the new Devstack external plugin method  https://review.openstack.org/16788505:06
*** insequent has joined #openstack-barbican05:14
*** insequent has quit IRC05:23
*** insequent has joined #openstack-barbican05:40
*** jaosorior has joined #openstack-barbican06:01
*** Nirupama has joined #openstack-barbican06:05
*** jaosorior has quit IRC06:12
*** DuncanT has quit IRC06:12
*** jillysciarilly has quit IRC06:13
*** DuncanT has joined #openstack-barbican06:15
*** jaosorior has joined #openstack-barbican06:16
*** jillysciarilly has joined #openstack-barbican06:18
*** yuanying has quit IRC06:25
*** Nirupama has quit IRC06:36
*** yuanying has joined #openstack-barbican06:44
*** tkelsey has joined #openstack-barbican07:04
openstackgerritChelsea Winfree proposed openstack/barbican: Add PUT support for generic container types  https://review.openstack.org/20724907:04
openstackgerritOpenStack Proposal Bot proposed openstack/barbican: Updated from global requirements  https://review.openstack.org/21656007:06
*** tkelsey has quit IRC07:08
*** insequent has quit IRC07:31
*** insequent has joined #openstack-barbican07:32
*** tkelsey has joined #openstack-barbican07:47
*** everjeje has joined #openstack-barbican08:00
*** shohel has joined #openstack-barbican08:04
*** shohel has quit IRC08:04
*** jaosorior has quit IRC08:05
*** Nirupama has joined #openstack-barbican08:29
*** insequent has quit IRC08:35
*** insequent has joined #openstack-barbican08:36
*** insequent has quit IRC08:43
*** pksingh has quit IRC08:58
*** jaosorior has joined #openstack-barbican08:59
*** pksingh has joined #openstack-barbican09:00
*** insequent has joined #openstack-barbican09:02
*** insequent has quit IRC09:59
*** mmdurrant has quit IRC10:09
*** everjeje has quit IRC10:12
*** insequent has joined #openstack-barbican10:17
*** dave-mcc_ has joined #openstack-barbican10:23
dave-mcc_jaosorior ping10:45
jaosoriordave-mcc_: pong10:46
dave-mcc_have you noticed the issues in the quotas functional tests running in parallel?10:46
jaosoriorNope. Gotta check it out. Got a log?10:48
dave-mcc_they're valid issues the way the tests are written.  all the tests share the same database, so counting the number of things in the DB won't be valid when more than one test runs at once.  the idea i had to "fix" them is to just not run certain tests in parallel.10:49
dave-mcc_i added a patch to run_tests.sh to stop parallel runs in the gate and that works fine.  but, if someone uses tox (like in the devstack gate), then the tests run in parallel again.10:50
dave-mcc_i have several ideas on ways to fix this, but i don't like any of them. :-)10:51
jaosoriorNot sure if that's an ideal solution10:51
dave-mcc_another example of this problem/fix is in SecretsPagingTestCase() in functional/test_secrets.py.10:52
dave-mcc_they have the same issue... counting the number of secrets in the DB.... they fix it by using the filter= param in the GET API, so each test has a unique set of data to count.10:53
dave-mcc_but, v1/project-quotas does not have a filter= param.  (i could add one to make the tests work) :-/10:53
dave-mcc_i could just skip/delete all the test cases that don't support concurrency. :-/10:54
dave-mcc_i've been looking for an attribute or flag or something to make tox/nose not run my suite in parallel, but could not find one. :-(10:55
dave-mcc_any other ideas?10:56
*** insequent has quit IRC10:57
*** insequent has joined #openstack-barbican11:15
jaosoriorGotta think about it :/ not a trivial thing11:21
*** insequent has quit IRC11:21
*** shohel has joined #openstack-barbican11:27
jaosoriordave-mcc_ another option would be to have different projects per test suite... But that would complicate the deployment quite a lot...11:35
jaosoriorwell, the test deployment11:35
jaosoriorso the most straight forward fixes are either we disable parallelism, or we filter somehow...11:37
jaosoriorlike you proposed11:37
