kryo - 将Collections $ UnmodifiableCollection与Apache Flink一起使用

Question

在使用带有以下代码的Apache Flink时:

DataStream<List<String>> result = source.window(Time.of(1, TimeUnit.SECONDS)).mapWindow(new WindowMapFunction<String, List<String>>() {

    @Override
    public void mapWindow(Iterable<String> iterable, Collector<List<String>> collector) throws Exception {
        List<String> top5 = Ordering.natural().greatestOf(iterable, 5);
        collector.collect(top5);
    }
}).flatten();

我有这个异常(exception)

Caused by: java.lang.UnsupportedOperationException
    at java.util.Collections$UnmodifiableCollection.add(Collections.java:1055)
    at com.esotericsoftware.kryo.serializers.CollectionSerializer.read(CollectionSerializer.java:116)
    at com.esotericsoftware.kryo.serializers.CollectionSerializer.read(CollectionSerializer.java:22)
    at com.esotericsoftware.kryo.Kryo.readClassAndObject(Kryo.java:761)
    at org.apache.flink.api.java.typeutils.runtime.kryo.KryoSerializer.deserialize(KryoSerializer.java:211)
    at org.apache.flink.streaming.runtime.streamrecord.StreamRecordSerializer.deserialize(StreamRecordSerializer.java:110)
    at org.apache.flink.streaming.runtime.streamrecord.StreamRecordSerializer.deserialize(StreamRecordSerializer.java:41)
    at org.apache.flink.runtime.plugable.NonReusingDeserializationDelegate.read(NonReusingDeserializationDelegate.java:55)
    at org.apache.flink.runtime.io.network.api.serialization.SpillingAdaptiveSpanningRecordDeserializer.getNextRecord(SpillingAdaptiveSpanningRecordDeserializer.java:125)
    at org.apache.flink.streaming.runtime.io.StreamInputProcessor.processInput(StreamInputProcessor.java:127)
    at org.apache.flink.streaming.runtime.tasks.OneInputStreamTask.run(OneInputStreamTask.java:56)
    at org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:172)
    at org.apache.flink.runtime.taskmanager.Task.run(Task.java:581)
    at java.lang.Thread.run(Thread.java:745)

如何使用Flink使用 UnmodifiableCollection？

Answer 1

问题是Kryo的默认CollectionSerializer无法再次反序列化集合，因为它不可修改(.add()调用失败)。

要解决此问题，我们可以使用kryo-serializers项目中的UnmodifiableCollectionsSerializer。 Flink可传递地依赖于项目，因此无需将其添加为依赖项。

接下来，我们必须向Flink的Kryo实例注册序列化程序。

StreamExecutionEnvironment see = StreamExecutionEnvironment.getExecutionEnvironment();
Class<?> unmodColl = Class.forName("java.util.Collections$UnmodifiableCollection");
see.getConfig().addDefaultKryoSerializer(unmodColl, UnmodifiableCollectionsSerializer.class);

通常，我们不必调用 Class.forName()来注册序列化程序，但是在这种情况下， java.util.Collections$UnmodifiableCollection是程序包可见的，因此我们无法直接访问该类。