public class StandbyTask extends AbstractTask
Modifier and Type | Method and Description |
---|---|
void |
close(boolean clean,
boolean isZombie)
-
commit()
- close state |
void |
closeSuspended(boolean clean,
boolean isZombie,
java.lang.RuntimeException e) |
void |
commit()
- flush store
- checkpoint store
- update offset limits
|
boolean |
initializeStateStores()
Initialize the task and return
true if the task is ready to run, i.e, it has not state stores |
void |
initializeTopology() |
void |
resume()
- update offset limits
|
void |
suspend()
- flush store
- checkpoint store
|
java.util.List<org.apache.kafka.clients.consumer.ConsumerRecord<byte[],byte[]>> |
update(org.apache.kafka.common.TopicPartition partition,
java.util.List<org.apache.kafka.clients.consumer.ConsumerRecord<byte[],byte[]>> records)
Updates a state store using records from one change log partition
|
activeTaskCheckpointableOffsets, applicationId, changelogPartitions, commitNeeded, context, getStore, hasStateStores, id, isClosed, isEosEnabled, partitions, topology, toString, toString, updateOffsetLimits
public boolean initializeStateStores()
Task
true
if the task is ready to run, i.e, it has not state storespublic void initializeTopology()
public void resume()
- update offset limits
public void commit()
- flush store - checkpoint store - update offset limits
public void suspend()
- flush store - checkpoint store
public void close(boolean clean, boolean isZombie)
- commit()
- close state
clean
- ignored by StandbyTask
as it can always try to close cleanly
(ie, commit, flush, and write checkpoint file)isZombie
- ignored by StandbyTask
as it can never be a zombiepublic void closeSuspended(boolean clean, boolean isZombie, java.lang.RuntimeException e)
public java.util.List<org.apache.kafka.clients.consumer.ConsumerRecord<byte[],byte[]>> update(org.apache.kafka.common.TopicPartition partition, java.util.List<org.apache.kafka.clients.consumer.ConsumerRecord<byte[],byte[]>> records)