How to drain/delete/expire existing messages in Kafka

Some time you might have a bad record in Kafka topic that you want to delete. Kafka does not provide direct option to delete specific record. Only way to delete records is to expire them. You can achieve this by setting data retention to say 1 second that expires all the old messages. You can follow these steps

First check the topic to find out value of retention.ms config parameter for the topic

1. 
   kafka_2.11-0.11.0.1 spatil$ bin/kafka-configs.sh --zookeeper localhost:2181 --describe --entity-name my-topic --entity-type topics
   Configs for topic 'my-topic' are retention.ms=86400000
   

2. Change value of retention.ms to 1 which means all messages older than 1 ms will be expired

   
   kafka_2.11-0.11.0.1 spatil$ bin/kafka-configs.sh --zookeeper localhost:2181 --alter --entity-name my-topic --entity-type topics --add-config retention.ms=1
   Completed Updating config for entity: topic 'my-topic'
   

3. Wait for few seconds and monitor logs of the Kafka server to make sure that messages from the topic are deleted
4. Now change the value of retention.ms back to its original value which was 86400000 (7 days)

   
   kafka_2.11-0.11.0.1 spatil$ bin/kafka-configs.sh --zookeeper localhost:2181 --alter --entity-name my-topic --entity-type topics --add-config retention.ms=86400000
   Completed Updating config for entity: topic 'my-topic'.
   

5. You can verify that your changes are saved by running this command

   
   kafka_2.11-0.11.0.1 spatil$ bin/kafka-configs.sh --zookeeper localhost:2181 --describe --entity-name my-topic --entity-type topics
   Configs for topic 'my-topic' are retention.ms=86400000
   

Kafka how to reset number of partitions in a topics

I wanted to figure out how to reset number of partitions in my topic in Kafka and I followed these steps

1. I did create a sample topic called my-topic with single partition

   
   spatil$ bin/kafka-topics.sh --create --zookeeper localhost:2181 --topic my-topic --replication-factor 1 --partitions 1
   Created topic "my-topic".
   

2. I used describe command to verify that my topic has actually single partition

   
   spatil$ bin/kafka-topics.sh --zookeeper localhost:2181 --describe --topic my-topic
   Topic:my-topic PartitionCount:1 ReplicationFactor:1 Configs:
    Topic: my-topic Partition: 0 Leader: 0 Replicas: 0 Isr: 0
   

3. Then I did execute alter command on my partition and changed number of partitions from 1 to 3

   
   spatil$ bin/kafka-topics.sh --zookeeper localhost:2181 --alter --topic my-topic --partitions 3
   WARNING: If partitions are increased for a topic that has a key, the partition logic or ordering of the messages will be affected
   Adding partitions succeeded!
   

4. I did execute describe command on my topic to verify that it actually has 3 topics

   
   spatil$ bin/kafka-topics.sh --zookeeper localhost:2181 --describe --topic my-topic
   Topic:my-topic PartitionCount:3 ReplicationFactor:1 Configs:
    Topic: my-topic Partition: 0 Leader: 0 Replicas: 0 Isr: 0
    Topic: my-topic Partition: 1 Leader: 0 Replicas: 0 Isr: 0
    Topic: my-topic Partition: 2 Leader: 0 Replicas: 0 Isr: 0
   

How to reset consumer group offset

First run describe on topic to check what it the current LAG its zero in this case

bin/kafka-consumer-groups.sh --bootstrap-server localhost:9092 --describe --group user.kafkaconsumer

Now run following command to just preview what will be the next offset if you reset

bin/kafka-consumer-groups.sh --bootstrap-server localhost:9092 --group user.kafkaconsumer --reset-offsets --to-earliest --all-topics

Shutdown the consumer so that you can restart the consumer Now go back/reset the offset so that it goes back to first message 3bin/kafka-consumer-groups.sh --bootstrap-server localhost:9092 --group user.kafkaconsumer --reset-offsets --to-earliest --all-topics --execute Go back and verify that the consumer offset actually went back by executing following command

bin/kafka-consumer-groups.sh --bootstrap-server localhost:9092 --describe --group user.kafkaconsumer

You should be able to see the offset set back to 22000 which is start of first message in Kafka.

