| *** chlong has joined #openstack-dns | 00:06 | |
| *** ryanpetrello has joined #openstack-dns | 00:35 | |
| *** ryanpetrello has quit IRC | 00:39 | |
| *** Stanley00 has joined #openstack-dns | 00:41 | |
| *** ryanpetrello has joined #openstack-dns | 01:00 | |
| *** ryanpetrello has quit IRC | 01:07 | |
| *** ryanpetrello has joined #openstack-dns | 01:09 | |
| *** rmoe has quit IRC | 01:17 | |
| *** ryanpetrello has quit IRC | 01:18 | |
| *** GonZo2000 has quit IRC | 01:32 | |
| *** ryanpetrello has joined #openstack-dns | 01:32 | |
| *** rmoe has joined #openstack-dns | 01:33 | |
| *** stanzgy has joined #openstack-dns | 01:35 | |
| *** ryanpetrello has quit IRC | 01:37 | |
| *** EricGonczer_ has joined #openstack-dns | 01:45 | |
| *** csoukup has quit IRC | 01:50 | |
| *** EricGonczer_ has quit IRC | 02:01 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/designate: Updated from global requirements https://review.openstack.org/147025 | 02:13 |
|---|---|---|
| *** burper has joined #openstack-dns | 02:32 | |
| *** burper has left #openstack-dns | 02:32 | |
| *** ryanpetrello has joined #openstack-dns | 02:58 | |
| *** vinod1 has joined #openstack-dns | 03:23 | |
| *** ryanpetrello has quit IRC | 03:54 | |
| *** GonZo2000 has joined #openstack-dns | 03:55 | |
| *** GonZo2000 has joined #openstack-dns | 03:55 | |
| *** chlong has quit IRC | 04:06 | |
| *** chlong has joined #openstack-dns | 04:11 | |
| *** chlong_ has joined #openstack-dns | 04:13 | |
| *** chlong has quit IRC | 04:17 | |
| *** richm has quit IRC | 04:17 | |
| *** chlong_ has quit IRC | 04:18 | |
| *** chlong has joined #openstack-dns | 04:24 | |
| *** chlong_ has joined #openstack-dns | 04:26 | |
| *** chlong has quit IRC | 04:27 | |
| *** chlong__ has joined #openstack-dns | 04:28 | |
| *** chlong_ has quit IRC | 04:29 | |
| *** chlong_ has joined #openstack-dns | 04:29 | |
| *** chlong_ has quit IRC | 04:32 | |
| *** chlong has joined #openstack-dns | 04:32 | |
| *** chlong__ has quit IRC | 04:33 | |
| *** chlong has quit IRC | 04:34 | |
| *** chlong has joined #openstack-dns | 05:00 | |
| *** chlong_ has joined #openstack-dns | 05:01 | |
| *** chlong has quit IRC | 05:04 | |
| *** chlong__ has joined #openstack-dns | 05:06 | |
| *** chlong_ has quit IRC | 05:09 | |
| *** GonZo2000 has quit IRC | 05:11 | |
| *** chlong has joined #openstack-dns | 05:12 | |
| *** chlong__ has quit IRC | 05:12 | |
| *** chlong has quit IRC | 05:12 | |
| *** chlong has joined #openstack-dns | 05:25 | |
| *** chlong has quit IRC | 05:31 | |
| *** chlong has joined #openstack-dns | 05:31 | |
| *** chlong has quit IRC | 05:34 | |
| *** chlong has joined #openstack-dns | 05:40 | |
| *** chlong_ has joined #openstack-dns | 05:41 | |
| *** chlong has quit IRC | 05:45 | |
| *** nihilifer has joined #openstack-dns | 06:26 | |
| *** chlong_ has quit IRC | 07:39 | |
| openstackgerrit | stanzgy proposed openstack/designate: Update oslo.db version to 1.4.0 or higher https://review.openstack.org/147407 | 07:43 |
| *** kbyrne has joined #openstack-dns | 08:57 | |
| *** k4n0 has joined #openstack-dns | 09:29 | |
| *** jordanP has joined #openstack-dns | 09:33 | |
| *** k4n0 has quit IRC | 09:36 | |
| openstackgerrit | Endre Karlson proposed openstack/designate: MDNS part of Secondary zones https://review.openstack.org/133683 | 10:05 |
| *** nihilifer has quit IRC | 10:18 | |
| *** Stanley00 has quit IRC | 10:22 | |
| *** nihilifer has joined #openstack-dns | 10:29 | |
| *** nihilifer has quit IRC | 10:36 | |
| *** nihilifer has joined #openstack-dns | 10:38 | |
| *** stanzgy has quit IRC | 10:53 | |
| *** untriaged-bot has joined #openstack-dns | 11:02 | |
| untriaged-bot | Untriaged bugs so far: | 11:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1289444 | 11:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1408202 | 11:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1409629 | 11:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1411132 | 11:02 |
| *** untriaged-bot has quit IRC | 11:02 | |
| *** GonZo2000 has joined #openstack-dns | 11:25 | |
| *** GonZo2000 has joined #openstack-dns | 11:25 | |
| *** GonZo2000 has quit IRC | 11:42 | |
| *** chlong_ has joined #openstack-dns | 12:08 | |
| *** nihilifer has quit IRC | 12:10 | |
| *** nihilifer has joined #openstack-dns | 12:13 | |
| *** EricGonczer_ has joined #openstack-dns | 12:16 | |
| *** EricGonczer_ has quit IRC | 12:18 | |
| *** mwagner_lap has quit IRC | 12:53 | |
| *** mwagner_lap has joined #openstack-dns | 13:38 | |
| *** chlong_ has quit IRC | 13:56 | |
| *** EricGonczer_ has joined #openstack-dns | 14:16 | |
| *** betsy has quit IRC | 14:16 | |
| openstackgerrit | Merged openstack/designate: Update README.md doc build instructions https://review.openstack.org/147037 | 14:21 |
| *** rjrjr_ has joined #openstack-dns | 14:26 | |
| *** csoukup has joined #openstack-dns | 14:30 | |
| *** nihilifer has quit IRC | 14:47 | |
| *** vinod1 has joined #openstack-dns | 14:50 | |
| *** richm has joined #openstack-dns | 14:52 | |
| *** vinod2 has joined #openstack-dns | 14:57 | |
| *** vinod1 has quit IRC | 14:57 | |
| *** paul_glass has joined #openstack-dns | 14:58 | |
| *** jmcbride has joined #openstack-dns | 15:01 | |
| *** timsim has joined #openstack-dns | 15:03 | |
| *** jmcbride has quit IRC | 15:04 | |
| *** jmcbride1 has joined #openstack-dns | 15:04 | |
| *** jmcbride1 has quit IRC | 15:04 | |
| *** jmcbride has joined #openstack-dns | 15:04 | |
| *** betsy has joined #openstack-dns | 15:12 | |
| *** jmcbride has quit IRC | 15:25 | |
| *** nkinder is now known as nkinder_away | 15:40 | |
| *** rjrjr has quit IRC | 15:40 | |
| *** rjrjr_ has quit IRC | 15:41 | |
| *** ryanpetrello has joined #openstack-dns | 15:50 | |
| openstackgerrit | Ron Rickard proposed openstack/designate: WIP Add Unit Tests for Pool Manager https://review.openstack.org/147241 | 15:58 |
| openstackgerrit | Ron Rickard proposed openstack/designate: Ensure Pool Manager Works for Multiple Backend Servers https://review.openstack.org/146246 | 15:58 |
| *** rjrjr has joined #openstack-dns | 16:01 | |
| rjrjr | i ran into 2 problems with unit testing Pool Manager. | 16:02 |
| rjrjr | 1) the designate database is not clean and contains quite a few domains | 16:02 |
| rjrjr | 2) i am unable to get test_delete_domain_backend_one_failure working | 16:04 |
| rjrjr | as for the second problem, I've done some investigating and don't see any reason why this method does not work. | 16:04 |
| rjrjr | all appearances are the FakeBackend is not being called. i've tested this manually, and there is nothing wrong with the Pool Manager code, as it works as the test is setup. | 16:06 |
| *** jmcbride has joined #openstack-dns | 16:06 | |
