You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Logging system tests are hanging on CircleCI #3237
I'm a bit confused as to under what circumstances these tests are failing? Is it with those error reporting changes or is it just sporadically failing everywhere (sorry Circle interface isn't obviously answering this question to me).
test_log_root_handler(self) is the most complex test and worth commenting out to see if it fixes anything.
The logging environment is completely isolated from the error_reporting changes, so they are unrelated. (Though logging is a dependency of error_reporting, so we want to make sure logging is good before doing a downstream release.)
Every CI run since the first failure has failed, i.e. it doesn't seem sporadic. To replicate we'd want to use the same env. vars. locally?
@chemelnucfin can I task you with taking care of the low-hanging fruit here? (there's some tests that refer to old bigquery names, so should be trivial to fix).
In terms of hanging, see if you can still make it happen after #4759 & I'll see what I can do to get it fixed.
Activity
dhermes commentedon Mar 29, 2017
Note this is "blocking" the release of
error_reporting == 0.23.3/cc @DizzeePascalldhermes commentedon Mar 29, 2017
Running
py.test tests/system.py -vv(under theloggingdirectory) the (deterministic) order of test cases is:(with numbers added by me).
waprin commentedon Mar 29, 2017
I'm a bit confused as to under what circumstances these tests are failing? Is it with those error reporting changes or is it just sporadically failing everywhere (sorry Circle interface isn't obviously answering this question to me).
test_log_root_handler(self)is the most complex test and worth commenting out to see if it fixes anything.dhermes commentedon Mar 29, 2017
The
loggingenvironment is completely isolated from theerror_reportingchanges, so they are unrelated. (Thoughloggingis a dependency oferror_reporting, so we want to make sureloggingis good before doing a downstream release.)Every CI run since the first failure has failed, i.e. it doesn't seem sporadic. To replicate we'd want to use the same env. vars. locally?
dhermes commentedon Mar 30, 2017
@waprin We've now got a fourth instance.
lukesneeringer commentedon Mar 30, 2017
Adding anNevermind, @dhermes beat me to it.issue-3237branch with-vvvso that we can see it more easily.dhermes commentedon Mar 30, 2017
lukesneeringer commentedon Mar 30, 2017
Closing, but merging #3244 to help us watch.
dhermes commentedon May 9, 2017
This issue has come back
test_log_handler_asyncin 2.7)test_log_text_with_timestampin 3.6, 2.7 was green)test_log_struct_w_metadatain 3.6, 2.7 was green)test_log_handler_syncin 2.7)test_log_text_with_timestampin 3.6, 2.7 was green)test_log_text_with_timestampin 3.6, 2.7 was green)test_log_handler_asyncin 2.7)test_log_handler_asyncin 2.7)test_log_handler_asyncin 2.7)The previous failures were:
test_log_structin 2.7)test_log_text_with_timestampin 2.7)test_log_textin 2.7)test_log_handler_asyncin 2.7)Third round of failures:
test_log_text_w_metadatain 2.7)test_log_handler_asyncin 2.7)test_log_textin 3.6)test_log_root_handlerin 2.7,test_log_text_with_resourcein 3.6)test_log_handler_asyncin 2.7,test_log_struct_w_metadatain 3.6)test_log_text_with_resourcein 3.6)test_log_handler_syncin 2.7)dhermes commentedon May 11, 2017
@waprin @jonparrott This is somewhat concerning. Do you guys have any ideas what could be causing this?
28 remaining items
theacodes commentedon Jan 19, 2018
This is now outright failures. We should get this fixed.
theacodes commentedon Jan 19, 2018
Whoever takes this on should wait until after #4759 is merged.
theacodes commentedon Jan 19, 2018
@chemelnucfin can I task you with taking care of the low-hanging fruit here? (there's some tests that refer to old bigquery names, so should be trivial to fix).
In terms of hanging, see if you can still make it happen after #4759 & I'll see what I can do to get it fixed.
chemelnucfin commentedon Jan 19, 2018
I saw the merge, working on the bigquery ones.
chemelnucfin commentedon Jan 22, 2018
Possibly fixed by #4768. Will reopen if that's not the case.