Skip to content

perf(kafka-connect): avoid per-record TopicPartition allocation in HoodieSinkTask.put #19018

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wombatu-kun wants to merge 1 commit into
base: master
Choose a base branch
from
+15 −5
Open

perf(kafka-connect): avoid per-record TopicPartition allocation in HoodieSinkTask.put #19018

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 15 additions & 5 deletions hudi-kafka-connect/src/main/java/org/apache/hudi/connect/HoodieSinkTask.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,6 +53,8 @@ public class HoodieSinkTask extends SinkTask {

private final Map<TopicPartition, TransactionCoordinator> transactionCoordinators;
private final Map<TopicPartition, TransactionParticipant> transactionParticipants;
// Secondary index for per-record routing in put(); avoids allocating a TopicPartition key per record.
private final Map<String, Map<Integer, TransactionParticipant>> participantsByTopicPartition;
private KafkaConnectControlAgent controlKafkaClient;
private KafkaConnectConfigs connectConfigs;

Expand All @@ -62,6 +64,7 @@ public class HoodieSinkTask extends SinkTask {
public HoodieSinkTask() {
transactionCoordinators = new HashMap<>();
transactionParticipants = new HashMap<>();
participantsByTopicPartition = new HashMap<>();
}

@Override
Expand Down Expand Up @@ -93,11 +96,9 @@ public void start(Map<String, String> props) {
@Override
public void put(Collection<SinkRecord> records) {
for (SinkRecord record : records) {
String topic = record.topic();
int partition = record.kafkaPartition();
TopicPartition tp = new TopicPartition(topic, partition);

TransactionParticipant transactionParticipant = transactionParticipants.get(tp);
Map<Integer, TransactionParticipant> participantsForTopic = participantsByTopicPartition.get(record.topic());
TransactionParticipant transactionParticipant =
participantsForTopic == null ? null : participantsForTopic.get(record.kafkaPartition());
if (transactionParticipant != null) {
transactionParticipant.buffer(record);
}
Expand Down Expand Up @@ -168,6 +169,13 @@ public void close(Collection<TopicPartition> partitions) {
}
}
TransactionParticipant worker = transactionParticipants.remove(partition);
Map<Integer, TransactionParticipant> participantsForTopic = participantsByTopicPartition.get(partition.topic());
if (participantsForTopic != null) {
participantsForTopic.remove(partition.partition());
if (participantsForTopic.isEmpty()) {
participantsByTopicPartition.remove(partition.topic());
}
}
if (worker != null) {
try {
log.debug("Closing data writer due to task start failure.");
Expand All @@ -194,6 +202,7 @@ private void bootstrap(Collection<TopicPartition> partitions) {
}
ConnectTransactionParticipant worker = new ConnectTransactionParticipant(connectConfigs, partition, controlKafkaClient, context);
transactionParticipants.put(partition, worker);
participantsByTopicPartition.computeIfAbsent(partition.topic(), k -> new HashMap<>()).put(partition.partition(), worker);
worker.start();
} catch (HoodieException exception) {
log.error("Fatal error initializing task {} for partition {}", taskId, partition.partition(), exception);
Expand All @@ -214,6 +223,7 @@ private void cleanup() {
}
}
transactionParticipants.clear();
participantsByTopicPartition.clear();
transactionCoordinators.forEach((topic, transactionCoordinator) -> transactionCoordinator.stop());
transactionCoordinators.clear();
}
Expand Down
Loading