| rjrjr | as for the first problem, do we have a method in the Database fixture to cleanup after a suite of tests are done? | 16:07 |
| rjrjr | i've been looking at these problems all night and could use some help. | 16:07 |
| *** EricGonczer_ has quit IRC | 16:08 | |
| timsim | Why does the database having things in it afffect the tests? I'm not saying it should, just curious. | 16:08 |
| rjrjr | i don't know. the nearest i can tell, these entries are leftovers from other tests. | 16:09 |
| rjrjr | this is pretty counter to how i would expect tests to behave (cleaning up after themselves.) | 16:09 |
| timsim | I'm leaning toward agreement, but you don't know why it breaks your tests? | 16:10 |
| rjrjr | i do know why it is breaking the tests. i'm trying to figure out how to cleanout the database so the tests don't break. | 16:11 |
| rjrjr | i'm expecting a clean database so i can count statuses, number of calls to notify_domain_changed and poll_for_serial_number. | 16:11 |
| rjrjr | needless to say, with a populated database, periodic_sync/periodic_recovery are running and mucking up the works. | 16:12 |
| timsim | Maybe the Pool Manager will have to handle not having a clean database to start? If someone had a deployment already and they wanted to integrate Server Pools they couldn't clean out heir database. | 16:13 |
| rjrjr | a clean database would go a long way to fixing that. | 16:13 |
| timsim | But I see your point. | 16:13 |
| rjrjr | sure, but that has nothing to do with unit tests. | 16:13 |
| timsim | Fair enough. | 16:13 |
| *** EricGonczer_ has joined #openstack-dns | 16:13 | |
| rjrjr | i checked in the code if someone wants to help out and take a look. | 16:14 |
| rjrjr | tox -e py27 -- designate.tests.test_pool_manager.test_service will show just problem #2. | 16:15 |
| rjrjr | tox -e py27 -r will show both problem #1 and #2. | 16:15 |
| rjrjr | when i run the pool manager tests without all the other tests, i just have the one problem, and don't have the issue with a prepopulated database. | 16:16 |
| paul_glass | just glancing at e.g. https://github.com/openstack/designate/blob/master/designate/tests/test_api/test_v2/test_zones.py#L29, I don't see any tear down function or anything that would cleanup those zones. | 16:18 |
| paul_glass | (the super class has tearDown, but it doesn't delete any zones) | 16:19 |
| Kiall | rjrjr / timsim - the DB should be recreated from scratch after each run.. should.. it's possible a bug a crept in? | 16:22 |
| timsim | Kiall: After each run of all the tests or after each test suite? I think rjrjr's issue is that if the PM tests run after other ones there is stuff in there from past ones, right rjrjr? | 16:23 |
| timsim | It seems like it wouldn't be too hard to create a cleanup method in https://github.com/openstack/designate/blob/master/designate/tests/__init__.py and call it in the teardowns to delete all the things. | 16:23 |
| rjrjr | i'm seeing the issue when the Pool Manager tests are run with the other tests. if i run the Pool Manager tests without the other tests, I don't see the issue with data in the database. | 16:24 |
| Kiall | In a meeting with Thierry - brb | 16:25 |
| timsim | Or just delete/recreate database after each test suite. | 16:25 |
| rjrjr | since the database exists for *all* tests, every component is suspect here. API, central, storage, even pool_manager_cache. | 16:25 |
| Kiall | So - the DB fixture creates 1 DB using the migrations per test runner (usually 1 per CPU core), then copies the "golden" DB once per test and uses that | 16:28 |
| paul_glass | I was just looking at that: https://github.com/openstack/designate/blob/master/designate/tests/__init__.py#L98 | 16:29 |
| Kiall | If you're seeing otherwise, then the PMC fixture is broke, or the PM tests are doing something odd :/ | 16:29 |
| Kiall | Looking at the PM tests you have now to check :) | 16:30 |
| rjrjr | i'm looking. i just shutdown my local mysql database to see if that is getting in the picture somehow. | 16:30 |
| *** jmcbride has quit IRC | 16:30 | |
| rjrjr | but that database *is* clean. | 16:30 |
| Kiall | It should be using the SQLite DB | 16:32 |
| Kiall | (was actually talking to devananda in the bar last night, turns out we can use MySQL during the gate runs instead...) | 16:32 |
| rjrjr | i agree. i have a bazillion files in /tmp name designate-*.sqlite | 16:32 |
| *** jordanP has quit IRC | 16:33 | |
| Kiall | Looking at your test code, I can't see anything obvious that might cause it - e.g. setup in __init__ rather than setUp, or global vars with the service in it eyc | 16:34 |
| Kiall | etc | 16:34 |
| Kiall | Also not seeing any obvious issues in the PM service itself.. | 16:35 |
| Kiall | or the PMC code .. grr.. Was hoping for somehting obvious -_- | 16:35 |
| rjrjr | with mysql database shutdown, i'm still seeing the issue. it does appear to be leftover from previous tests. | 16:36 |
| Kiall | And - You're sure all the tests are passing? I've seen cases that look like leftover crud when a test fails, it's rare, but happens in some circumstances I haven't figured out.. | 16:38 |
| rjrjr | example domains: 2.168.192.in-addr.arpa, example.com, example.net, www.example.com, etc. | 16:38 |
| rjrjr | just need to figure out which tests are working with these domains. | 16:38 |
| Kiall | Ah - That could be it.. http://paste.openstack.org/show/158139/ | 16:39 |
| Kiall | First failing test is usually it when I've seen that weirdness in the past | 16:39 |
| Kiall | past | 16:39 |
| rjrjr | i'm wondering if the problem is just mine? | 16:41 |
| rjrjr | kiall, what do you get if you run 'tox -e py27 -- designate.tests.test_pool_manager.test_service' | 16:41 |
| Kiall | Just waiting for pip to download stuff | 16:42 |
| Kiall | Think the office proxy is killing me -_- | 16:42 |
| rjrjr | just the pool manager tests by themselves? you should only see one test fail, test_delete_domain_backend_one_failure. | 16:42 |
| Kiall | Get's like half the stuff from pypi then stalls | 16:43 |
| *** jmcbride has joined #openstack-dns | 16:44 | |
| Kiall | tox -e py27 -- designate.tests.test_pool_manager.test_service.PoolManagerServiceTest.test_delete_domain <-- that fails too.. Runs just 3 tests | 16:44 |
| Kiall | tox -e py27 -- designate.tests.test_pool_manager.test_service.PoolManagerServiceTest.test_delete_domain$ passes however (run's just 1 test) | 16:44 |
| Kiall | tox -e py27 -- designate.tests.test_pool_manager.test_service.PoolManagerServiceTest.test_delete_domain_backend_one_failure <-- Fails, runs a single test | 16:45 |
| Kiall | tox -e py27 -- designate.tests.test_pool_manager.test_service <-- Now passes, after deleting the test_delete_domain_backend_one_failure test. | 16:46 |
| rjrjr | that is the broken test. | 16:47 |
| Kiall | Seems to | 16:47 |
