zchovan commented on code in PR #8: URL: https://github.com/apache/flink-connector-kudu/pull/8#discussion_r1973474719
########## flink-connector-kudu/src/main/java/org/apache/flink/connector/kudu/source/enumerator/KuduSourceEnumerator.java: ########## @@ -0,0 +1,230 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.flink.connector.kudu.source.enumerator; + +import org.apache.flink.api.connector.source.SourceEvent; +import org.apache.flink.api.connector.source.SplitEnumerator; +import org.apache.flink.api.connector.source.SplitEnumeratorContext; +import org.apache.flink.connector.kudu.connector.KuduTableInfo; +import org.apache.flink.connector.kudu.connector.reader.KuduReaderConfig; +import org.apache.flink.connector.kudu.source.split.KuduSourceSplit; +import org.apache.flink.connector.kudu.source.split.SplitFinishedEvent; + +import org.apache.kudu.util.HybridTimeUtil; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import javax.annotation.Nonnull; +import javax.annotation.Nullable; + +import java.io.IOException; +import java.time.Duration; +import java.util.ArrayList; +import java.util.Iterator; +import java.util.List; +import java.util.Optional; +import java.util.concurrent.TimeUnit; +import java.util.function.Supplier; + +/** + * The Kudu source enumerator is responsible for periodically discovering and assigning new splits + * when possible. + * + * <p>To provide CDC-like functionality, the enumeration works as follows: Initially, we perform a + * snapshot read of the table and mark the snapshot time as t0. From that point onward, we perform + * differential scans in the time intervals t0 - t1, t1 - t2, and so on. + * + * <p>This approach means that new splits can only be enumerated once the current time range is + * fully processed. + * + * <p>The process is controlled as follows: + * + * <ul> + * <li>Once a set of splits is enumerated for a time range, we track: + * <ul> + * <li><b>Unassigned splits</b>: Discovered but not yet assigned to readers. + * <li><b>Pending splits</b>: Assigned but not yet fully processed. + * </ul> + * <li>A new set of splits is generated only when there are no remaining unassigned or pending + * splits. + * </ul> + */ +public class KuduSourceEnumerator + implements SplitEnumerator<KuduSourceSplit, KuduSourceEnumeratorState> { + private static final Logger LOG = LoggerFactory.getLogger(KuduSourceEnumerator.class); + + private final SplitEnumeratorContext<KuduSourceSplit> context; + private final List<Integer> readersAwaitingSplit; + private final List<KuduSourceSplit> unassigned; + private final List<KuduSourceSplit> pending; + private final Duration period; + private final KuduSplitGenerator splitGenerator; + private long lastEndTimestamp; + + public KuduSourceEnumerator( + KuduTableInfo tableInfo, + KuduReaderConfig readerConfig, + Duration period, + SplitEnumeratorContext<KuduSourceSplit> context, + KuduSourceEnumeratorState enumState) { + this.period = period; + this.context = context; + this.readersAwaitingSplit = new ArrayList<>(); + this.unassigned = enumState == null ? new ArrayList<>() : enumState.getUnassigned(); + this.pending = enumState == null ? new ArrayList<>() : enumState.getPending(); + this.splitGenerator = new KuduSplitGenerator(readerConfig, tableInfo); + this.lastEndTimestamp = enumState == null ? -1L : enumState.getLastEndTimestamp(); + } + + @Override + public void start() { + context.callAsync( + () -> this.enumerateNewSplits(() -> pending.isEmpty() && unassigned.isEmpty()), + this::assignSplits, + 0, + this.period.toMillis()); + } + + @Override + public void handleSplitRequest(int subtaskId, @Nullable String requesterHostname) { + readersAwaitingSplit.add(subtaskId); + assignSplitsToReaders(); + } + + @Override + public void addSplitsBack(List<KuduSourceSplit> splits, int subtaskId) { + LOG.debug("File Source Enumerator adds splits back: {}", splits); + unassigned.addAll(splits); + if (context.registeredReaders().containsKey(subtaskId)) { + readersAwaitingSplit.add(subtaskId); + } + assignSplitsToReaders(); + } + + @Override + public void addReader(int subtaskId) { + // The source is purely lazy-pull-based, nothing to do upon registration + } + + @Override + public KuduSourceEnumeratorState snapshotState(long checkpointId) throws