Github user aljoscha commented on a diff in the pull request:
https://github.com/apache/flink/pull/1984#discussion_r63523040
--- Diff:
flink-streaming-java/src/main/java/org/apache/flink/streaming/api/functions/source/FileSplitReadOperator.java
---
@@ -0,0 +1,368 @@
+/*
+ * 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.streaming.api.functions.source;
+
+import org.apache.flink.api.common.io.CheckpointableInputFormat;
+import org.apache.flink.api.common.io.FileInputFormat;
+import org.apache.flink.api.common.typeinfo.TypeInformation;
+import org.apache.flink.api.common.typeutils.TypeSerializer;
+import org.apache.flink.api.java.tuple.Tuple2;
+import org.apache.flink.api.java.tuple.Tuple3;
+import org.apache.flink.configuration.Configuration;
+import org.apache.flink.core.fs.FileInputSplit;
+import org.apache.flink.core.memory.DataInputViewStreamWrapper;
+import org.apache.flink.runtime.state.AbstractStateBackend;
+import org.apache.flink.runtime.state.StreamStateHandle;
+import org.apache.flink.streaming.api.operators.AbstractStreamOperator;
+import org.apache.flink.streaming.api.operators.OneInputStreamOperator;
+import org.apache.flink.streaming.api.operators.TimestampedCollector;
+import org.apache.flink.streaming.api.watermark.Watermark;
+import org.apache.flink.streaming.runtime.streamrecord.StreamRecord;
+import org.apache.flink.streaming.runtime.tasks.StreamTaskState;
+import org.apache.flink.util.Preconditions;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.io.ObjectInputStream;
+import java.io.ObjectOutputStream;
+import java.io.Serializable;
+import java.util.ArrayList;
+import java.util.LinkedList;
+import java.util.List;
+import java.util.Queue;
+
+import static org.apache.flink.util.Preconditions.checkNotNull;
+
+/**
+ * This is the operator that reads the splits received from {@link
FileSplitMonitoringFunction}.
+ * This operator will receive just the split descriptors and then read and
emit records. This may lead
+ * to backpressure. To avoid this, we will have another thread actually
reading the splits and
+ * another forwarding the checkpoint barriers. The two should sync so that
the checkpoints reflect the
+ * current state.
+ * */
+public class FileSplitReadOperator<OUT, S extends Serializable> extends
AbstractStreamOperator<OUT>
+ implements OneInputStreamOperator<FileInputSplit, OUT> {
+
+ private static final Logger LOG =
LoggerFactory.getLogger(FileSplitReadOperator.class);
+
+ private static final FileInputSplit EOF = new FileInputSplit(-1, null,
-1, -1, null);
+
+ private transient SplitReader<S, OUT> reader;
+ private transient TimestampedCollector<OUT> collector;
+
+ private Configuration configuration;
+ private FileInputFormat<OUT> format;
+ private TypeInformation<OUT> typeInfo;
+
+ private Tuple3<List<FileInputSplit>, FileInputSplit, S> readerState;
+
+ public FileSplitReadOperator(FileInputFormat<OUT> format,
TypeInformation<OUT> typeInfo, Configuration configuration) {
+ this.format = checkNotNull(format);
+ this.typeInfo = checkNotNull(typeInfo);
+ this.configuration = checkNotNull(configuration);
+ }
+
+ @Override
+ public void open() throws Exception {
+ super.open();
+
+ this.format.configure(configuration);
+ this.collector = new TimestampedCollector<>(output);
+
+ TypeSerializer<OUT> serializer =
typeInfo.createSerializer(getRuntimeContext().getExecutionConfig());
+ Object checkpointLock = getContainingTask().getCheckpointLock();
+
+ this.reader = new SplitReader<>(format, serializer, collector,
checkpointLock);
+ this.reader.setReaderState(this.readerState);
+ this.reader.start();
+ this.readerState = null;
+ }
+
+ @Override
+ public void processElement(StreamRecord<FileInputSplit> element) throws
Exception {
+ reader.addSplit(element.getValue());
+ }
+
+ @Override
+ public void processWatermark(Watermark mark) throws Exception {
+ output.emitWatermark(mark);
+ }
+
+ @Override
+ public void dispose() {
+ super.dispose();
+
+ // first try to cancel it properly and
+ // give it some time until it finishes
+ reader.cancel();
+ try {
+ reader.join(200);
+ } catch (InterruptedException e) {
+ // we can ignore this
+ }
+
+ // if the above did not work, then interrupt the thread
repeatedly
+ while (reader.isAlive()) {
+
+ StringBuilder bld = new StringBuilder();
+ StackTraceElement[] stack = reader.getStackTrace();
+ for (StackTraceElement e : stack) {
+ bld.append(e).append('\n');
+ }
+ LOG.warn("The reader is stuck in method:\n {}",
bld.toString());
+
+ reader.interrupt();
+ try {
+ reader.join(50);
+ } catch (InterruptedException e) {
+ // we can ignore this
+ }
+ }
+ reader = null;
+ collector = null;
+ configuration = null;
+ format = null;
+ typeInfo = null;
+ }
+
+ @Override
+ public void close() throws Exception {
+ super.close();
+
+ // signal that no more splits will come, wait for the reader to
finish
+ // and close the collector. Further cleaning up is handled by
the dispose().
+
+ if (reader != null && reader.isAlive() && reader.isRunning()) {
+ // add a dummy element to signal that no more splits
will
+ // arrive and wait until the reader finishes
+ reader.addSplit(EOF);
+ reader.join();
+ }
+ collector.close();
+ }
+
+ private class SplitReader<S extends Serializable, OT> extends Thread {
+
+ private volatile boolean isRunning;
+
+ private final FileInputFormat<OT> format;
+ private final TypeSerializer<OT> serializer;
+
+ private final Object checkpointLock;
+ private final TimestampedCollector<OT> collector;
+
+ private final Object lock = new Object();
+
+ private final Queue<FileInputSplit> pendingSplits;
+
+ SplitReader(FileInputFormat<OT> format,
+ TypeSerializer<OT> serializer,
+ TimestampedCollector<OT> collector,
+ Object checkpointLock) {
+
+ this.format = checkNotNull(format, "Unspecified
FileInputFormat.");
+ this.serializer = checkNotNull(serializer, "Unspecified
Serialized.");
+
+ this.pendingSplits = new LinkedList<>();
+ this.collector = collector;
+ this.checkpointLock = checkpointLock;
+ this.isRunning = true;
+ }
+
+ void addSplit(FileInputSplit split) {
+ Preconditions.checkNotNull(split);
+ synchronized (lock) {
+ this.pendingSplits.add(split);
+ }
+ }
+
+ public boolean isRunning() {
+ return this.isRunning;
+ }
+
+ @Override
+ public void run() {
+ FileInputSplit split = null;
+ try {
+ while (this.isRunning) {
+
+ // get the next split to read.
+ // locking is needed because
checkpointing will
+ // ask for a consistent snapshot of the
list.
+ synchronized (lock) {
--- End diff --
True, I didn't have that in mind. We could use a concurrent queue but this
is also fine.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
