search

sql-machine-learning / sqlflow

Brings SQL and AI together.
https://sqlflow.org
Apache License 2.0
5.09k stars 699 forks source link

EDL integration test is failing #1203

Closed tonyyang-svail closed 4 years ago

tonyyang-svail commented 4 years ago

Build log: https://travis-ci.com/sql-machine-learning/sqlflow/jobs/257933006

Error message:

...
Error from server (NotFound): pods "elasticdl-test-odps-master" not found
Error from server (NotFound): pods "elasticdl-test-odps-worker-0" not found
Error from server (NotFound): pods "elasticdl-test-odps-worker-1" not found
Master: , Worker0: , Worker1: . Continue checking...
...

This error not always happens. For example, this build succeeded. Build log:

Error from server (NotFound): pods "elasticdl-test-odps-worker-0" not found
Error from server (NotFound): pods "elasticdl-test-odps-worker-1" not found
Master: Pending, Worker0: , Worker1: . Continue checking...
48%
Master: Running, Worker0: Running, Worker1: Running. Continue checking...
47.8%
Master: Running, Worker0: Succeeded, Worker1: Succeeded. Continue checking...
47.4%
Master: Running, Worker0: Succeeded, Worker1: Succeeded. Continue checking...
47.1%
ElasticDL job succeeded.
pod "elasticdl-test-odps-master" deleted
terrytangyuan commented 4 years ago

This is not related to ElasticDL codegen. It's a failing test case before ElasticDL integration test:

=== RUN   TestEnd2EndMaxCompute/TestTrainSQL
time="2019-11-18T20:55:00Z" level=error msg="runSQLProgram error: thirdPartyParse failed: unsupported driver type maxcompute" package=sql
2019/11/18 20:55:00 stream read err: rpc error: code = Unknown desc = thirdPartyParse failed: unsupported driver type maxcompute
FAIL    sqlflow.org/sqlflow/cmd/sqlflowserver   50.999s
FAIL
terrytangyuan commented 4 years ago

This is why I added set -e in #1194 to prevent misleading error messages since there's a test failing before ElasticDL's integration test.

tonyyang-svail commented 4 years ago

@terrytangyuan I see. Thanks.