LiJie20190102 commented on issue #10123: URL: https://github.com/apache/seatunnel/issues/10123#issuecomment-3677765752
> [@LiJie20190102](https://github.com/LiJie20190102) Can you add a specific implementation plan? Then we can discuss it. Currently, when obtaining data from starrocks, the query plan is first obtained from the FE node, and then the obtained query plan is used as parameters, mainly `be_cost` and `be_port`, and distributed to the BE node. Finally, the data returned by the BE node is obtained.The main process is as follows: <img width="522" height="335" alt="Image" src="https://github.com/user-attachments/assets/27c3144e-651f-4a2c-9618-230af8995905"; /> However, there may be issues at this point, such as when starrocks is deployed in the k8s scenario, if the computing cluster and starrocks cluster are not the same cluster, meaning that the computing cluster cannot access starrocks' be domain name and `be_port`, an 'Unknown HostException' will be thrown。 To make up for the above scenario, my idea is to add a `beHostPortMapping` configuration, which is an optional configuration. When the computing cluster is able to fully access `be_ host` and `be_ port`, this configuration can be ignored; If the computing cluster cannot recognize `be_ip` and `be_port`, this configuration can be used to obtain the accessible `be_ip` and `be_port`, The improved flowchart is as follows: <img width="571" height="381" alt="Image" src="https://github.com/user-attachments/assets/fed3c38c-4a9b-4366-99c9-d625cbcd3da2"; /> -- 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]