dave-mcc_jaosorior project quotas are per-project, so for the paging tests there are already 20+ projects used to verify limit/offset actions.  (barbican does not verify with keystone if a project exists in keystone for the purpose of quotas)11:37
jaosoriorah, true, true, was thinking more about the secretPaging test. But yeah, you're right11:38
*** insequent has joined #openstack-barbican11:38
dave-mcc_jaosorior i think skipping is right way to go.  adding filtering seems nice, but we'd wind up just testing filtering code (which is not the intended code under test).11:41
*** shohel has quit IRC11:43
jaosoriordave-mcc_ well, we do have attributes for the functional tests http://testtools.readthedocs.org/en/latest/for-test-authors.html#test-attributes11:43
jaosoriorI'm trying to check if we could mark the tests (even the test-suits) which can be ran concurrently11:44
jaosoriorand just run those11:44
jaosoriorthen sequentially run the ones which can't be easily parallelized11:44
jaosoriornosetests can do that, but not sure if testr11:44
*** insequent has quit IRC11:45
*** insequent has joined #openstack-barbican11:46
dave-mcc_jaosorior that would be good.  i got testr to skip my paging tests using command line params.11:47
jaosoriordave-mcc_ exactly. I'm still checking if it's possible11:48
*** Kevin_Zheng_ has joined #openstack-barbican11:51
thervedave-mcc_, You can tell testr to not use parallelism11:55
jaosoriordave-mcc_ seems to me like that attr is not really taken into account by testr :/11:55
therveIdeally the tests would be fixed, though11:55
jaosoriortherve: while that is possible, and probably an easy solution, what I would dislike about that is how slow the gate runs will get11:55
therveRight11:55
dave-mcc_therve the tests aren't really broken, they just don't support concurrency.11:56
*** shohel has joined #openstack-barbican11:57
*** tkelsey has quit IRC11:57
dave-mcc_to get them to support concurrency i'd either have to water-down certain tests to the point they aren't testing completely anymore, or modify the code under test to the point i'm not testing the real feature any more.  (or so it seems so far)11:58
dave-mcc_therve, for the gate, i told testr to skip the parallel run on my test cases that don't support concurrency.  that's been working for the gate.  but, rm_work's patch for a future devstack gate uses tox and nose.11:59
dave-mcc_jaosorior did you say you have a fix that works for nosetests?  or are you still looking?12:01
thervedave-mcc_, OK. In my experience it's rarely impossible, but I don't know enough of your feature.12:01
jaosoriordave-mcc_ The documentation says one thing and the options available in nosetests say another :/12:02
dave-mcc_therve yea, definitely possible.  it's "just software".  :-)12:02
therve:)12:02
Kevin_Zheng_I have tried to config the sql_connection in barbican.conf, I found there are multiple barbican.conf in different path, is it just me have done something wrong or it is just like that?12:04
thervejaosorior, One option may be to have 2 testr runs, one with concurrency and one without, using filtering to differentiate12:05
dave-mcc_Kevin_Zheng_ I think most of us will edit the one in etc/barbican/barbican.conf.  then when we start barbican, we use "bin/barbican.sh install".  that script with the install option will copy all of the config files to another directory where they are used by the running code.12:06
jaosoriortherve: I was looking into doing that filtering, and actually having that solution. But I haven't found a way to do it12:06
jaosoriortherve: Any ideas?12:06
therveLet's see12:07
Kevin_Zheng_dave-mcc_ Ok, I see, so one is for install and one is for the actual program12:08
dave-mcc_jaosorior therve: the secrets test cases use filtering, but that's because the secrets API supports filtering.  i could easily copy that solution, but that would mean adding a filter parameter to the project-quotas API.   (which is weird to change the code under test to get contrived test cases to pass) :-)12:09
dave-mcc_jaosorior therve, i guess i could add some post-filtering to the test cases, if i'm careful about project naming.12:10
