mxm commented on code in PR #18303:
URL: https://github.com/apache/iceberg/pull/18303#discussion_r4399842104
##########
flink/v2.3/flink/src/main/java/org/apache/iceberg/flink/sink/dynamic/DynamicWriteResultAggregator.java:
##########
@@ -72,21 +86,41 @@ class DynamicWriteResultAggregator
private transient Map<String, Tuple2<ManifestOutputFileFactory, Integer>>
outputFileFactoriesAndFormatVersions;
private transient String flinkJobId;
+ private transient String sinkId;
+ private transient ListState<String> sinkIdState;
private transient String operatorId;
private transient int subTaskId;
private transient int attemptId;
private transient Catalog catalog;
- DynamicWriteResultAggregator(CatalogLoader catalogLoader, int
cacheMaximumSize) {
+ DynamicWriteResultAggregator(
+ CatalogLoader catalogLoader, int cacheMaximumSize, String initialSinkId)
{
this.catalogLoader = catalogLoader;
this.cacheMaximumSize = cacheMaximumSize;
+ this.initialSinkId = initialSinkId;
}
@Override
public void initializeState(StateInitializationContext context) throws
Exception {
context
.getRestoredCheckpointId()
.ifPresent(checkpointId -> this.lastCheckpointId = checkpointId);
+
+ this.sinkIdState =
context.getOperatorStateStore().getUnionListState(SINK_ID_DESCRIPTOR);
+ if (!CheckpointingOptions.isCheckpointingEnabled(
+ getContainingTask().getEnvironment().getJobConfiguration())) {
+ this.sinkId = getContainingTask().getEnvironment().getJobID().toString();
Review Comment:
Do we need the jobId here? `initialSinkId` is part of the serialized
operator, so batch retries within the job already see the same id.
##########
flink/v2.3/flink/src/main/java/org/apache/iceberg/flink/sink/dynamic/DynamicWriteResultAggregator.java:
##########
@@ -72,21 +86,41 @@ class DynamicWriteResultAggregator
private transient Map<String, Tuple2<ManifestOutputFileFactory, Integer>>
outputFileFactoriesAndFormatVersions;
private transient String flinkJobId;
+ private transient String sinkId;
+ private transient ListState<String> sinkIdState;
private transient String operatorId;
private transient int subTaskId;
private transient int attemptId;
private transient Catalog catalog;
- DynamicWriteResultAggregator(CatalogLoader catalogLoader, int
cacheMaximumSize) {
+ DynamicWriteResultAggregator(
+ CatalogLoader catalogLoader, int cacheMaximumSize, String initialSinkId)
{
this.catalogLoader = catalogLoader;
this.cacheMaximumSize = cacheMaximumSize;
+ this.initialSinkId = initialSinkId;
}
@Override
public void initializeState(StateInitializationContext context) throws
Exception {
context
.getRestoredCheckpointId()
.ifPresent(checkpointId -> this.lastCheckpointId = checkpointId);
+
+ this.sinkIdState =
context.getOperatorStateStore().getUnionListState(SINK_ID_DESCRIPTOR);
+ if (!CheckpointingOptions.isCheckpointingEnabled(
+ getContainingTask().getEnvironment().getJobConfiguration())) {
+ this.sinkId = getContainingTask().getEnvironment().getJobID().toString();
+ } else {
+ this.sinkId = Iterables.getFirst(sinkIdState.get(), initialSinkId);
Review Comment:
Do we need to restore the sinkId? The committables already carry it in the
committer state, so replayed committables are deduplicated correctly.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