| rjrjr | makes no sense to me why it is broken either. | 16:47 |
| Kiall | be | 16:47 |
| *** jmcbride has quit IRC | 16:47 | |
| rjrjr | manually running the code, everything makes sense. | 16:47 |
| rjrjr | manually executing the test (designate domain-delete XXX) works. | 16:47 |
| Kiall | Yea, I think the issue is somewhere in that test | 16:48 |
| rjrjr | let me take out the one test and run the entire suite to see if it makes a diff. | 16:48 |
| Kiall | So, this test has somehow triggered the issue I mentioned with the DB fixtures not cleaning up.. It's painfully non-obvious what's happening.. I've hit it 2 or 3 times before, but never figured it out... | 16:49 |
| *** jmcbride has joined #openstack-dns | 16:50 | |
| *** jmcbride has quit IRC | 16:50 | |
| rjrjr | running 'tox -e py27 -r' now to see if that test is causing the issue. | 16:51 |
| rjrjr | i hope so. <fingers crossed> | 16:51 |
| Kiall | :) | 16:51 |
| Kiall | I think the issue might be | 16:52 |
| Kiall | mock_update_status.assert_called_once_with(261 | 16:52 |
| Kiall | self.admin_context, domain.id, 'ERROR', domain.serial) | 16:52 |
| Kiall | The update status method will be called twice, so that'll bail | 16:52 |
| Kiall | What you likely want is.. | 16:53 |
| Kiall | assert_has_calls([(self.admin_context, domain.id, 'ERROR', domain.serial), (self.admin_context, domain.id, 'SUCCESS', domain.serial),] | 16:53 |
| Kiall | and assert that call_count == 2 | 16:54 |
| Kiall | brb | 16:54 |
| rjrjr | i'm still seeing a bunch of failures with 'tox -e py27 -r'. again, from looking at the output, it looks like the database has domains in it. this is without the broken test in PM. | 16:55 |
| rjrjr | man, i hope my env isn't just messed up. can someone else confirm? | 16:55 |
| *** rmoe has quit IRC | 16:59 | |
| timsim | rjrjr: what do you need me to do? | 17:00 |
| rjrjr | can you remove the bad test and run a 'tox -e py27 -r'? does it pass or fail? | 17:00 |
| Kiall | Running that now.. | 17:01 |
| rjrjr | cool. | 17:02 |
| *** jmcbride has joined #openstack-dns | 17:02 | |
| *** untriaged-bot has joined #openstack-dns | 17:02 | |
| untriaged-bot | Untriaged bugs so far: | 17:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1289444 | 17:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1408202 | 17:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1409629 | 17:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1411132 | 17:02 |
| *** untriaged-bot has quit IRC | 17:02 | |
| Kiall | fails... might be another bad test? looking now.. | 17:03 |
| rjrjr | but the unit tests for PM isolated works? | 17:03 |
| Kiall | Yep, not sure why yet | 17:04 |
| rjrjr | okay, so you are in the same state as me. when i look at the output, i see domains that i do not create in the PM tests. the only domain in the PM tests is example.org. | 17:04 |
| rjrjr | grep for 'Executing' to see the domains. | 17:05 |
| Kiall | Yea, the bug is a little weird - when one test fails, it can cause some really weird results... | 17:05 |
| Kiall | I suspect we still have a dodgy test in the new tests | 17:06 |
| rjrjr | but the PM unit tests succeed on their own. | 17:06 |
| rjrjr | you are probably right. i guess i could add one PM test in at a time until i see the failure. | 17:07 |
| rjrjr | is there a decorator to ignore tests? | 17:07 |
| *** jmcbride has quit IRC | 17:07 | |
| Kiall | Ehh - I think there is? But that could have been the old test suite/runner | 17:08 |
| rjrjr | i'm thinking of something like JUnits @Ignore annotation. | 17:09 |
| rjrjr | i'll do it the hard way. comment out the tests and add them back one at a time. | 17:09 |
| *** mwagner_lap has quit IRC | 17:10 | |
| Kiall | Yea - I went to redo the DB fixture a while back, and ran into some weirdness that blocked me (It was like 11pm though at the time, I suspect that had something to do with it ;)) | 17:11 |
| betsy | rjrjr: You can activate testr by typing ‘source .tox/py27/bin/activate’ | 17:11 |
| betsy | then you can say: testr list-tests | 17:11 |
| betsy | to get all the tests | 17:11 |
| betsy | and testr run <test name> | 17:11 |
| betsy | to try an individual test | 17:11 |
| betsy | All the debug messages are still there | 17:12 |
| rjrjr | betsy: the individual PM tests run fine. it is when the run with the other tests that there is a problem. | 17:12 |
| betsy | rjrjr: Ah | 17:12 |
| rjrjr | that is why i thought the database was sticking around from a previous test. i'm seeing PM try to work with domains that are not in the PM unit tests. it has one domain, example.org. | 17:13 |
| rjrjr | but i'm seeing example.com, example.net, www.example.com, 2.168.192.in-addr.arpa, etc. | 17:13 |
| betsy | hmm | 17:13 |
| rjrjr | the joys of Python I guess... | 17:14 |
| Kiall | I'm 99% sure it's a bug in our DB fixture code that I haven't managed to trace yet :'( | 17:14 |
| rjrjr | the problem or why it is hard to see the root of the problem? | 17:15 |
| Kiall | Something triggers a failure to reset the DB I think, never managed to figure out what though :( | 17:16 |
| *** rmoe has joined #openstack-dns | 17:17 | |
| rjrjr | okay, this might take a while to find the test causing this. going to run 'tox -e py27 -r' enabling one test at a time until a test causes the failure. | 17:18 |
| rjrjr | what's a few more hours? :) | 17:18 |
| *** vinod2 has quit IRC | 17:19 | |
| Kiall | Broken tests with these commented start working: | 17:19 |
| Kiall | # mock_notify_zone_changed.assert_called_once_with( | 17:19 |
| Kiall | # self.admin_context, domain, '10.0.0.2:53', 30, 2, 3, 0) | 17:19 |
| Kiall | # mock_poll_for_serial_number.assert_called_once_with( | 17:19 |
| Kiall | # self.admin_context, domain, '10.0.0.2:53', 30, 2, 3, 1) | 17:19 |
| Kiall | Also - no need to kep doing to the -r on the tox calls, unless requirements have changed, just adds a few mins to each run | 17:20 |
| rjrjr | which test? you are able to run *all* the tests with just those lines commented? or just the PM tests? | 17:21 |
| Kiall | Slowing re-enabling the lot, most are woking without that - or with call_count assertions changed | 17:21 |
| rjrjr | if you look at the code, they should only be executed once. that is troubling they are being executed more than once. | 17:22 |
| rjrjr | those asserts and poll_counts are the crux of the unit tests. :) | 17:22 |
| Kiall | Yea.. Not sure what's happening yet | 17:23 |
| rjrjr | gotta say, up until i ran all the tests, i was digging mock. now, not so much. :( | 17:24 |
| Kiall | lol | 17:24 |
| Kiall | I'm wondering if this could be part of the issue? | 17:24 |
| Kiall | # Reset the mock call attributes. | 17:24 |
| Kiall | mock_notify_zone_changed.reset_mock() | 17:24 |
