背景

GenericWriteAheadSink是可以用于几乎是精准一次输出的场景，为什么说是几乎精准一次呢？我们从源码的角度分析一下

GenericWriteAheadSink做不到精准一次输出的原因

首先我们看一下flink检查点完成后通知GenericWriteAheadSink开始进行分段的记录输出并提交事务的代码

pubblic void notifyCheckpointComplete(long checkpointId) throws Exception {super.notifyCheckpointComplete(checkpointId);synchronized (pendingCheckpoints) {Iterator<PendingCheckpoint> pendingCheckpointIt = pendingCheckpoints.iterator();while (pendingCheckpointIt.hasNext()) {PendingCheckpoint pendingCheckpoint = pendingCheckpointIt.next();long pastCheckpointId = pendingCheckpoint.checkpointId;int subtaskId = pendingCheckpoint.subtaskId;long timestamp = pendingCheckpoint.timestamp;StreamStateHandle streamHandle = pendingCheckpoint.stateHandle;if (pastCheckpointId <= checkpointId) {try {if (!committer.isCheckpointCommitted(subtaskId, pastCheckpointId)) {try (FSDataInputStream in = streamHandle.openInputStream()) {//开始把分段记录列表的记录进行输出boolean success =sendValues(new ReusingMutableToRegularIteratorWrapper<>(new InputViewIterator<>(new DataInputViewStreamWrapper(in),serializer),serializer),pastCheckpointId,timestamp);if (success) {//把分段记录列表输出成功后提交事务committer.commitCheckpoint(subtaskId, pastCheckpointId);streamHandle.discardState();pendingCheckpointIt.remove();}}} else {streamHandle.discardState();pendingCheckpointIt.remove();}} catch (Exception e) {// we have to break here to prevent a new (later) checkpoint// from being committed before this oneLOG.error("Could not commit checkpoint.", e);break;}}}}}

从上面的源码可以看出，sendValue方法和提交事务commitCheckpoint方法并不能保证原子性，这就意味着如果sendValue执行了一部分或者全部，而提交事务方法commitCheckpoint失败，那么此时这个检查点对应的事务相当于就没有完成，在下一个检查点的通知消息中，会把历史检查点的事务重新sendValue然后进行commit一次，这就意味着相同的记录会执行两次sendValue操作，这就是GenericWriteAheadSink不能保证精准一次的原因