poorbarcode commented on code in PR #19585:
URL: https://github.com/apache/pulsar/pull/19585#discussion_r1119669778
##########
pulsar-client/src/main/java/org/apache/pulsar/client/impl/ConsumerImpl.java:
##########
@@ -1075,6 +1076,45 @@ private void closeConsumerTasks() {
}
negativeAcksTracker.close();
stats.getStatTimeout().ifPresent(Timeout::cancel);
+ releaseMsgIfEnabledPooledMsg();
+ }
+
+ /**
+ * If enabled pooled messages, we should release the messages after
closing consumer.
+ * 1. Call "clearIncomingMessages" regardless of whether
"internalPinnedExecutor" has shutdown or not.
+ * 2. Increment epoch will auto release all messages which received
after close.
+ * 3. Call "clearIncomingMessages" in "internalPinnedExecutor" is used
to clear messages in flight.
+ */
+ private void releaseMsgIfEnabledPooledMsg() {
+ if (!poolMessages) {
+ return;
+ }
+ // Increment epoch.
+ CONSUMER_EPOCH.incrementAndGet(this);
+ // Try clear the incoming queue in internalPinnedExecutor-thread.
+ boolean executorIsShutdown = false;
+ try {
+ if (!internalPinnedExecutor.isShutdown()) {
+ internalPinnedExecutor.execute(this::clearIncomingMessages);
+ }
+ } catch (RejectedExecutionException rejectedExecutionException) {
+ executorIsShutdown = true;
+ }
+ // If internalPinnedExecutor is shutdown, try await termination and
clear the incoming queue.
+ if (executorIsShutdown) {
+ boolean awaitTerminationSuccess = false;
+ try {
+ awaitTerminationSuccess =
internalPinnedExecutor.awaitTermination(1, TimeUnit.SECONDS);
Review Comment:
> We cannot guarantee that all tasks can be executed after 1 second
Yes, but there is another mechanism that ensures that:
We increase the consumer epoch before `awaitTermination`, which prevents
subsequent messages from entering the queue.
--
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]
