You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flink.apache.org by "Martijn Visser (Jira)" <ji...@apache.org> on 2026/04/20 14:15:00 UTC
[jira] [Created] (FLINK-39500) test_batch_wordcount.sh presto and presto_minio variants fail due to FileSink batch mode requiring recoverable writer
Martijn Visser created FLINK-39500:
--------------------------------------
Summary: test_batch_wordcount.sh presto and presto_minio variants fail due to FileSink batch mode requiring recoverable writer
Key: FLINK-39500
URL: https://issues.apache.org/jira/browse/FLINK-39500
Project: Flink
Issue Type: Technical Debt
Components: Tests
Reporter: Martijn Visser
- The presto and presto_minio variants of test_batch_wordcount.sh fail at plan translation with UnsupportedOperationException: This s3 file system implementation does not support recoverable writers
- flink-examples-streaming WordCount.jar uses FileSink which calls createRecoverableWriter unconditionally during getCommittableSerializer, even in batch mode.
- Presto S3 deliberately doesn't support recoverable writers
- These variants aren't in run-nightly-tests.sh, so the breakage went unnoticed. Likely regressed when WordCount was migrated to FileSink (around FLINK-24830 era)
We should consider removing the presto and presto_minio cases from test_batch_wordcount.sh
--
This message was sent by Atlassian Jira
(v8.20.10#820010)