tomscut edited a comment on pull request #3538: URL: https://github.com/apache/hadoop/pull/3538#issuecomment-955108462
> How about using the word of "port" instead of "clientPort" here, and adding "clientPort" as the actual client server port for RBF in another JIRA? Thank you @tasanuma for your advice. Based on your suggestion, I have an idea like this: We leave the ClientPort field for now. 1. When a client sends a request directly to Namenode, the ```clientPort``` records the port of the real client. 2. When the client sends a request to Router and then forwards to NameNode, we set the ```clientPort``` (the real clientPort) in the ```CallerContext``` of the Router. Before NameNode prints the audit log, if the ```CallerContext``` already contains the ```clientPort``` field, we will not set port again. (I plan to do this in another PR) In both cases, only one ```clientPort``` field is left in the CallerContext, which holds the actual clientPort. What do you think of this? Looking forward to your comments. Thank you. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