| Kiall | usually, you would use the contextmanager syntax instead of that... But with 3 mocks in some places that can be painful ;) | 17:25 |
| rjrjr | i added that to exclude the previous calls in the test. i just want to count the calls in the method being unit tested. | 17:25 |
| rjrjr | i started with the 'with patch.object ...' and changed it probably 4-5 hours ago while trying to fix the problem with the one unit test. | 17:26 |
| Kiall | Ah.. Okay, problem was pre-existing that change then. | 17:27 |
| rjrjr | the decorator has the benefit of using a new mock object. the context manager does not and reuses the mock object. (at least that is the way i understand what i read in the documentation.) | 17:27 |
| rjrjr | i've been working on this for about 8 hours. so i've tried lots of things. :) | 17:27 |
| rjrjr | didn't sleep last night. that will catch up to me sometime today. | 17:27 |
| Kiall | LOL | 17:28 |
| Kiall | Go get some sleep | 17:28 |
| rjrjr | i have to get this working. lots of pressure on me right now to get this working. | 17:28 |
| rjrjr | the first test i uncomment failed. it looks like the database has more domains than what i created. i'll paste the output. just a second. | 17:30 |
| rjrjr | http://paste.openstack.org/show/158150/ | 17:32 |
| rjrjr | so, it is trying to create example.com. which has nothing to do with the Pool Manager tests. this is too weird. | 17:33 |
| rjrjr | Kiall are you seeing other domains in our output too? | 17:34 |
| rjrjr | our = your | 17:34 |
| Kiall | Yep, not sure what's going on.. | 17:35 |
| Kiall | brb | 17:35 |
| rjrjr | it could be the pool manager cache tests. i didn't have nearly the background then as i do now on Python unit tests. | 17:36 |
| rjrjr | i 'borrowed' code we use for storage for those tests. | 17:36 |
| rjrjr | i'm going to disable the pool manager cache tests and see if that fixes the issue. | 17:50 |
| rjrjr | okay, even if i remove the pool manager cache tests, the same problem exists. | 17:55 |
| *** kbyrne has quit IRC | 18:07 | |
| *** kbyrne has joined #openstack-dns | 18:09 | |
| rjrjr | i see where 192.168.2 is defined. it is in the fake network_api. | 18:12 |
| *** vinod1 has joined #openstack-dns | 18:22 | |
| Kiall | rjrjr: sorry, on a call right now.. I'll be back in 30 mins... | 18:27 |
| Kiall | TZ changes suck - keep realizing at the last second I have a meeting -_- | 18:28 |
| rjrjr | np. i'm continuing my investigation of the code. | 18:33 |
| *** GonZo2000 has joined #openstack-dns | 18:55 | |
| *** GonZo2000 has joined #openstack-dns | 18:55 | |
| *** timsim_ has joined #openstack-dns | 19:01 | |
| mugsie | so, whos here? | 19:02 |
| Kiall | o/ | 19:02 |
| timsim_ | o/ | 19:02 |
| vinod1 | o/ | 19:03 |
| mugsie | https://plus.google.com/hangouts/_/g2gweiakazvdcf57hsj5j5a3aua | 19:03 |
| Kiall | vinod1 / betsy / rjrjr about? | 19:03 |
| betsy | yep | 19:03 |
| Kiall | and https://etherpad.openstack.org/p/designate-testing-sprint | 19:04 |
| *** timsim has quit IRC | 19:04 | |
| rjrjr | o/ | 19:06 |
| mugsie | rjrjr: we are the hagnout ^ | 19:08 |
| mugsie | hangout* | 19:08 |
| mugsie | we are just dividing up the work | 19:10 |
| mugsie | at the bottom of the etherpad ^ we are each claiming bits of the code, if you have time for the sprint | 19:11 |
| Kiall | Just dropped off the hangout BTW | 19:12 |
| betsy | Should we open a bug for the ones we’re working on and do a branch based on that bug? | 19:13 |
| betsy | Or what branch should we use? | 19:13 |
| rjrjr | so, no hangout? | 19:13 |
| Kiall | rjrjr: planned for the HO just to be a quick intro, then move to IRC.. | 19:14 |
| Kiall | betsy: Good Q - Yes, Let's file bugs so it's tracked for k2 :) | 19:14 |
| Kiall | re branch - probably best to branch off master for when other peoples changes start merging in.. | 19:15 |
| rjrjr | gotcha. i am past the problem withe the one test. i'm now looking at why the pool manager unit tests do not work with all the tests. | 19:15 |
| Kiall | rjrjr: cool... I'm a tad stuped, I think adding just 1 test back in at a time, and running with all tests until it fails is the only way to identify your issue :( | 19:16 |
| rjrjr | i'm pretty sure the problem is NOT with PM> | 19:17 |
| Kiall | :/ | 19:17 |
| rjrjr | the PM unit tests work fine by themselves. they all pass now. when i try running them with the other tests, they fail. | 19:17 |
| Kiall | ekarlso-: about BTW? | 19:17 |
| Kiall | rjrjr: maybe something else is storing global state and causing the crud? But, I'd expect things like the central tests to fail if there were extra domains - e.g. list domains test checks the # of domains after creating the same # etc | 19:18 |
| rjrjr | they are definitely picking up crud. maybe another test is silently failing, causing this problem. | 19:20 |
| Kiall | Possibly, but I'm not sure how that could happen without a being a bug in testr or oslo.test or something | 19:20 |
| Kiall | rjrjr: try adding something to list all domains in the DB into the base test's setUp method, at the end? | 19:25 |
| Kiall | If it ever has a domain, it's broke | 19:25 |
| rjrjr | will do. | 19:25 |
| Kiall | lol - discoverd a bug in the PDNS backend already -_- | 19:26 |
| *** jmcbride has joined #openstack-dns | 19:35 | |
| ekarlso- | Kiall: ... | 19:36 |
| ekarlso- | Kiall: ? | 19:36 |
| Kiall | I know it's late @ home, but we're sarting the unittest sprint if you want to join ;) | 19:36 |
| ekarlso- | awww fack :p | 19:37 |
| ekarlso- | hangout or ? | 19:37 |
| * ekarlso- prods Kiall with a lightning stick | 19:38 | |
| Kiall | ekarlso-: We had a 5 min HO at the start, IRC now ;) | 19:38 |
| Kiall | https://etherpad.openstack.org/p/designate-testing-sprint | 19:38 |
| Kiall | Idea is to pick a small area, write the tests, submit and get near-instant review+merge to help keep momentum up ;) | 19:39 |
| Kiall | List of "claimed" areas at the bottom, list of suggested areas towards the middle.. | 19:39 |
| Kiall | "Edge cases / branches within SQL Storage Driver" tis still open.. I'll also have the barebones for pDNS tests up in 20 mins, which should give a framework for the others | 19:41 |
| ekarlso- | Kiall: meh, but unit tests for dyn f ex is bork due to the also_notify stuffs | 19:43 |
| Kiall | Yep, there are other areas on the list | 19:43 |
| ekarlso- | gimme 30 mins Kiall :p | 19:43 |
| ekarlso- | I just got in -,m,- | 19:43 |
| Kiall | (and other areas we've not listed out I'm sure..( | 19:43 |
| ekarlso- | Kiall: suggestions for a fitting area ? | 19:49 |
| Kiall | Ehh - Edge cases / branches within SQL Storage Driver ? | 19:49 |
| ekarlso- | we got an overview of them ? :D | 19:49 |
| openstackgerrit | Ron Rickard proposed openstack/designate: WIP Add Unit Tests for Pool Manager https://review.openstack.org/147241 | 19:49 |