Spark program to read data from RDBMS

I wanted to figure out how to connect to RDBMS from spark and extract data, so i followed these steps. You can download this project form github
First i did create Address table in my local mysql like this

CREATE TABLE `address` (
  `addressid` int(11) NOT NULL AUTO_INCREMENT,
  `contactid` int(11) DEFAULT NULL,
  `line1` varchar(300) NOT NULL,
  `city` varchar(50) NOT NULL,
  `state` varchar(50) NOT NULL,
  `zip` varchar(50) NOT NULL,
  `lastmodified` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
  PRIMARY KEY (`addressid`),
  KEY `contactid` (`contactid`),
  CONSTRAINT `address_ibfk_1` FOREIGN KEY (`contactid`) REFERENCES `CONTACT` (`contactid`)
) ENGINE=InnoDB AUTO_INCREMENT=6 DEFAULT CHARSET=utf8;

Then i did add 5 sample records to the address table. When i query address table on my local this is what i get

After that i did create a Spark Scala project that has mysql-connector-java as one of the dependencies

	<?xml version="1.0" encoding="UTF-8"?>
	<project xmlns="http://maven.apache.org/POM/4.0.0"
	xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
	xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
	<modelVersion>4.0.0</modelVersion>
	<groupId>com.spnotes.spark</groupId>
	<artifactId>JDBCSpark</artifactId>
	<version>1.0-SNAPSHOT</version>
	<properties>
	<maven.compiler.target>1.7</maven.compiler.target>
	<encoding>UTF-8</encoding>
	<scala.tools.version>2.10</scala.tools.version>
	<scala.version>2.10.4</scala.version>
	<spark.version>1.5.2</spark.version>
	</properties>
	<dependencies>
	<dependency>
	<groupId>org.scala-lang</groupId>
	<artifactId>scala-library</artifactId>
	<version>${scala.version}</version>
	</dependency>
	<dependency>
	<groupId>org.apache.spark</groupId>
	<artifactId>spark-core_${scala.tools.version}</artifactId>
	<version>${spark.version}</version>
	</dependency>
	<dependency>
	<groupId>org.apache.spark</groupId>
	<artifactId>spark-hive_${scala.tools.version}</artifactId>
	<version>${spark.version}</version>
	</dependency>
	<dependency>
	<groupId>mysql</groupId>
	<artifactId>mysql-connector-java</artifactId>
	<version>5.1.38</version>
	</dependency>
	</dependencies>
	<build>
	<sourceDirectory>src/main/scala</sourceDirectory>
	<plugins>
	<plugin>
	<groupId>org.scala-tools</groupId>
	<artifactId>maven-scala-plugin</artifactId>
	<version>2.15.2</version>
	<executions>
	<execution>
	<goals>
	<goal>compile</goal>
	</goals>
	</execution>
	</executions>
	</plugin>
	</plugins>
	</build>
	</project>

view raw pom.xml hosted with ❤ by GitHub

The last step was to create a simple Spark program like this,

	package com.spnotes.spark
	import java.sql.{Connection, DriverManager, ResultSet}
	import org.apache.spark.rdd.JdbcRDD
	import org.apache.spark.{SparkConf, SparkContext}
	/**
	* Created by sunilpatil on 4/19/16.
	*/
	object JDBCRDDClient {
	case class Address(addressId: Int, contactId: Int, line1: String, city: String, state: String, zip: String)
	def main(argv: Array[String]): Unit = {
	val sparkConf = new SparkConf().setMaster("local[2]").setAppName("HelloJDBC")
	val sparkContext = new SparkContext(sparkConf)
	val jdbcRdd = new JdbcRDD(sparkContext, getConnection,
	"select * from address limit ?,?",
	0, 5, 1, convertToAddress)
	jdbcRdd.foreach(println)
	}
	def getConnection(): Connection = {
	Class.forName("com.mysql.jdbc.Driver")
	DriverManager.getConnection("jdbc:mysql://localhost/test1?" + "user=test1&password=test1")
	}
	def convertToAddress(rs: ResultSet): Address = {
	new Address(rs.getInt("addressid"), rs.getInt("contactid"), rs.getString("line1"),
	rs.getString("city"), rs.getString("state"), rs.getString("zip"))
	}
	}

view raw JDBCRDDClient.scala hosted with ❤ by GitHub

My program has 4 main sections

1. First is Address as case class with same schema as that of Address table, without lastmodified field
2. Next is this call to create object of JdbcRDD that says query everything from address with addressid between 1 and 5. new JdbcRDD(sparkContext, getConnection, "select * from address limit ?,?", 0, 5, 1, convertToAddress)
3. Then i did define getConnection() method that creates JDBC connection to my database and returns it
4. Last is the convertToAddress() method that knows how to take a ResultSet and convert it into object of Address

