lianetm commented on code in PR #22617:
URL: https://github.com/apache/kafka/pull/22617#discussion_r3438677166
##########
clients/src/main/java/org/apache/kafka/clients/FetchSessionHandler.java:
##########
@@ -526,11 +526,15 @@ private String
responseDataToLogString(Set<TopicPartition> topicPartitions) {
*/
public boolean handleResponse(FetchResponse response, short version) {
if (response.error() != Errors.NONE) {
- log.info("Node {} was unable to process the fetch request with {}:
{}.",
- node, nextMetadata, response.error());
if (response.error() == Errors.FETCH_SESSION_ID_NOT_FOUND) {
Review Comment:
updated.
The invalid epoch should not be as common I would imagine but still makes
sense to log as in session not found as the handling and impact on the client
is the same (none, self-healing). I imagine it's less common because the
consumer sends a full fetch request upon errors, so would start a new session
even if it carries an old epoch it seems (just sharing because I didn't know
myself this on fetch, same approach in the consumer and heartbeats: full
request upon errors with the GC). Btw, I ended up consolidating all other
errors like the invalid epoch (all are top-level errors with that are handled
the same: re-send full fetch with nextCloseExistingAttemptNew)
--
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]