| Kiall | Yep - `tox -e cover` or stolen from another review - http://logs.openstack.org/46/146246/4/check/designate-coverage/95faedb/cover/ | 19:50 |
| Kiall | Red and yellow bits of http://logs.openstack.org/46/146246/4/check/designate-coverage/95faedb/cover/designate_storage_impl_sqlalchemy___init__.html | 19:50 |
| Kiall | rjrjr: just noticed this.. http://paste.openstack.org/show/158155/ | 20:02 |
| Kiall | check the order of the decoarators and the mock passed into the method | 20:02 |
| Kiall | backwards -_- | 20:02 |
| ekarlso- | Kiall: when it says red on http://logs.openstack.org/46/146246/4/check/designate-coverage/95faedb/cover/designate_storage_impl_sqlalchemy___init__.html#n109 f ex that means ? | 20:03 |
| Kiall | It means no test is executing that code - which is right, since betsy's recent change landed and removed "servers" | 20:03 |
| Kiall | That code must be leftover and in need of removal | 20:04 |
| ekarlso- | so what, remove the server methods ? ^ | 20:04 |
| betsy | Kiall: I thought I had removed all the extra code, but I must have missed some | 20:05 |
| Kiall | Yea, Double check, But I think that's dead code in need of removal. | 20:05 |
| ekarlso- | betsy: u covering that or shall I ? :) | 20:07 |
| ekarlso- | already doing it :p | 20:07 |
| betsy | ok | 20:07 |
| Kiall | betsy: Yea, I'm spotting lots of dead code :) | 20:08 |
| ekarlso- | :D | 20:08 |
| ekarlso- | cleanup on aisle 5 | 20:08 |
| betsy | Kiall: Ugh. All from my change or from others, too? | 20:08 |
| Kiall | No, just in general :) | 20:08 |
| Kiall | Including many of my own changes! | 20:08 |
| ekarlso- | Kiall: spring clean allmost :p | 20:08 |
| betsy | Kiall: Oh, good. I thought you meant I had left lots of dead code :P | 20:09 |
| betsy | Good thing we’re doing this then | 20:09 |
| ekarlso- | Kiall: should delete_server stuff be removed from backends ? | 20:09 |
| Kiall | None of the new backends have those as far as I remember, and the old style will possibly need them still.. | 20:10 |
| ekarlso- | Kiall: a reason that PoolBackend has all the methods then ? | 20:11 |
| *** jmcbride has quit IRC | 20:11 | |
| ekarlso- | ah just pas atm | 20:11 |
| Kiall | Yea, to fill the interface | 20:11 |
| Kiall | the PoolBackend and Backend classes need to merge at some point, with the extra stuff removed. | 20:11 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: WIP - PowerDNS Backend Tests https://review.openstack.org/147648 | 20:25 |
| Kiall | Start of PDNS backend tests ^ | 20:25 |
| Kiall | I'm going to skip writing tests for the _create / _delete etc methods - since I believe we can consolidate with the designate.sqlalchemy.base versions of those methods.. there nearly identical | 20:27 |
| rjrjr | Kiall: that is the correct order. bottom up on the decorators. | 20:27 |
| Kiall | Yea, It got me ;) Wasn't sure if it got you too! | 20:28 |
| rjrjr | i don't believe the setUp method is called outside the base testing. | 20:28 |
| ekarlso- | file:///home/karlsone/projects/dnsaas/designate/cover/designate_storage_impl_sqlalchemy___init__.html#n841 < that's dead too Kiall ? | 20:29 |
| rjrjr | i'm printing out the domains in setUp method and even though i have failed tests, i never see the output. | 20:29 |
| Kiall | I can't see your HD ekarlso- :P | 20:29 |
| Kiall | rjrjr: printing? or Logging? I wonder if logging from before the test starts (e.g. setUp) is hidden | 20:29 |
| ekarlso- | LOL | 20:29 |
| rjrjr | i'm using print('... | 20:29 |
| ekarlso- | hah, http://logs.openstack.org/46/146246/4/check/designate-coverage/95faedb/cover/designate_storage_impl_sqlalchemy___init__.html#n896 | 20:30 |
| rjrjr | is logging guaranteed to be output? | 20:30 |
| rjrjr | apparently printing isn't. | 20:30 |
| Kiall | ekarlso-: not sure - I don't think it's dead, likely just untested | 20:31 |
| Kiall | mugsie: would know for sure | 20:31 |
| mugsie | know what? | 20:31 |
| mugsie | oh | 20:31 |
| Kiall | rjrjr: Mmm, It should be captured and included - But try this: | 20:31 |
| Kiall | OS_OS_STDOUT_CAPTURE=0 OS_OS_STDERR_CAPTURE=0 tox -e py27 | 20:32 |
| mugsie | ekarlso-: IT IS NOT DEAD | 20:32 |
| rjrjr | print or LOG? | 20:32 |
| mugsie | -Caps | 20:32 |
| Kiall | That'll prevent the test runner from trying to capture stdout / stderr at all | 20:32 |
| Kiall | 20:32 | |
| rjrjr | i see we define a LOG, but it is never ever used. | 20:32 |
| mugsie | not sure why it is shoing as red thouh# | 20:32 |
| Kiall | 20:43 | |
| Kiall | grr | 20:43 |
| Kiall | wrong window ;) | 20:43 |
| rjrjr | LOL - seeing this on output - Segmentation fault (core dumped) | 20:44 |
| Kiall | Noo, what?? | 20:44 |
| Kiall | That's just .. Noo.. :/ | 20:44 |
| rjrjr | the tests run and look like they did before, but now i'm seeing that in the output stream. | 20:45 |
| rjrjr | with OS_STDOUT_CAPTURE=0 ... | 20:46 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: PowerDNS Backend Tests https://review.openstack.org/147648 | 20:46 |
| *** paul_glass has quit IRC | 20:48 | |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: PowerDNS Backend Tests https://review.openstack.org/147648 | 20:48 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: PowerDNS Backend Tests https://review.openstack.org/147648 | 20:49 |
| Kiall | vinod1 / betsy / timsim_ / ekarlso- / rjrjr - how are you guys getting on? | 20:57 |
| betsy | Kiall: good | 20:58 |
| ekarlso- | Kiall: with my current brain power trying :p | 20:58 |
| Kiall | :D | 20:58 |
| vinod1 | trying to decipher mock methods and figuring out how to make sure methods are called - so far tackling untested mdns service | 20:58 |
| Kiall | ^ review has lots of use of mock, might be worth an eyeball? | 20:59 |
| timsim_ | Alright, started writing tests for https://github.com/openstack/designate/blob/master/designate/objects/base.py#L386 then realized maybe that's not as important because that'll probably be going away soon. Switching to the DesignateObject class :P | 20:59 |
| Kiall | Is it? I didn't think it was.. lol | 21:00 |
| timsim_ | Well maybe not. I thought we wanted to swtich to only Object.attr notation, rather than Dict or List style | 21:00 |
| Kiall | Ah, List i think stays - Dict i think goes | 21:00 |
| timsim_ | Ah, well lovely then :) | 21:01 |
| rjrjr | i'm still staring at a wall. i think something is really, really wrong here, i just cannot figure out what. | 21:01 |
| timsim_ | Oh I get it now. Dumb. | 21:02 |
| ekarlso- | meh Kiall in a try block the else: part does get executed when ? | 21:02 |
| Kiall | in a try/except/else - the else is executed when there is no exception.. | 21:03 |
| rjrjr | it really looks like the database is not being cleaned between tests to me. | 21:03 |