When i run this program in IDE this is the output i get

How to implement cache (LRU Cache) using LinkedHashMap in java

Recently i wanted to implement a simple Least recently used (LRU) cache in one my applications. But my use case is simple enough that instead of going for something ehcache i decided to build it on own by using java.util.LinkedHashMap
As you can see from the code its very simple. All you have to do is extend java.util.LinkedHashMap and override its protected removeEldestEntry() method so that it checks if the size of map is greater than a size you specified while creating the Map if yes remove the eldest entry

	import java.util.Collections;
	import java.util.Iterator;
	import java.util.LinkedHashMap;
	import java.util.Map;
	public class LRUCache<Key,Value> extends LinkedHashMap<Key,Value>{
	private int cacheSize;
	public LRUCache(int cacheSize) {
	super(cacheSize, 0.75f,true);
	this.cacheSize = cacheSize;
	}
	@Override
	protected boolean removeEldestEntry(Map.Entry<Key, Value> eldest) {
	return size() > cacheSize;
	}
	public static void main(String[] argv){
	LRUCache<String, String> cache = new LRUCache<String,String>(3);
	Map<String,String> map = Collections.synchronizedMap(cache);
	cache.put("First","1");
	cache.put("Second","2");
	cache.put("Third","3");
	System.out.println(cache);
	cache.get("First");
	cache.get("First");
	cache.put("Fourth","4");
	System.out.println(cache);
	cache.put("Fifth","5");
	System.out.println(cache);
	}
	}

view raw LRUCache.java hosted with ❤ by GitHub

Now the question is when Map is full which entry will it remove, you have 2 options

1. Eldest: If you just want to remove the first entry that you inserted in the Map when adding a new entry then in your constructor you could use super(cacheSize, 0.75f);, so LinkedHashMap wont keep track of when a particular entry were accessed.
2. Least recently used (LRU): But if you want to make sure that the entry that was least recently used should be removed then call super(cacheSize, 0.75f, true); from constructor of your LRUCache so that LinkedHashMap keeps track of when entry was accessed and removes the Least recently used entry

Spark Streaming Kafka 10 API Word Count application Scala

In Spark Kafka Streaming Java program Word Count using Kafka 0.10 API blog entry i talked about how you create a simple java program that uses Spark Streaming's Kafka10 API using Java. This blog entry does the same thing but using Scala. You can download the complete application from github
You can run this sample by first downloading Kafka 0.10.* from Apache Kafka WebSite, then you can create and start a test topic and send messages to it by following this Kafka Quick start document

	package com.spnotes.spark
	import org.apache.kafka.clients.consumer.ConsumerConfig
	import org.apache.spark.SparkConf
	import org.apache.spark.streaming.kafka010.{ConsumerStrategies, KafkaUtils, LocationStrategies}
	import org.apache.spark.streaming.{Durations, StreamingContext}
	import scala.collection.mutable
	/**
	* Created by sunilpatil on 1/11/17.
	*/
	object Kafka10 {
	def main(argv: Array[String]): Unit = {
	// Configure Spark to connect to Kafka running on local machine
	val kafkaParam = new mutable.HashMap[String, String]()
	kafkaParam.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, "localhost:9092")
	kafkaParam.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG,
	"org.apache.kafka.common.serialization.StringDeserializer")
	kafkaParam.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG,
	"org.apache.kafka.common.serialization.StringDeserializer")
	kafkaParam.put(ConsumerConfig.GROUP_ID_CONFIG, "group1")
	kafkaParam.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, "latest")
	kafkaParam.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG, "true")
	val conf = new SparkConf().setMaster("local[2]").setAppName("Kafka10")
	//Read messages in batch of 30 seconds
	val sparkStreamingContext = new StreamingContext(conf, Durations.seconds(30))
	//Configure Spark to listen messages in topic test
	val topicList = List("test")
	// Read value of each message from Kafka and return it
	val messageStream = KafkaUtils.createDirectStream(sparkStreamingContext,
	LocationStrategies.PreferConsistent,
	ConsumerStrategies.Subscribe[String, String](topicList, kafkaParam))
	val lines = messageStream.map(consumerRecord => consumerRecord.value().asInstanceOf[String])
	// Break every message into words and return list of words
	val words = lines.flatMap(_.split(" "))
	// Take every word and return Tuple with (word,1)
	val wordMap = words.map(word => (word, 1))
	// Count occurance of each word
	val wordCount = wordMap.reduceByKey((first, second) => first + second)
	//Print the word count
	wordCount.print()
	sparkStreamingContext.start()
	sparkStreamingContext.awaitTermination()
	}
	}

view raw Kafka10.scala hosted with ❤ by GitHub

Spark Kafka Streaming Java program Word Count using Kafka 0.10 API

Kafka API went through a lot of changes starting Kafka 0.9. Spark Kafka Streaming API also was changed to better support Kafka 0.9. i wanted to try that out so i built this simple Word Count application using Kafka 0.10 API. This blog entry does the same thing but using Scala. You can download the complete application from github
You can run this sample by first downloading Kafka 0.10.* from Apache Kafka WebSite, then you can create and start a test topic and send messages to it by following this Kafka Quick start document First thing i did was to include Kafka 0.10 API dependencies for the Spark Project. As you can see i am using Spark 2.1 version

