Caizhi Weng created FLINK-24677:
-----------------------------------
Summary: JdbcBatchingOutputFormat should not generate circulate
chaining of exceptions when flushing fails in timer thread
Key: FLINK-24677
URL: https://issues.apache.org/jira/browse/FLINK-24677
Project: Flink
Issue Type: Improvement
Components: Connectors / JDBC
Affects Versions: 1.15.0
Reporter: Caizhi Weng
This is reported from the [user mailing
list|https://lists.apache.org/thread.html/r3e725f52e4f325b9dcb790635cc642bd6018c4bca39f86c71b8a60f4%40%3Cuser.flink.apache.org%3E].
In the timer thread created in {{JdbcBatchingOutputFormat#open}},
{{flushException}} field will be recorded if the call to {{flush}} throws an
exception. This exception is used to fail the job in the main thread.
However {{JdbcBatchingOutputFormat#flush}} will also check for this exception
and will wrap it with a new layer of runtime exception. This will cause a super
long stack when the main thread finally discover the exception and fails.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