| Kiall | Yea, there is a bug somewhere that means the DB doesn't get cleaned between runs, but every time I've seen it, it's been caused by another broken test :/ | 21:04 |
| *** rjrjr has quit IRC | 21:07 | |
| *** rjrjr has joined #openstack-dns | 21:10 | |
| rjrjr | but not necessarily a bug in a pool manager test. | 21:11 |
| rjrjr | it could be a bug in the tests that run before the pool manager tests. | 21:11 |
| Kiall | Yes, not necessarily - but I've always seen it in the context of me writing/adding a broken test | 21:12 |
| Kiall | Yes, but not necessarily* | 21:12 |
| rjrjr | Kiall: i could use some help here. i'm at a loss on what to do to debug this. half my unit tests fail if i run them by themselves with the other tests and that makes no sense to me. | 21:13 |
| rjrjr | running these unit tests on their own, they pass just fine. | 21:13 |
| rjrjr | (as a group.) | 21:13 |
| rjrjr | i can change what i'm looking at, but then i feel i'm just masking an underlying issue here. | 21:14 |
| Kiall | Humm - Yea, actually... thinking of that it probably means there OK.. We must have some state we're leaving somewhere, or more code than you expect running - did you trace where the example.com. came from in your logs? | 21:14 |
| rjrjr | well, periodic_sync and periodic_recovery are picking up these domains from the database. | 21:16 |
| rjrjr | up these = these up | 21:16 |
| rjrjr | i'm disabling these threads. my hunch, things will work fine. but like i said, i believe i would be masking an underlying issue. | 21:17 |
| timsim_ | Kiall: Want me to just fix this tiny bug inline (https://github.com/openstack/designate/blob/master/designate/objects/base.py#L474) should be self.objects.insert(index, value), yeah? | 21:18 |
| rjrjr | there is a reason to disable these threads for unit testing anyway, but they are what is picking up these stray domains from the database. | 21:18 |
| Kiall | rjrjr: humm, that suggests theres an issue with the PM and clearing all state between tests | 21:18 |
| rjrjr | <sigh> | 21:19 |
| Kiall | But - Yes, they can lead to unpredicible call counts etc | 21:19 |
| Kiall | So - Better to disable those for the generate tests, and test them separatly. | 21:19 |
| rjrjr | my guess central tests are running, queueing up notifications and those notification are not getting cleared or something similar. | 21:19 |
| rjrjr | central is calling pool manager now. | 21:20 |
| Kiall | timsim_: yep, makes sense | 21:20 |
| rjrjr | might explain why i'm seeing a request to create a domain that isn't coming from the PM tests. | 21:20 |
| Kiall | rjrjr: humm - that's very possible actually | 21:20 |
| *** chlong_ has joined #openstack-dns | 21:21 | |
| rjrjr | is there a way to clear the notification. i'll look. | 21:21 |
| rjrjr | so, probably a combination of turning off the timers and clearing the notifications... | 21:22 |
| vinod1 | kiall: for checking the arguments, instead of calling mock.assert_called_once_with, can I get the arguments instead | 21:22 |
| vinod1 | the reason here is that part of the 1st argument could be different each time it is called | 21:23 |
| rjrjr | vinod1, there is a list to get the arguments from the mock. just a sec... | 21:23 |
| vinod1 | so I want to chop of the variable part of the argument and check that the fixed argument matches what I think it should | 21:23 |
| Kiall | vinod1: yes, you can - session_mock.execute.call_args_list | 21:23 |
| Kiall | You get back something like this: | 21:23 |
| Kiall | # [call(<sqlalchemy.sql.dml.Insert object at 0x7f065f2c5510>, {'account': None, 'type': 'SLAVE', 'designate_id': None, 'master': '127.0.1.1:53,127.0.1.2:53', 'name': 'example.com'}), | 21:23 |
| Kiall | # call(<sqlalchemy.sql.selectable.Select at 0x7f065f26a410; Select object>)] is not false | 21:23 |
| Kiall | array of call(arg1, arg2, kwarg1=foo) etc | 21:24 |
| rjrjr | call_args_list | 21:24 |
| Kiall | or call_args to get just the latest call's args | 21:24 |
| rjrjr | Kiall beat me to it! :) | 21:24 |
| rjrjr | there is a helper method called call you can use to create and compare. | 21:25 |
| Kiall | vinod1: example - http://paste.openstack.org/show/oyR1Dl5u4Ma6xiPOxvCz/ | 21:26 |
| rjrjr | i'm convinced my problem is probably notifications. :) that would explain the multiple attempts to create example.com with different IDs. | 21:26 |
| Kiall | rjrjr: checking how that all works again ;) | 21:26 |
| * ekarlso- is sleeping in his chair soon | 21:27 | |
| Kiall | So, we tell oslo.messaging to use a "fake" transport.. But I'm not sure we're clearing it out | 21:27 |
| ekarlso- | uh | 21:29 |
| ekarlso- | u mean notifications is not being cleared ? | 21:29 |
| Kiall | RPC messages | 21:30 |
| Kiall | Rather than actual notifications | 21:30 |
| ekarlso- | what u wanna clear ? | 21:30 |
| rjrjr | ekarlso, ideally, between tests, all RPC messages would be cleared. | 21:31 |
| ekarlso- | rjrjr: it doens't atm ? | 21:33 |
| Kiall | rjrjr: looking into that, https://github.com/openstack/oslo.messaging/blob/master/oslo_messaging/_drivers/impl_fake.py#L91 | 21:33 |
| Kiall | ^ that's the buffer, maybe check that at the start of your tests, and explode if it's not empty? (Not 100% sure how you would access it, but it's available somewhere...) | 21:34 |
| rjrjr | ekarlso: i don't know. just thinking about what could be causing this issue. | 21:34 |
| ekarlso- | if you're evil | 21:35 |
| Kiall | Yea, and it seems reasonable that this could be the cause - lots of cast rather than call's in PM | 21:35 |
| Kiall | fm central -> PM | 21:35 |
| Kiall | from* | 21:35 |
| rjrjr | what about this notifications.clear() in the base test class? | 21:35 |
| ekarlso- | rjrjr: diff things | 21:35 |
| rjrjr | okay. | 21:36 |
| ekarlso- | you can access it I think if you can reach the driver then driver._exchange_manager has all the "exchanges" | 21:36 |
| ekarlso- | once you get to that you can do whatever | 21:36 |
| rjrjr | greek so far, but i'll figure it out. | 21:37 |
| rjrjr | i'm guessing mdns testing completes before pool_manager testing, otherwise i'd be leaving lots of messages for mdns. | 21:37 |
| Kiall | rjrjr: `source .tox/py27/bin/activate && python setup.py testr --testr-args="--isolated" | 21:37 |
| Kiall | run all tests in separate runners | 21:38 |
| Kiall | Meaning no state possible, can you check if everything still passes? | 21:38 |
| rjrjr | sure. just a second. | 21:38 |
| Kiall | I'll take longer than a second ;) | 21:38 |
| Kiall | It'll* | 21:39 |
| ekarlso- | Kiall: all the "missing" things for transfers etc why are they missing ? :| | 21:39 |
| rjrjr | running. | 21:39 |
| Kiall | ekarlso-: missing? They might not be execised by any of the tests.. hence excercising them ;) | 21:39 |
| ekarlso- | oh | 21:40 |