	<dependency>
	<groupId>org.apache.spark</groupId>
	<artifactId>spark-core_2.11</artifactId>
	<version>2.1.0</version>
	</dependency>
	<dependency>
	<groupId>org.apache.spark</groupId>
	<artifactId>spark-streaming_2.11</artifactId>
	<version>2.1.0</version>
	</dependency>
	<dependency>
	<groupId>org.apache.spark</groupId>
	<artifactId>spark-streaming-kafka-0-10_2.11</artifactId>
	<version>2.1.0</version>
	</dependency>

view raw pom.xml hosted with ❤ by GitHub

Then i did create a SparkKafka10.java file that looks like this. Please take a look at comments inside the code for what i am doing. Now if you create test topic and send messages to it, you should be able to see the wordcount on console

	package com.test;
	import com.test.schema.ContactType;
	import org.apache.kafka.clients.consumer.ConsumerConfig;
	import org.apache.kafka.clients.consumer.ConsumerRecord;
	import org.apache.spark.SparkConf;
	import org.apache.spark.api.java.JavaPairRDD;
	import org.apache.spark.api.java.function.*;
	import org.apache.spark.streaming.Durations;
	import org.apache.spark.streaming.api.java.JavaDStream;
	import org.apache.spark.streaming.api.java.JavaInputDStream;
	import org.apache.spark.streaming.api.java.JavaPairDStream;
	import org.apache.spark.streaming.api.java.JavaStreamingContext;
	import org.apache.spark.streaming.kafka010.ConsumerStrategies;
	import org.apache.spark.streaming.kafka010.KafkaUtils;
	import org.apache.spark.streaming.kafka010.LocationStrategies;
	import scala.Tuple2;
	import java.util.*;
	/**
	* Created by sunilpatil on 1/11/17.
	*/
	public class SparkKafka10 {
	public static void main(String[] argv) throws Exception{
	// Configure Spark to connect to Kafka running on local machine
	Map<String, Object> kafkaParams = new HashMap<>();
	kafkaParams.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG,"localhost:9092");
	kafkaParams.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG,
	"org.apache.kafka.common.serialization.StringDeserializer");
	kafkaParams.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG,
	"org.apache.kafka.common.serialization.StringDeserializer");
	kafkaParams.put(ConsumerConfig.GROUP_ID_CONFIG,"group1");
	kafkaParams.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG,"latest");
	kafkaParams.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG,true);
	//Configure Spark to listen messages in topic test
	Collection<String> topics = Arrays.asList("test");
	SparkConf conf = new SparkConf().setMaster("local[2]").setAppName("SparkKafka10WordCount");
	//Read messages in batch of 30 seconds
	JavaStreamingContext jssc = new JavaStreamingContext(conf, Durations.seconds(30));
	// Start reading messages from Kafka and get DStream
	final JavaInputDStream<ConsumerRecord<String, String>> stream =
	KafkaUtils.createDirectStream(jssc, LocationStrategies.PreferConsistent(),
	ConsumerStrategies.<String,String>Subscribe(topics,kafkaParams));
	// Read value of each message from Kafka and return it
	JavaDStream<String> lines = stream.map(new Function<ConsumerRecord<String,String>, String>() {
	@Override
	public String call(ConsumerRecord<String, String> kafkaRecord) throws Exception {
	return kafkaRecord.value();
	}
	});
	// Break every message into words and return list of words
	JavaDStream<String> words = lines.flatMap(new FlatMapFunction<String, String>() {
	@Override
	public Iterator<String> call(String line) throws Exception {
	return Arrays.asList(line.split(" ")).iterator();
	}
	});
	// Take every word and return Tuple with (word,1)
	JavaPairDStream<String,Integer> wordMap = words.mapToPair(new PairFunction<String, String, Integer>() {
	@Override
	public Tuple2<String, Integer> call(String word) throws Exception {
	return new Tuple2<>(word,1);
	}
	});
	// Count occurance of each word
	JavaPairDStream<String,Integer> wordCount = wordMap.reduceByKey(new Function2<Integer, Integer, Integer>() {
	@Override
	public Integer call(Integer first, Integer second) throws Exception {
	return first+second;
	}
	});
	//Print the word count
	wordCount.print();
	jssc.start();
	jssc.awaitTermination();
	}
	}

view raw SparkKafka10.java hosted with ❤ by GitHub