Exception { + return new KuduSourceEnumeratorState(lastEndTimestamp, unassigned, pending); + } + + @Override + public void close() throws IOException { + try { + splitGenerator.close(); + } catch (Exception e) { + throw new IOException("Error closing split generator", e); + } + } + + @Override + public void handleSourceEvent(int subtaskId, SourceEvent sourceEvent) { + if (sourceEvent instanceof SplitFinishedEvent) { + SplitFinishedEvent splitFinishedEvent = (SplitFinishedEvent) sourceEvent; + LOG.debug( + "Received SplitFinishedEvent from subtask {} for splits: {}", + subtaskId, + splitFinishedEvent.getFinishedSplits()); + pending.removeAll(splitFinishedEvent.getFinishedSplits()); + readersAwaitingSplit.add(subtaskId); + } + } + + // This function is invoked repeatedly according to this.period if there are no outstanding + // splits. + // Outstanding meaning that there are no pending splits, and no enumerated but not assigned + // splits for the + // current period. + private List<KuduSourceSplit> enumerateNewSplits(@Nonnull Supplier<Boolean> shouldGenerate) { + if (shouldGenerate.get() == Boolean.FALSE) { + return null; + } + List<KuduSourceSplit> newSplits; + + if (isFirstSplitGeneration()) { + this.lastEndTimestamp = getCurrentHybridTime(); + newSplits = this.splitGenerator.generateFullScanSplits(this.lastEndTimestamp); + } else { + long startHT = this.lastEndTimestamp; + long endHT = getCurrentHybridTime(); + newSplits = this.splitGenerator.generateIncrementalSplits(startHT, endHT); + this.lastEndTimestamp = endHT; + } + + return newSplits; + } + + private void assignSplits(List<KuduSourceSplit> splits, Throwable error) { + if (error != null) { + LOG.error("Failed to enumerate Kudu splits.", error); + return; + } + + if (splits != null) { + this.unassigned.addAll(splits); + } + assignSplitsToReaders(); + } + + private void assignSplitsToReaders() { + final Iterator<Integer> awaitingSubtasks = readersAwaitingSplit.iterator(); + + while (awaitingSubtasks.hasNext()) { + final int awaitingSubtask = awaitingSubtasks.next(); + + // If the reader that requested another split has failed in the meantime, remove + // it from the list of waiting readers + if (!context.registeredReaders().containsKey(awaitingSubtask)) { + awaitingSubtasks.remove(); + continue; + } + + final Optional<KuduSourceSplit> nextSplit = getNextSplit(); + if (nextSplit.isPresent()) { + context.assignSplit(nextSplit.get(), awaitingSubtask); + awaitingSubtasks.remove(); + pending.add(nextSplit.get()); + } else { + break; + } + } + } + + private Optional<KuduSourceSplit> getNextSplit() { + if (unassigned.isEmpty()) { + return Optional.empty(); + } + Iterator<KuduSourceSplit> iterator = unassigned.iterator(); + KuduSourceSplit next = iterator.next(); + iterator.remove(); + return Optional.of(next); + } + + // Helper method to get the current Hybrid Time + private long getCurrentHybridTime() { + long fromMicros = TimeUnit.MILLISECONDS.toMicros(System.currentTimeMillis()); + return HybridTimeUtil.physicalAndLogicalToHTTimestamp(fromMicros, 0) + 1; Review Comment: not sure about the +1, @martongreber ? otherwise done ########## flink-connector-kudu/src/main/java/org/apache/flink/connector/kudu/source/enumerator/KuduSourceEnumerator.java: ########## @@ -0,0 +1,230 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.flink.connector.kudu.source.enumerator; + +import org.apache.flink.api.connector.source.SourceEvent; +import org.apache.flink.api.connector.source.SplitEnumerator; +import org.apache.flink.api.connector.source.SplitEnumeratorContext; +import org.apache.flink.connector.kudu.connector.KuduTableInfo; +import org.apache.flink.connector.kudu.connector.reader.KuduReaderConfig; +import org.apache.flink.connector.kudu.source.split.KuduSourceSplit; +import org.apache.flink.connector.kudu.source.split.SplitFinishedEvent; + +import org.apache.kudu.util.HybridTimeUtil; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import javax.annotation.Nonnull; +import javax.annotation.Nullable; + +import java.io.IOException; +import java.time.Duration; +import java.util.ArrayList; +import java.util.Iterator; +import java.util.List; +import java.util.Optional; +import java.util.concurrent.TimeUnit; +import java.util.function.Supplier; + +/** + * The