Github user tdas commented on a diff in the pull request:
https://github.com/apache/spark/pull/1434#discussion_r15724923
--- Diff:
extras/kinesis-asl/src/main/scala/org/apache/spark/streaming/kinesis/KinesisRecordProcessor.scala
---
@@ -0,0 +1,214 @@
+/*
+ * 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.spark.streaming.kinesis
+
+import java.util.List
+
+import scala.collection.JavaConversions.asScalaBuffer
+import scala.util.Random
+
+import org.apache.spark.Logging
+
+import
com.amazonaws.services.kinesis.clientlibrary.exceptions.InvalidStateException
+import
com.amazonaws.services.kinesis.clientlibrary.exceptions.KinesisClientLibDependencyException
+import
com.amazonaws.services.kinesis.clientlibrary.exceptions.ShutdownException
+import
com.amazonaws.services.kinesis.clientlibrary.exceptions.ThrottlingException
+import
com.amazonaws.services.kinesis.clientlibrary.interfaces.IRecordProcessor
+import
com.amazonaws.services.kinesis.clientlibrary.interfaces.IRecordProcessorCheckpointer
+import com.amazonaws.services.kinesis.clientlibrary.types.ShutdownReason
+import com.amazonaws.services.kinesis.model.Record
+
+/**
+ * Kinesis-specific implementation of the Kinesis Client Library (KCL)
IRecordProcessor.
+ * This implementation operates on the Array[Byte] from the
KinesisReceiver.
+ * The Kinesis Worker creates an instance of this KinesisRecordProcessor
upon startup.
+ *
+ * @param receiver Kinesis receiver
+ * @param workerId for logging purposes
+ * @param checkpointState represents the checkpoint state including the
next checkpoint time.
+ * It's injected here for mocking purposes.
+ */
+private[kinesis] class KinesisRecordProcessor(
+ receiver: KinesisReceiver,
+ workerId: String,
+ checkpointState: KinesisCheckpointState) extends IRecordProcessor with
Logging {
+
+ /** shardId to be populated during initialize() */
+ var shardId: String = _
+
+ /**
+ * The Kinesis Client Library calls this method during IRecordProcessor
initialization.
+ *
+ * @param shardId assigned by the KCL to this particular RecordProcessor.
+ */
+ override def initialize(shardId: String) {
+ logInfo(s"Initialize: Initializing workerId $workerId with shardId
$shardId")
+ this.shardId = shardId
+ }
+
+ /**
+ * This method is called by the KCL when a batch of records is pulled
from the Kinesis stream.
+ * This is the record-processing bridge between the KCL's
IRecordProcessor.processRecords()
+ * and Spark Streaming's Receiver.store().
+ *
+ * @param batch list of records from the Kinesis stream shard
+ * @param checkpointer used to update Kinesis when this batch has been
processed/stored
+ * in the DStream
+ */
+ override def processRecords(batch: List[Record], checkpointer:
IRecordProcessorCheckpointer) {
+ if (!receiver.isStopped()) {
+ try {
+ /**
+ * Note: If we try to store the raw ByteBuffer from
record.getData(), the Spark Streaming
+ * Receiver.store(ByteBuffer) attempts to deserialize the
ByteBuffer using the
+ * internally-configured Spark serializer (kryo, etc).
+ * This is not desirable, so we instead store a raw Array[Byte]
and decouple
+ * ourselves from Spark's internal serialization strategy.
+ */
+ batch.foreach(record =>
+
KinesisRecordProcessor.retry(receiver.store(record.getData().array()), 4, 500)
--- End diff --
Backing off by 500 here is a probably a bad idea. Maybe reduce it to 10 ms.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
