CatOrLeader commented on PR #4071:
URL: https://github.com/apache/flink-cdc/pull/4071#issuecomment-3302242100

   Hello, @lvyanquan! Let me explain a little bit of the rationale behind this: 
for the engineer, it is common sense to structure the environment he is working 
in, so there occur situations like the following
   
   A developer/engineer/DevOps creates a custom Flink CDC image and puts 
multiple .jar files into different folders, for example:
   
   ```bash
   .
   ├── lib
   |   ├── some-custom-code.jar
   │   ├── metric
   │   │   ├── oshi-core-6.1.5.jar
   │   │   ├── jna-5.10.0.jar
   │   │   └── jna-platform-5.10.0.jar
   │   ├── cdc
   │   │   ├── flink-cdc-common-3.4.0.jar
   │   │   ├── flink-cdc-runtime-3.4.0.jar
   │   │   ├── flink-sql-connector-postgres-cdc-3.4.0.jar
   │   │   ├── ojdbc8-19.3.0.0.jar
   │   │   └── xdb-19.3.0.0.jar
   │   └── utils
   │       ├── jackson-databind-2.18.2.jar
   │       ├── jackson-core-2.18.2.jar
   │       ├── jackson-annotations-2.18.2.jar
   │       ├── sshj-0.40.0.jar
   │       └── bcprov-jdk18on-1.80.jar
   ```
   
   Then, we experience a small inconvenience when someone tries to run a Flink 
CDC job and an exception occurs due to not finding any dependencies or even a 
CDC connector. It can take (and has taken) a significant amount of time to 
debug and identify the issue, even if the developer sees that all dependencies 
are present. Additionally, Flink does not prevent the placement of JAR files in 
subdirectories, and it is not clear when the CDC library requires exactly one 
level of depth for storing dependencies.
   


-- 
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]

Reply via email to