cmccabe opened a new pull request #10395: URL: https://github.com/apache/kafka/pull/10395
Previously, if we assigned one or more calls to a remote node, but it never became available, we would block until the calls hit their the API timeout. This was particularly unfortunate in the case where the calls could have been sent to a different node in the cluster. This PR fixes this behavior by timing out pending connections to remote nodes if they take longer than the request timeout. There are a few other small cleanups in this PR: it removes the unecessary Call#aborted, sets Call#curNode to null after the call has failed to avoid confusion when debugging or logging, and adds a "closing" boolean rather than setting newCalls to null when the client is closed. Also, it increases the log level of the log message that indicates that we timed out some calls because AdminClient closed. Finally, it simplifies the type of callsInFlight. -- 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