thervedave-mcc_, Not sure we're talking about the same thing. I'm talking about filtering test names.12:10
dave-mcc_therve ah.  got it.  for testr, i've used regex to skip the offending ones. i grouped them all into one class to make them easy to pick out by name.   so, we just need a solution for nose.12:11
jaosoriortherve: I was hoping to be able to filter tests through test attributes... since we have plenty defined. But looking forward into that it seems to me like they are not useful at all with the tools we are using12:12
thervedave-mcc_, Why not use just one tool?12:13
jaosoriortherve: filtering by name should be possible12:13
jaosoriorbut, not very easily maintainable12:13
jaosoriorsince we need to keep that list updated12:13
thervejaosorior, RIght that's the "easy" path12:13
*** zz_dimtruck is now known as dimtruck12:13
jaosoriortherve: Easy as in straight-forward12:14
therveYeah12:14
jaosoriorwhat I would have liked is to use this http://testtools.readthedocs.org/en/latest/for-test-authors.html#test-attributes12:14
dave-mcc_therve barbican uses both tools already.  not sure why.  many people, many gates, long history...12:14
jaosoriorI don't really like testr... but not much I can do about it :P12:15
thervejaosorior, Interesting, let me see12:15
therveIMO, testr is not great, but nose is worse12:15
*** alee has quit IRC12:16
jaosoriortherve: Have had less issues with nose... buuut anyway, probably it is a good idea to have just one tool. Would probably help people to focus and just learn one for this project12:16
dave-mcc_it would be nice if the skipping could be done in the code, instead of on the command line,  something like: if getattr(os, "running_parallel"): self.skipTest("No concurrency support")12:20
thervejaosorior, Ah. test run '\[tag-name\]' seems to work12:36
therves/test/testr12:36
jaosoriortherve: not bad! Then that's an approach we could use12:38
dave-mcc_therve, jaosorior that's close to what i have for the gate run. https://github.com/openstack/barbican/blob/master/functionaltests/run_tests.sh#L3812:46
dave-mcc_but for "tox -e functional", which other runs use: https://github.com/openstack/barbican/blob/master/tox.ini#L58   it's nose based.12:47
jaosoriordave-mcc_: sooo... maybe we could switch it to use testr also12:51
*** mmdurrant has joined #openstack-barbican12:53
dave-mcc_jaosorior in tox.ini the processes=-1 was just added in March.  another option is to revert that change.12:59
*** insequent has quit IRC12:59
*** insequent has joined #openstack-barbican13:00
*** Nirupama has quit IRC13:03
*** dimtruck is now known as zz_dimtruck13:05
*** nelsnelson has joined #openstack-barbican13:08
*** rellerreller has joined #openstack-barbican13:12
*** alee has joined #openstack-barbican13:27
*** zz_dimtruck is now known as dimtruck13:38
openstackgerritDave McCowan proposed openstack/barbican: Skip Functional Tests that don't support concurrency when using concurrency  https://review.openstack.org/21671513:46
dave-mcc_therve, jaosorior, rm_work: ^^^ let's give this a shot.  for nose, skip by attribute.  for testr, skip by name (already merged).13:47
*** spotz_zzz is now known as spotz13:59
*** spotz is now known as spotz_zzz14:08
*** xaeth_afk is now known as xaeth14:13
*** pglass has joined #openstack-barbican14:20
*** jlhinson has joined #openstack-barbican14:21
*** insequent has quit IRC14:21
*** spotz_zzz is now known as spotz14:24
*** lisaclark1 has joined #openstack-barbican14:38
*** edtubill has joined #openstack-barbican14:38
openstackgerritDave McCowan proposed openstack/barbican: Skip Functional Tests that don't support concurrency when using concurrency  https://review.openstack.org/21671514:39
*** chadlung has joined #openstack-barbican14:40
*** lisaclark1 has quit IRC14:51
*** diazjf has joined #openstack-barbican14:51
*** lisaclark1 has joined #openstack-barbican14:55
*** lisaclark1 has quit IRC14:56
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Allow Barbican Client Secret Update Functionality  https://review.openstack.org/19687614:56
