cadonna commented on a change in pull request #9615:
URL: https://github.com/apache/kafka/pull/9615#discussion_r533687031
##########
File path: streams/src/main/java/org/apache/kafka/streams/KafkaStreams.java
##########
@@ -894,19 +885,88 @@ private KafkaStreams(final InternalTopologyBuilder
internalTopologyBuilder,
queryableStoreProvider = new QueryableStoreProvider(storeProviders,
globalStateStoreProvider);
stateDirCleaner = setupStateDirCleaner();
- oldHandler = false;
maybeWarnAboutCodeInRocksDBConfigSetter(log, config);
rocksDBMetricsRecordingService =
maybeCreateRocksDBMetricsRecordingService(clientId, config);
}
+ private StreamThread createStreamThread(final long cacheSizePerThread,
final int threadIdx) {
+ final StreamThread streamThread = StreamThread.create(
+ internalTopologyBuilder,
+ config,
+ clientSupplier,
+ adminClient,
+ processId,
+ clientId,
+ streamsMetrics,
+ time,
+ streamsMetadataState,
+ cacheSizePerThread,
+ stateDirectory,
+ delegatingStateRestoreListener,
+ threadIdx,
+ KafkaStreams.this::closeToError,
+ streamsUncaughtExceptionHandler
+ );
+ threads.add(streamThread);
+ threadState.put(streamThread.getId(), streamThread.state());
+ storeProviders.add(new StreamThreadStateStoreProvider(streamThread));
+ return streamThread;
+ }
+
+ /**
+ * Adds and starts a stream thread in addition to the stream threads that
are already running in this
+ * Kafka Streams client.
+ *
+ * Since the number of stream threads increases, the sizes of the caches
in the new stream thread
+ * and the existing stream threads are adapted so that the sum of the
cache sizes over all stream
+ * threads does not exceed the total cache size specified in configuration
+ * {@code cache.max.bytes.buffering}.
+ *
+ * Stream threads can only be added if this Kafka Streams client is in
state RUNNING or REBALANCING.
+ *
+ * @return name of the added stream thread or empty if a new stream thread
could not be added
+ */
+ public Optional<String> addStreamThread() {
+ if (isRunningOrRebalancing()) {
+ final int threadIdx = getNextThreadIndex();
+ final long cacheSizePerThread =
getCacheSizePerThread(threads.size() + 1);
+ resizeThreadCache(cacheSizePerThread);
+ final StreamThread streamThread =
createStreamThread(cacheSizePerThread, threadIdx);
+ streamThread.setStateListener(streamStateListener);
+ synchronized (stateLock) {
+ if (isRunningOrRebalancing()) {
+ streamThread.start();
Review comment:
Yes, currently this assumption is correct, but if the state transitions
change in future, we would be safe if we do the cleanup.
On a second thought, we are probably not 100% safe because if a transition
from `NOT_RUNNING` to `RUNNING` is added (or any other transition that goes
from the above mentioned states to `RUNNING` or `REBALANCING`), we would still
not do the clean up.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
