[jira] [Created] (FLINK-35474) During the process of catching up after a restart, Flink CDC uses the latest table information to parse old binlogs.

[stevedengzhi (Jira)]

stevedengzhi created FLINK-35474:
------------------------------------

             Summary: During the process of catching up after a restart, Flink 
CDC uses the latest table information to parse old binlogs.
                 Key: FLINK-35474
                 URL: https://issues.apache.org/jira/browse/FLINK-35474
             Project: Flink
          Issue Type: Bug
          Components: Flink CDC
    Affects Versions: 1.17.2
            Reporter: stevedengzhi
         Attachments: image-2024-05-28-16-13-58-566.png

!image-2024-05-28-16-13-58-566.png!

 

I don't know how to handle it.
This DDL issue can prevent recovery from checkpoints, leading to infinite 
restarts.
How to reproduce this issue?（翻译by chatGPT，English not well ）
Step 1: Turn off Flink CDC.
Step 2: Perform DDL operations on multiple tables, including deleting columns, 
reducing columns, and adding columns.
Step 3: Restore Flink CDC from the last checkpoint before it was turned off.
 
 
 

 



--
This message was sent by Atlassian Jira
(v8.20.10#820010)