*** lisaclark1 has joined #openstack-barbican14:56
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Add Unit Tests for Store and Update Payload when Payload is zero  https://review.openstack.org/21213014:58
*** kebray has joined #openstack-barbican15:00
jaosoriordave-mcc_: let me know when it passes15:02
*** rellerreller has quit IRC15:03
*** rellerreller has joined #openstack-barbican15:12
*** dave-mcc_ has quit IRC15:12
*** Kevin_Zheng_ has quit IRC15:15
*** ccneill has joined #openstack-barbican15:17
*** SheenaG has joined #openstack-barbican15:20
*** rm_work is now known as rm_work|away15:24
*** dave-mcc_ has joined #openstack-barbican15:26
*** vivek-ebay has joined #openstack-barbican15:32
*** kebray has quit IRC15:38
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Remove test behaviors abstraction for secrets  https://review.openstack.org/21341415:55
*** lisaclark1 has quit IRC16:00
*** everjeje has joined #openstack-barbican16:07
*** lisaclark1 has joined #openstack-barbican16:11
*** vivek-ebay has quit IRC16:19
*** spotz is now known as spotz_zzz16:21
*** spotz_zzz is now known as spotz16:21
*** DTadrzak has quit IRC16:34
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Remove test behaviors abstraction for secrets  https://review.openstack.org/21341416:36
*** xaeth is now known as xaeth_afk16:37
*** gyee has joined #openstack-barbican16:43
*** insequent has joined #openstack-barbican16:50
*** vivek-ebay has joined #openstack-barbican16:50
*** rellerreller has quit IRC17:03
*** xaeth_afk is now known as xaeth17:08
*** dave-mccowan has left #openstack-barbican17:08
*** shohel has quit IRC17:10
*** lisaclark1 has quit IRC17:11
*** jaosorior has quit IRC17:15
*** lisaclark1 has joined #openstack-barbican17:20
*** xaeth is now known as xaeth_afk17:28
*** xaeth_afk is now known as xaeth17:35
*** ccneill has quit IRC17:40
*** ccneill has joined #openstack-barbican17:49
*** vivek-ebay has quit IRC17:51
*** vivek-ebay has joined #openstack-barbican18:04
openstackgerritArun Kant proposed openstack/python-barbicanclient: Part 1: Adding ACL support for Client API.  https://review.openstack.org/20669918:06
openstackgerritArun Kant proposed openstack/python-barbicanclient: Part 2: Adding ACL support for CLI commands and docs  https://review.openstack.org/20834318:11
openstackgerritArun Kant proposed openstack/python-barbicanclient: Part 3: Adding ACL functional tests.  https://review.openstack.org/20834418:12
*** vivek-ebay has quit IRC18:15
*** vivek-ebay has joined #openstack-barbican18:16
*** lisaclark1 has quit IRC18:19
*** lisaclark1 has joined #openstack-barbican18:21
*** silos has joined #openstack-barbican18:41
*** lisaclark1 has quit IRC18:56
*** vivek-ebay has quit IRC18:59
*** vivek-ebay has joined #openstack-barbican19:02
*** pglbutt has joined #openstack-barbican19:02
*** pglbutt has quit IRC19:02
*** pglass has quit IRC19:04
*** rm_work|away is now known as rm_work19:23
rm_workdave-mcc_: sorry, actually there is no part of my change that REALLY needs to use nose19:23
rm_workdave-mcc_: it just happens that is how i am testing currently19:23
rm_workdave-mcc_: really my only change to the gate is how the devstack plugin is loaded, you should ignore the specific testing I'm doing19:23
rm_workso probably we're solving for something that isn't a real problem19:23
rm_workI can just switch to testing however DOES wor19:24
rm_work*work19:24
dave-mcc_rm_work no worries.  tox uses nose, and that's what most people use locally.  thanks for catching that... it would have come up eventually.  i was able to recreate it locally about 50% of the time with "tox -e functional".19:25
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Remove test behaviors abstraction for orders  https://review.openstack.org/21341919:25
rm_workdave-mcc_: ok, so you fixed it with that patch?19:25
rm_workis that the same patch or a new one? can I just rebase and it'll work?19:26
