Copilot commented on code in PR #23004:
URL: https://github.com/apache/kafka/pull/23004#discussion_r3979082726


##########
streams/src/main/java/org/apache/kafka/streams/processor/internals/StreamThread.java:
##########
@@ -1926,15 +1921,29 @@ private long advanceNowAndComputeLatency() {
      * (e.g., in testing), hence the state is set only the first time
      *
      * @param operation the group membership operation to apply on shutdown. 
Must be one of LEAVE_GROUP or REMAIN_IN_GROUP.
+     * @return true if this call initiated the shutdown, i.e., transitioned 
the thread to
+     *         {@code PENDING_SHUTDOWN}; false if the thread was already 
shutting down or dead,
+     *         in which case the group membership operation of the earlier 
shutdown request is kept
+     *         (use {@link #updateGroupMembershipOperation} to change it)
      */
-    public void shutdown(final 
org.apache.kafka.streams.CloseOptions.GroupMembershipOperation operation) {
+    public boolean shutdown(final 
org.apache.kafka.streams.CloseOptions.GroupMembershipOperation operation) {
         log.info("Informed to shut down");
         final State oldState = setState(State.PENDING_SHUTDOWN);
+        if (oldState == null) {
+            // Shutdown was already requested by another caller (a concurrent 
removal, thread
+            // replacement, or client close); that caller owns this thread's 
death.
+            return false;
+        }
         leaveGroupRequested.set(operation);
         if (oldState == State.CREATED) {
             // The thread may not have been started. Take responsibility for 
shutting down
             completeShutdown(true);
         }
+        return true;
+    }
+
+    public void updateGroupMembershipOperation(final 
org.apache.kafka.streams.CloseOptions.GroupMembershipOperation operation) {
+        leaveGroupRequested.set(operation);

Review Comment:
   This update is not coordinated with `completeShutdown()`, which reads 
`leaveGroupRequested` before closing the consumer. After `shutdown(operation)` 
returns `false`, the target thread can read the old operation before this 
method executes, so a concurrent client close can still silently lose its 
requested group-membership behavior. Please make the “already pending + 
override operation” action atomic with the operation consumed during shutdown 
(rather than splitting it across `shutdown` and this setter).



-- 
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]

Reply via email to