merlau commented on PR #12258: URL: https://github.com/apache/seatunnel/pull/12258#issuecomment-5652877709
Thanks for checking the CI runs and logs! We investigated both issues and addressed them in a77b9375: 1. mariadb:10.11 404: The image was not evicted by runner disk pressure, but explicitly deleted by AbstractJdbcIT.tearDown() (dockerClient.removeImageCmd(...)) run by prior sibling tests in the module. Since JdbcMariaDBTimestampIT relies on Testcontainers' in-memory JVM cache, it didn't re-pull the deleted image. We followed the existing SeaTunnel pattern (AbstractJdbcIT:338, JdbcMysqlMultipleTablesIT:287) by adding .withImagePullPolicy(PullPolicy.alwaysPull()). 2. MariaDbCDCCheckpointRestoreIT index mismatch: Upstream bbde2509143 is unrelated here since it only modifies SnapshotSplitAssigner (snapshot phase), whereas this failure happened at Checkpoint 2 during pure binlog streaming (IncrementalSplitAssigner). In fact, all 8 rows were restored without data loss (asserted by row counts in lines 327–333). The mismatch at index [6][0] was caused by SOURCE_SQL_TEMPLATE and SINK_SQL_TEMPLATE lacking an ORDER BY id clause on a table without a primary key after an ALTER TABLE ... DROP PRIMARY KEY rebuild, causing non-deterministic scan ordering. We added deterministic ORDER BY id to the query templates. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