| rjrjr | this is going to take a while... | 21:40 |
| betsy | hmm. It now shows 88% for designate/api/v2/controllers/extensions/counts | 21:40 |
| * ekarlso- thought mugsie added all tests | 21:40 | |
| betsy | The two lines it says are not covered are | 21:40 |
| betsy | return self._view.show(context, request, counts) | 21:40 |
| betsy | and return self._view.show(context, request, counts) | 21:41 |
| mugsie | ekarlso-: I did, but I could have made a mistake ;) | 21:41 |
| betsy | Anybody know why the last return line in each wouldn’t be covered? | 21:41 |
| ekarlso- | :D | 21:41 |
| mugsie | are they definitly not being called? | 21:42 |
| Kiall | guy who's desk I've robbed is standing behind me.. eek | 21:42 |
| mugsie | :D | 21:42 |
| mugsie | move over the wall, there is 2 deskes there atm | 21:42 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/designate: Updated from global requirements https://review.openstack.org/147025 | 21:42 |
| Kiall | Yea, I offered.. He said no ;) | 21:42 |
| betsy | How could the return statement of the functions not be covered? | 21:42 |
| ekarlso- | Kiall: https://bpaste.net/show/ebd375aff5e1 file:///home/karlsone/projects/dnsaas/designate/cover/designate_storage_impl_sqlalchemy___init__.html#n841 why doesn't that hit the resultproxy.fetchone ? | 21:46 |
| ekarlso- | gawd... | 21:47 |
| ekarlso- | http://logs.openstack.org/46/146246/4/check/designate-coverage/95faedb/cover/designate_storage_impl_sqlalchemy___init__.html#n494 | 21:48 |
| openstackgerrit | Vinod Mangalpally proposed openstack/designate: WIP: Add more unit tests to mdns https://review.openstack.org/147669 | 21:48 |
| ekarlso- | same thing.. | 21:48 |
| rjrjr | thanks for talking this through with me. i'm 99% sure lingering RPC messages is going to be my problem. it's going to be about 2 hours before these isolated tests complete. in the meantime, i'll dig into oslo messaging. | 21:48 |
| Kiall | rjrjr: really?? You need a SSD | 21:49 |
| vinod1 | kiall: I see call_args_list as an empty string at line 55 - https://review.openstack.org/#/c/147669/1/designate/tests/test_mdns/test_service.py | 21:49 |
| rjrjr | it is running about 10 tests a minute. we have close to 600 tests. okay, 1 hour. | 21:49 |
| vinod1 | Do you know what I am doing wrong | 21:49 |
| *** chlong_ has quit IRC | 21:50 | |
| ekarlso- | rjrjr: u need this thing | 21:50 |
| ekarlso- | https://www.google.no/search?q=road+runner&espv=2&biw=1745&bih=1013&tbm=isch&imgil=8S-jUTtLcdgZ7M%253A%253BY-C3zYmYfJIgJM%253Bhttp%25253A%25252F%25252Fen.wikipedia.org%25252Fwiki%25252FWile_E._Coyote_and_The_Road_Runner&source=iu&pf=m&fir=8S-jUTtLcdgZ7M%253A%252CY-C3zYmYfJIgJM%252C_&usg=__AtsoIKWwAp6jZ8QrJJLN6IK-Kp0%3D&ved=0CCsQyjc&ei=Gza4VNLKIejnygO56oL4CQ#imgdii=_&imgrc=8S-jUTtLcdgZ7M%253A%3BY-C3zYmY | 21:50 |
| ekarlso- | fJIgJM%3Bhttp%253A%252F%252Fupload.wikimedia.org%252Fwikipedia%252Fen%252Fe%252Fee%252FRoadrunner_looney_tunes.png%3Bhttp%253A%252F%252Fen.wikipedia.org%252Fwiki%252FWile_E._Coyote_and_The_Road_Runner%3B350%3B270 | 21:50 |
| Kiall | ekarlso-: Looks like you're patching the execute method, so: | 21:50 |
| ekarlso- | rofl | 21:50 |
| Kiall | session.execute.fetch_one() would return None | 21:50 |
| Kiall | session.execute.fetch_one() would return None | 21:50 |
| Kiall | rather than | 21:50 |
| Kiall | session.execute().fetch_one() would return None | 21:50 |
| ekarlso- | Kiall: what to patch then ? :p | 21:51 |
| ekarlso- | I was thinking resultproxy but | 21:51 |
| Kiall | with mock.patch.object(self.storage.session, 'execute') as execute_mock: | 21:51 |
| Kiall | rp_mock = mock.Mock() | 21:51 |
| Kiall | rp_mock.fetch_one.return_value = None | 21:51 |
| Kiall | execute_mock.return_value = rp_mock | 21:51 |
| Kiall | something like that I think? | 21:51 |
| Kiall | vinod1: looking | 21:52 |
| Kiall | what happens if you get rid of the , new_callable=mock.MagicMock from the decorator? | 21:53 |
| Kiall | and replace with return_value = XX if we check the return value (Bytes sent) | 21:54 |
| Kiall | if we check the return value.. | 21:54 |
| vinod1 | if we specify the return_value, then we would not be doing the right checks - but maybe we can patch some other function to do that - will check | 21:56 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: PowerDNS Backend Tests https://review.openstack.org/147648 | 22:03 |
| Kiall | vinod1: humm, not sure what you mean? | 22:03 |
| Kiall | OO wait | 22:04 |
| Kiall | self.assertEqual(1, sendto_mock.call_count) | 22:04 |
| Kiall | call_args_list below is an empty list for me 53 | 22:04 |
| Kiall | call_args_list = sendto_mock.execute.call_args_list | 22:04 |
| Kiall | you're asserting that sendto_mock was called once, then checking for args to sendto.execute() - which hasn't been called | 22:04 |
| openstackgerrit | Vinod Mangalpally proposed openstack/designate: WIP: Add more unit tests to mdns https://review.openstack.org/147669 | 22:05 |
| Kiall | Try changing the last line to | 22:05 |
| Kiall | call_args_list = sendto_mock.call_args_list | 22:05 |
| vinod1 | Changed it to something simpler https://review.openstack.org/#/c/147669/2/designate/tests/test_mdns/test_service.py | 22:05 |
| Kiall | I see, that looks like it'll do too :) | 22:06 |
| ekarlso- | Kiall: so for http://logs.openstack.org/46/146246/4/check/designate-coverage/95faedb/cover/designate_storage_impl_sqlalchemy___init__.html#n437 how the crap do we make that get covered ? | 22:07 |
| ekarlso- | I've tried just adding a recordset.records.append(record) but doesn't seem to have nudged it | 22:07 |
| Kiall | Looks like in all the tests, that condition always evaluates to true | 22:08 |
| Kiall | getting that covered would mean getting it to eval to False, i.e. no records attr set at all | 22:08 |
| ekarlso- | :/ | 22:09 |
| vinod1 | Kiall: One more question - I was looking at ways to test _handle_tcp - https://github.com/openstack/designate/blob/master/designate/mdns/service.py#L112 - which has an infinite loop - is there a way to call it in test and have it terminate after one loop | 22:09 |
| Kiall | ehh - that probably needs a little re-work to be testable.. Thinking how would be the best way to handle that.. | 22:11 |
| Kiall | ekarlso-: BTW - that partial probably doesn't warrant the effort to cover (100.00% isn't the goal ;)) | 22:12 |
| ekarlso- | Kiall: what tests you want to cover then ? ;P | 22:12 |
| Kiall | The red is the most important - the yellow are "only one of True/False on this condition has been executed" - and in that case, there's no important difference when that evaluates to False | 22:13 |
| Kiall | vinod1: I think that method needs to be renamed to _handle_tcp_thread, and everything bar the loop moved to _handle_tcp or something? | 22:14 |