Kudu source enumerator is responsible for periodically discovering and assigning new splits + * when possible. + * + * <p>To provide CDC-like functionality, the enumeration works as follows: Initially, we perform a + * snapshot read of the table and mark the snapshot time as t0. From that point onward, we perform + * differential scans in the time intervals t0 - t1, t1 - t2, and so on. + * + * <p>This approach means that new splits can only be enumerated once the current time range is + * fully processed. + * + * <p>The process is controlled as follows: + * + * <ul> + * <li>Once a set of splits is enumerated for a time range, we track: + * <ul> + * <li><b>Unassigned splits</b>: Discovered but not yet assigned to readers. + * <li><b>Pending splits</b>: Assigned but not yet fully processed. + * </ul> + * <li>A new set of splits is generated only when there are no remaining unassigned or pending + * splits. + * </ul> + */ +public class KuduSourceEnumerator + implements SplitEnumerator<KuduSourceSplit, KuduSourceEnumeratorState> { + private static final Logger LOG = LoggerFactory.getLogger(KuduSourceEnumerator.class); + + private final SplitEnumeratorContext<KuduSourceSplit> context; + private final List<Integer> readersAwaitingSplit; + private final List<KuduSourceSplit> unassigned; + private final List<KuduSourceSplit> pending; + private final Duration period; + private final KuduSplitGenerator splitGenerator; + private long lastEndTimestamp; + + public KuduSourceEnumerator( + KuduTableInfo tableInfo, + KuduReaderConfig readerConfig, + Duration period, + SplitEnumeratorContext<KuduSourceSplit> context, + KuduSourceEnumeratorState enumState) { + this.period = period; + this.context = context; + this.readersAwaitingSplit = new ArrayList<>(); + this.unassigned = enumState == null ? new ArrayList<>() : enumState.getUnassigned(); + this.pending = enumState == null ? new ArrayList<>() : enumState.getPending(); + this.splitGenerator = new KuduSplitGenerator(readerConfig, tableInfo); + this.lastEndTimestamp = enumState == null ? -1L : enumState.getLastEndTimestamp(); + } + + @Override + public void start() { + context.callAsync( + () -> this.enumerateNewSplits(() -> pending.isEmpty() && unassigned.isEmpty()), + this::assignSplits, + 0, + this.period.toMillis()); + } + + @Override + public void handleSplitRequest(int subtaskId, @Nullable String requesterHostname) { + readersAwaitingSplit.add(subtaskId); + assignSplitsToReaders(); + } + + @Override + public void addSplitsBack(List<KuduSourceSplit> splits, int subtaskId) { + LOG.debug("File Source Enumerator adds splits back: {}", splits); + unassigned.addAll(splits); + if (context.registeredReaders().containsKey(subtaskId)) { + readersAwaitingSplit.add(subtaskId); + } + assignSplitsToReaders(); + } + + @Override + public void addReader(int subtaskId) { + // The source is purely lazy-pull-based, nothing to do upon registration + } + + @Override + public KuduSourceEnumeratorState snapshotState(long checkpointId) throws Exception { + return new KuduSourceEnumeratorState(lastEndTimestamp, unassigned, pending); + } + + @Override + public void close() throws IOException { + try { + splitGenerator.close(); + } catch (Exception e) { + throw new IOException("Error closing split generator", e); + } + } + + @Override + public void handleSourceEvent(int subtaskId, SourceEvent sourceEvent) { + if (sourceEvent instanceof SplitFinishedEvent) { + SplitFinishedEvent splitFinishedEvent = (SplitFinishedEvent) sourceEvent; + LOG.debug( + "Received SplitFinishedEvent from subtask {} for splits: {}", + subtaskId, + splitFinishedEvent.getFinishedSplits()); + pending.removeAll(splitFinishedEvent.getFinishedSplits()); + readersAwaitingSplit.add(subtaskId); + } + } + + // This function is invoked repeatedly according to this.period if there are no outstanding + // splits. + // Outstanding meaning that there are no pending splits, and no enumerated but not assigned + // splits for the + // current period. + private List<KuduSourceSplit> enumerateNewSplits(@Nonnull Supplier<Boolean> shouldGenerate) { Review Comment: done -- 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: issues-unsubscr...@flink.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