rm_workah it's not on the chain with the other fix19:26
rm_workso I can't depend on both :/19:26
rm_workdave-mcc_: i am in a meeting though, will be back later -- but could you make them be in a chain so i can depend on both?19:27
dave-mcc_there are two patch that fix two different intermittent problems.  i can add the dependency.19:27
*** rm_work is now known as rm_work|away19:27
silosredrobot: ping19:32
openstackgerritDave McCowan proposed openstack/barbican: Cleanup database after functional test runs  https://review.openstack.org/21625519:37
dave-mcc_rm_work done.  i made your CR's parent the child of the other CR, so you should be good to go.19:39
*** gyee has quit IRC19:42
*** insequent has quit IRC19:47
*** pglass has joined #openstack-barbican19:48
*** pglbutt has joined #openstack-barbican19:50
*** pglass has quit IRC19:53
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Remove test behaviors abstraction for orders  https://review.openstack.org/21341919:57
*** vivek-ebay has quit IRC20:00
*** vivek-ebay has joined #openstack-barbican20:00
*** pglbutt has quit IRC20:04
*** pglass has joined #openstack-barbican20:09
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Remove test behaviors abstraction for containers  https://review.openstack.org/21342020:09
*** pglbutt has joined #openstack-barbican20:10
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Remove test behaviors abstraction for orders  https://review.openstack.org/21341920:12
*** pglass has quit IRC20:13
*** lisaclark1 has joined #openstack-barbican20:15
*** vivek-ebay has quit IRC20:19
*** lisaclark1 has quit IRC20:36
*** lisaclark1 has joined #openstack-barbican20:37
*** gyee has joined #openstack-barbican20:41
*** ngupta__ has joined #openstack-barbican20:44
*** kfarr has joined #openstack-barbican20:45
kfarrdave-mcc_ I was going to +W https://review.openstack.org/#/c/212967/, did you need someone else to look it first?20:46
dave-mcc_kfarr thanks kaitlin, that one is good to go.20:56
*** vivek-ebay has joined #openstack-barbican21:20
openstackgerritFernando Diaz proposed openstack/python-barbicanclient: Remove test behaviors abstraction for containers  https://review.openstack.org/21342021:21
*** dave-mcc_ has quit IRC21:21
*** lisaclark1 has quit IRC21:24
*** vivek-ebay has quit IRC21:24
openstackgerritMerged openstack/barbican: Add Quota Enforcement API  https://review.openstack.org/21296721:28
openstackgerritMerged openstack/barbican: Updated from global requirements  https://review.openstack.org/21656021:37
*** silos has left #openstack-barbican21:40
*** ngupta__ has quit IRC21:44
*** diazjf has left #openstack-barbican21:49
*** edtubill has quit IRC21:50
*** kfarr has quit IRC21:56
*** insequent has joined #openstack-barbican22:06
*** vivek-ebay has joined #openstack-barbican22:07
*** alee is now known as alee_afk22:17
*** xaeth is now known as xaeth_afk22:18
*** a2hill has joined #openstack-barbican22:21
*** alee_afk has quit IRC22:22
*** a2hill has quit IRC22:25
*** crc32| has joined #openstack-barbican22:26
*** crc32- has joined #openstack-barbican22:27
*** crc32- has quit IRC22:27
*** jorgem has joined #openstack-barbican22:27
*** pglbutt has quit IRC22:32
*** jlhinson has quit IRC22:35
*** spotz is now known as spotz_zzz22:52
*** darrenmoffat has quit IRC23:04
*** crc32- has joined #openstack-barbican23:04
*** darrenmoffat has joined #openstack-barbican23:04
*** crc32- has quit IRC23:05
*** kebray has joined #openstack-barbican23:06
*** fnaval| has joined #openstack-barbican23:06
*** kebray has quit IRC23:07
*** fnaval| is now known as fnaval23:09
*** kebray has joined #openstack-barbican23:18
*** dimtruck is now known as zz_dimtruck23:21
*** chadlung has quit IRC23:22
*** alee_afk has joined #openstack-barbican23:25
*** bitblt has joined #openstack-barbican23:28
*** bitblt has quit IRC23:28
*** SheenaG has quit IRC23:32
*** SheenaG has joined #openstack-barbican23:32
*** dave-mccowan has joined #openstack-barbican23:49

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