| openstackgerrit | Tim Simmons proposed openstack/designate: Add more unit tests to Base Object class https://review.openstack.org/147679 | 22:14 |
| Kiall | the while(True): self._handle_tcp() is perfectly OK to leave untested.. | 22:14 |
| Kiall | (IMO) | 22:14 |
| Kiall | It's too trival a method at that point to need it - 2 statements, 1 builtin, 1 a call to another tested function | 22:15 |
| vinod1 | will do that | 22:15 |
| timsim_ | I'd like to get a few more things in that^ but wanted to get it up there. | 22:16 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: PowerDNS Backend Tests https://review.openstack.org/147648 | 22:16 |
| Kiall | timsim_: yea, I think finishing out what we have - getting reviews up so we can start reviewing them... | 22:16 |
| Kiall | Not perfect - but powerdns backend is up to 74% ish (ignoring the migrations, not really sure why they don't get 100% since we do run them.. Just as part of setUp rather than as part of an actual test) | 22:17 |
| Kiall | And it's done without needing a DB or testing code outside the powerdns backend :P Woo :P | 22:18 |
| Kiall | mugsie / timsim_ / vinod1 / betsy / ekarlso- - let's aim to have reviews up that'll pass by say half past, even if there's still more test's yous | 22:20 |
| Kiall | you would want to add - so we can get them reviewed :) | 22:20 |
| vinod1 | ok | 22:22 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: PowerDNS Backend Tests https://review.openstack.org/147648 | 22:28 |
| openstackgerrit | Vinod Mangalpally proposed openstack/designate: Add more unit tests to mdns service https://review.openstack.org/147669 | 22:30 |
| vinod1 | ok i am now switching to reviewing mode | 22:30 |
| Kiall | Yep - Me too | 22:31 |
| Kiall | timsim_ / betsy / mugsie - got the latest up? | 22:31 |
| openstackgerrit | Tim Simmons proposed openstack/designate: Add more unit tests to Base Object class https://review.openstack.org/147679 | 22:31 |
| timsim_ | Kiall: yep^ | 22:31 |
| Kiall | K - I've added all the review links I know of to the etherpad, if I'm missing any .. let me know.. | 22:32 |
| *** ryanpetrello has quit IRC | 22:32 | |
| Kiall | (or - better yet, add em) | 22:32 |
| betsy | mine’s coming | 22:32 |
| Kiall | Cool :) | 22:33 |
| openstackgerrit | Endre Karlson proposed openstack/designate: Cleanup dead code and add some tests for coverage https://review.openstack.org/147687 | 22:35 |
| Kiall | timsim_: since you managed the highest # of LOC, you're officially our new test guy. | 22:36 |
| timsim_ | lololol | 22:36 |
| timsim_ | I lucked out, that was the easiest thing to test ever :P | 22:36 |
| Kiall | ;) | 22:36 |
| openstackgerrit | Graham Hayes proposed openstack/designate: Add more tests to the central service https://review.openstack.org/147688 | 22:37 |
| ekarlso- | ok, im gonna gonna hit the sack.. | 22:37 |
| openstackgerrit | Betsy Luzader proposed openstack/designate: Add Unit Tests for v2 API Reports https://review.openstack.org/147689 | 22:37 |
| ekarlso- | cya guys | 22:37 |
| betsy | I’m blaming the new (to me) keyboard on my slowness ;) | 22:38 |
| betsy | It’s a PC keyboard I’m using on my mac and driving me nuts | 22:38 |
| Kiall | Lol - Like me using a apple keyboard.. I usually want to smash it ;) | 22:39 |
| *** EricGonczer_ has quit IRC | 22:40 | |
| Kiall | ekarlso-: enjoy :) | 22:41 |
| *** openstack has joined #openstack-dns | 23:02 | |
| *** rediskin has joined #openstack-dns | 23:06 | |
| *** ryanpetrello has joined #openstack-dns | 23:06 | |
| *** rektide_ has joined #openstack-dns | 23:06 | |
| *** rmoe_ has joined #openstack-dns | 23:06 | |
| *** arn_ has joined #openstack-dns | 23:06 | |
| *** mikal has joined #openstack-dns | 23:06 | |
| *** eandersson has joined #openstack-dns | 23:06 | |
| *** mariusv_ has joined #openstack-dns | 23:06 | |
| *** ekarlso has joined #openstack-dns | 23:06 | |
| *** gohko has joined #openstack-dns | 23:06 | |
| *** fouxm has joined #openstack-dns | 23:06 | |
| *** dhp_ has joined #openstack-dns | 23:06 | |
| *** mikal_ has joined #openstack-dns | 23:06 | |
| *** timfreun1 has joined #openstack-dns | 23:06 | |
| *** rjrjr has joined #openstack-dns | 23:06 | |
| *** GonZo2000 has joined #openstack-dns | 23:06 | |
| *** kbyrne has joined #openstack-dns | 23:06 | |
| *** rmoe has joined #openstack-dns | 23:06 | |
| *** csoukup has joined #openstack-dns | 23:06 | |
| *** serverascode has joined #openstack-dns | 23:06 | |
| *** pfreund has joined #openstack-dns | 23:06 | |
| *** arn has joined #openstack-dns | 23:06 | |
| *** openstackgerrit has joined #openstack-dns | 23:06 | |
| *** mugsie has joined #openstack-dns | 23:06 | |
| *** vipul has joined #openstack-dns | 23:06 | |
| *** ChanServ has joined #openstack-dns | 23:06 | |
| *** sendak.freenode.net sets mode: +vo openstackgerrit ChanServ | 23:06 | |
| *** bauruine has quit IRC | 23:06 | |
| *** Kiall has joined #openstack-dns | 23:06 | |
| *** ahu has joined #openstack-dns | 23:06 | |
| *** mikehn has joined #openstack-dns | 23:07 | |
| *** bauruine has joined #openstack-dns | 23:07 | |
| *** mikal_ has quit IRC | 23:07 | |
| *** rmoe has quit IRC | 23:07 | |
| *** openstackgerrit has quit IRC | 23:07 | |
| *** arn has quit IRC | 23:07 | |
| rjrjr | Kiall: the individual tests completed successfully. | 23:10 |
| *** f1ller has joined #openstack-dns | 23:11 | |
| *** pk has joined #openstack-dns | 23:12 | |
| *** simonmcc has joined #openstack-dns | 23:12 | |
| *** nkinder_away has joined #openstack-dns | 23:12 | |
| *** richm has joined #openstack-dns | 23:12 | |
| *** HenryG has joined #openstack-dns | 23:12 | |
| *** openstackgerrit has joined #openstack-dns | 23:13 | |
| *** baffle has joined #openstack-dns | 23:13 | |
| *** jbratton has joined #openstack-dns | 23:13 | |
| *** CaptTofu_ has joined #openstack-dns | 23:13 | |
| *** puck` has joined #openstack-dns | 23:13 | |
| *** briancline has joined #openstack-dns | 23:13 | |
| *** crc32 has joined #openstack-dns | 23:16 | |
| *** vinod has joined #openstack-dns | 23:16 | |
| *** ryanpetrello has quit IRC | 23:22 | |
| openstackgerrit | Vinod Mangalpally proposed openstack/designate: Add more unit tests to mdns service https://review.openstack.org/147669 | 23:23 |
| *** CaptTofu_ has quit IRC | 23:31 | |
| *** CaptTofu_ has joined #openstack-dns | 23:31 | |
| *** chlong has joined #openstack-dns | 23:40 | |
| openstackgerrit | Merged openstack/designate: Add more unit tests to Base Object class https://review.openstack.org/147679 | 23:43 |
| *** vinod has quit IRC | 23:53 | |
| openstackgerrit | Merged openstack/designate: Add Unit Tests for v2 API Reports https://review.openstack.org/147689 | 23:54 |
| *** crc32 has quit IRC | 23:54 | |
| *** csoukup has quit IRC | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!