wcarlson5 commented on code in PR #13927:
URL: https://github.com/apache/kafka/pull/13927#discussion_r1256076934
##########
streams/src/main/java/org/apache/kafka/streams/processor/internals/TaskManager.java:
##########
@@ -1619,7 +1639,14 @@ private List<Task> standbyTaskIterable() {
}
private Stream<Task> standbyTaskStream() {
- return tasks.allTasks().stream().filter(t -> !t.isActive());
+ if (stateUpdater != null) {
+ return Stream.concat(
+ stateUpdater.getStandbyTasks().stream(),
+ tasks.allTasks().stream().filter(t -> !t.isActive())
+ );
+ } else {
+ return tasks.allTasks().stream().filter(t -> !t.isActive());
Review Comment:
This is the same line right above it. Can we simplify them?
##########
streams/src/test/java/org/apache/kafka/streams/integration/NamedTopologyIntegrationTest.java:
##########
@@ -173,9 +173,9 @@ public static void closeCluster() {
private static final List<KeyValue<String, Long>> STANDARD_INPUT_DATA =
asList(pair("A", 100L), pair("B", 200L), pair("A", 300L), pair("C",
400L), pair("C", -50L));
private static final List<KeyValue<String, Long>> COUNT_OUTPUT_DATA =
- asList(pair("B", 1L), pair("A", 2L), pair("C", 2L)); // output of
count operation with caching
Review Comment:
I don't think that output should have changed. That is worrying
##########
streams/src/main/java/org/apache/kafka/streams/processor/internals/TaskManager.java:
##########
@@ -571,7 +571,9 @@ private Map<TaskId, Set<TopicPartition>>
pendingTasksToCreate(final Map<TaskId,
while (iter.hasNext()) {
final Map.Entry<TaskId, Set<TopicPartition>> entry = iter.next();
final TaskId taskId = entry.getKey();
- if (taskId.topologyName() != null &&
!topologyMetadata.namedTopologiesView().contains(taskId.topologyName())) {
+ final boolean taskIsOwned = tasks.allTaskIds().contains(taskId)
Review Comment:
If we are going to just invert this in the next line and not use it anywhere
else can we move that inversion here?
##########
streams/src/test/java/org/apache/kafka/streams/integration/KTableKTableForeignKeyInnerJoinCustomPartitionerIntegrationTest.java:
##########
@@ -207,11 +207,12 @@ public void
shouldThrowIllegalArgumentExceptionWhenCustomPartitionerReturnsMulti
});
}
- startApplicationAndWaitUntilRunning(kafkaStreamsList, ofSeconds(120));
+ for (final KafkaStreams stream: kafkaStreamsList) {
+ stream.start();
+ }
Review Comment:
That makes sense, but do we still need the synchronization point or even the
timeout here?
What I'm asking is is maybe we have a `waitForApplicationState` for running
after calling start for them all
##########
streams/src/main/java/org/apache/kafka/streams/processor/internals/TaskManager.java:
##########
@@ -1600,7 +1617,21 @@ List<Task> activeTaskIterable() {
return activeTaskStream().collect(Collectors.toList());
}
+ List<Task> activeRunningTaskIterable() {
+ return activeRunningTaskStream().collect(Collectors.toList());
+ }
+
private Stream<Task> activeTaskStream() {
+ if (stateUpdater != null) {
+ return Stream.concat(
+ activeRunningTaskStream(),
+ stateUpdater.getTasks().stream().filter(Task::isActive)
+ );
+ }
Review Comment:
Should the state updater have active tasks at all? If so who owns them?
--
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]
