[FLINK-8148] Test instability in YarnFileStageTest

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 1.4.0
• Fix Version/s: 1.4.0
• Component/s: Deployment / YARN, Tests
• Labels: test-stability
• Assignee: Nico Kruber
• Reporter: Nico Kruber
• Created: 24/Nov/17 13:53
• Updated: 04/Dec/17 14:56
• Resolved: 04/Dec/17 14:56

Description

Running org.apache.flink.yarn.YarnFileStageTestS3ITCase
Tests run: 3, Failures: 1, Errors: 0, Skipped: 1, Time elapsed: 13.152 sec <<< FAILURE! - in org.apache.flink.yarn.YarnFileStageTestS3ITCase
testRecursiveUploadForYarnS3(org.apache.flink.yarn.YarnFileStageTestS3ITCase)  Time elapsed: 8.515 sec  <<< FAILURE!
java.lang.AssertionError: null
	at org.junit.Assert.fail(Assert.java:86)
	at org.junit.Assert.assertTrue(Assert.java:41)
	at org.junit.Assert.assertFalse(Assert.java:64)
	at org.junit.Assert.assertFalse(Assert.java:74)
	at org.apache.flink.yarn.YarnFileStageTestS3ITCase.testRecursiveUploadForYarn(YarnFileStageTestS3ITCase.java:171)
	at org.apache.flink.yarn.YarnFileStageTestS3ITCase.testRecursiveUploadForYarnS3(YarnFileStageTestS3ITCase.java:192)

from https://travis-ci.org/apache/flink/jobs/305861539

YarnFileStageTestS3ITCase.testRecursiveUploadForYarn verifies that the test directory used is cleaned up by YarnFileStageTest.testCopyFromLocalRecursive which should clean up the directory (in a finally block). However, for S3, we may not always see our own deletes.

Quoting from https://aws.amazon.com/s3/faqs/ here:

Q: What data consistency model does Amazon S3 employ?
Amazon S3 buckets in all Regions provide read-after-write consistency for PUTS of new objects and eventual consistency for overwrite PUTS and DELETES.

Issue Links

• links to: GitHub Pull Request #5066