You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@flink.apache.org by "FaxianZhao (JIRA)" <ji...@apache.org> on 2019/05/13 09:54:00 UTC

[jira] [Created] (FLINK-12499) JDBCAppendTableSink support retry

FaxianZhao created FLINK-12499:
----------------------------------

             Summary: JDBCAppendTableSink support retry
                 Key: FLINK-12499
                 URL: https://issues.apache.org/jira/browse/FLINK-12499
             Project: Flink
          Issue Type: Improvement
          Components: Connectors / JDBC
            Reporter: FaxianZhao


JDBCOutputFormat will try to restart the whole job while it catch any SQLException depend on the Restart Strategy. 
So, the network spike is a terrible thing for streaming job which has a jdbc sink. Sometimes one more try will save it. 
Currently, the common case for streaming job delivering messages at-least-once is to run idempotent queries, retry more than once is acceptable.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)