pnowojski commented on a change in pull request #16589:
URL: https://github.com/apache/flink/pull/16589#discussion_r680147864



##########
File path: 
flink-streaming-java/src/main/java/org/apache/flink/streaming/runtime/tasks/SourceStreamTask.java
##########
@@ -281,15 +320,50 @@ protected void declineCheckpoint(long checkpointId) {
         public void run() {
             try {
                 mainOperator.run(lock, operatorChain);
-                if (!wasStoppedExternally && !isCanceled()) {
-                    synchronized (lock) {
-                        operatorChain.setIgnoreEndOfInput(false);
-                    }
-                }
+                completeProcessing();
                 completionFuture.complete(null);
             } catch (Throwable t) {
                 // Note, t can be also an InterruptedException
-                completionFuture.completeExceptionally(t);
+                if (isCanceled()
+                        && ExceptionUtils.findThrowable(t, 
InterruptedException.class)
+                                .isPresent()) {
+                    completionFuture.completeExceptionally(new 
CancelTaskException(t));
+                } else if (finishingReason == 
FinishingReason.STOP_WITH_SAVEPOINT_DRAIN
+                        && ExceptionUtils.findThrowable(t, 
InterruptedException.class)
+                                .isPresent()) {
+                    // if we are stopping the source thread for 
stop-with-savepoint
+                    // we may actually return from run with an 
InterruptedException which
+                    // should be ignored (e.g. Kinesis case see FLINK-23528)
+                    try {
+                        // clear the interrupted status for the thread
+                        Thread.interrupted();
+                        completeProcessing();
+                        completionFuture.complete(null);
+                    } catch (Throwable e) {
+                        completionFuture.completeExceptionally(e);
+                    }

Review comment:
       I presume it's a reminiscent of previous version of the code, where we 
were expecting failures/exceptions and we were interrupting the code. 
   
   Can you maybe point to the code it was handling this case before?




-- 
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