kpatelatwork commented on a change in pull request #10530: URL: https://github.com/apache/kafka/pull/10530#discussion_r617556919
########## File path: connect/runtime/src/main/java/org/apache/kafka/connect/runtime/rest/RestServer.java ########## @@ -369,9 +370,34 @@ else if (serverConnector != null && serverConnector.getHost() != null && serverC else if (serverConnector != null && serverConnector.getPort() > 0) builder.port(serverConnector.getPort()); - log.info("Advertised URI: {}", builder.build()); + URI uri = builder.build(); + validateUriHost(uri); + log.info("Advertised URI: {}", uri); - return builder.build(); + return uri; + } + + /** + * Parses the uri and throws a more definitive error + * when the internal node to node communication can't happen due to an invalid host name. + */ + static void validateUriHost(URI uri) { + //java URI parsing will fail silently returning null in the host if the host name contains invalid characters like _ + //this bubbles up later when the Herder tries to communicate on the advertised url and the current HttpClient fails with an ambiguous message + //we need to revisit this when we upgrade to a better HttpClient that can communicate with such host names or throws a better error message Review comment: @tombentley excellent points about not parsing the hostname from URI and I also felt Utils.getHost was broken as it didn't worked on a case with path in the URL and that's why I passed authority to it. Let me work on a change using the approach you suggested and get back to you. I was doing uri.getHost null check to avoid the blast radius in case we didn't anticipated some pattern so I think we should still use the null check but as you said use the hostname that is available to us instead of parsing it from authority and uri. -- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org