[GitHub] [pulsar] michaeljmarshall commented on a change in pull request #13023: Fix consume message order issue when use listener.

[GitBox <gi...@apache.org>]

michaeljmarshall commented on a change in pull request #13023:
URL: https://github.com/apache/pulsar/pull/13023#discussion_r768080748



##########
File path: pulsar-client/src/main/java/org/apache/pulsar/client/impl/ConsumerBase.java
##########
@@ -916,30 +916,32 @@ private void doPendingBatchReceiveTask(Timeout timeout) {
     protected void triggerListener() {
         // Trigger the notification on the message listener in a separate thread to avoid blocking the networking
         // thread while the message processing happens
-        try {
-            // Control executor to call MessageListener one by one.
-            if (executorQueueSize.get() < 1) {
-                final Message<T> msg = internalReceive(0, TimeUnit.MILLISECONDS);
-                if (msg != null) {
-                    executorQueueSize.incrementAndGet();
-                    if (SubscriptionType.Key_Shared == conf.getSubscriptionType()) {
-                        executorProvider.getExecutor(peekMessageKey(msg)).execute(() ->
-                                callMessageListener(msg));
-                    } else {
-                        getExternalExecutor(msg).execute(() -> {
-                            callMessageListener(msg);
-                        });
+        internalPinnedExecutor.execute(() -> {
+            try {
+                // Control executor to call MessageListener one by one.
+                if (executorQueueSize.get() < 1) {
+                    final Message<T> msg = internalReceive(0, TimeUnit.MILLISECONDS);
+                    if (msg != null) {
+                        executorQueueSize.incrementAndGet();

Review comment:
       Here is the race condition. Lines 922 and 925. The non-synchronized get and subsequent update would explain messages processed out of order. By putting this on the same thread, `internalPinnedExecutor`, the race is no longer possible. However, I wonder if we want the extra thread switching here and if we want the `internalReceive` method called on the `internalPinnedExecutor` for _all_ method calls. It seems like we should have instead solved this data race.




-- 
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: commits-unsubscribe@pulsar.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org