Create a Realtime Data Stream with Apache Storm

Apache Storm is a distributed stream processing computation framework. It integrates with various systems and libraries, including Apache Kafka, Apache Cassandra, Redis, Amazon Kinesis, Kestrel, RabbitMQ/Advanced Message Queuing Protocol (AMQP), and Java Message Service (JMS).

This guide discusses key concepts and terminology associated with Apache Storm, and includes instructions for using it to create a real-time data stream.

Apache Storm Components & Terminology

• Tuples: In Storm, tuples serve as the main data structure. Tuples are named lists that can contain any type of values and are dynamically typed. They have helper methods to retrieve field values without casting. By default, Storm can serialize tuple values for primitives, strings, and byte arrays, though a custom serializer is needed for other types.

• Topologies: A Storm topology is the logical framework for a real-time application. It functions similarly to a MapReduce job, but is designed to run indefinitely. Topologies are composed of spouts and bolts connected by stream groupings, as illustrated in the diagram below:

• Streams: A stream is a sequence of tuples processed in parallel, capable of containing various data types. Each stream has a defined schema that specifies the fields within its tuples and can be assigned a unique ID.

• Spouts: A spout acts as a source of streams in a topology. It reads tuples from an external source and emits them into the topology. Spouts are categorized as either “reliable” or “unreliable”. Reliable spouts have the capability to replay a tuple if it fails to be processed, whereas unreliable spouts do not retain the tuple after it is emitted.

• Bolts: Bolts perform the processing tasks in Storm, including filtering, aggregations, joins, and interactions with databases. Bolts can transform streams, and complex transformations often involve multiple bolts working together.

• Nodes: Apache Storm consists of two primary types of nodes:

  • Nimbus (Master Node): The central component responsible for running Storm topologies. Nimbus analyzes the topology and determines which tasks to execute.

  • Supervisor (Worker Node): This node manages worker processes, monitors their health, ensures they are running correctly, and restarts them if necessary. It receives heartbeats from workers, reports to Nimbus, distributes tasks, and facilitates communication between workers and Nimbus. The Supervisor acts as a bridge between Nimbus and workers, ensuring tasks are executed correctly and keeping Nimbus updated on task status.

• ZooKeeper: Apache ZooKeeper acts as a centralized server that helps manage and coordinate services for distributed applications. In this guide, ZooKeeper is installed on a separate, single node, however ZooKeeper clusters can be configured for larger workloads.

Before You Begin

1. Using the instructions in our Creating a Compute Instance guide and specifications below, create the four necessary instances to run an Apache Storm cluster (one for ZooKeeper, one for Nimbus, and two Storm Supervisors):

   • Images: Use the latest Long Term Support (LTS) version of Ubuntu available for all nodes. The examples in this guide use Ubuntu 24.04 LTS.

   • Region: Choose the geographic region best suited for your use case. The examples in this guide use the Miami (us-mia) region.

   • Linode Plan: Below are the minimum specifications recommended for each node:

     • ZooKeeper Node: Linode Shared 2 GB
     • Nimbus Node: Linode Shared 4 GB
     • Storm Supervisor Node 1: Linode Shared 4 GB
     • Storm Supervisor Node 2: Linode Shared 4 GB

   • Linode Label: Enter a descriptive label for each instance. The examples in this guide use the following:

     • ZooKeeper Node: storm-zoo
     • Nimbus Node: storm-nimbus
     • Storm Supervisor Node 1: storm-super-1
     • Storm Supervisor Node 2: storm-super-2

2. Once deployed, follow our Setting Up and Securing a Compute Instance guide to update your system. You may also wish to set the timezone, configure your hostname, create a limited user account, and harden SSH access.

Install Java On All Nodes

Both Python and Java are prerequisites for Storm and ZooKeeper. Ubuntu 24.04 LTS comes with Python version 3.12.3 installed by default, but Java must be installed manually.

On each deployed instance, follow the steps below to install Java.

1. Install the Java Development Kit (JDK):

   storm-zoo, storm-nimbus, storm-super-1, and storm-super-2

   sudo apt install default-jdk

2. Display the installed version to verify the Java installation:

   storm-zoo, storm-nimbus, storm-super-1, and storm-super-2

   java --version

   openjdk 21.0.4 2024-07-16
   OpenJDK Runtime Environment (build 21.0.4+7-Ubuntu-1ubuntu224.04)
   OpenJDK 64-Bit Server VM (build 21.0.4+7-Ubuntu-1ubuntu224.04, mixed mode, sharing)

Install ZooKeeper

Follow the steps in this section to install ZooKeeper on the storm-zoo instance.

1. Use wget to download the latest stable version of ZooKeeper available on the ZooKeeper release page. The examples in this guide use ZooKeeper version 3.8.4. If necessary, adjust the commands below according to your version:

   storm-zoo

   wget https://dlcdn.apache.org/zookeeper/zookeeper-3.8.4/apache-zookeeper-3.8.4-bin.tar.gz

2. Unpack the ZooKeeper tarball:

   storm-zoo

   tar -zxvf apache-zookeeper-3.8.4-bin.tar.gz

3. Create a zoo.cfg ZooKeeper configuration file:

   storm-zoo

   nano ~/apache-zookeeper-3.8.4-bin/conf/zoo.cfg

   Give the file the following content, and save your changes:

   File: /opt/zookeeper/conf/zoo.cfg

   1 2 3

   tickTime=2000 dataDir=~/apache-zookeeper-3.8.4-bin/data clientPort=2181

4. Create the ~/apache-zookeeper-3.8.4-bin/data directory for data storage:

   storm-zoo

   mkdir ~/apache-zookeeper-3.8.4-bin/data

5. Start ZooKeeper:

   storm-zoo

   ~/apache-zookeeper-3.8.4-bin/bin/zkServer.sh start

   /usr/bin/java
   ZooKeeper JMX enabled by default
   Using config: /home/USERNAME/apache-zookeeper-3.8.4-bin/bin/../conf/zoo.cfg
   Starting zookeeper ... STARTED

Set Up Storm on Nimbus and Supervisor Instances

This section provides instructions for downloading and installing Apache Storm on the storm-nimbus, storm-super-1, and storm-super-2 instances.

1. Download the latest Storm release on each specified instance. This guide uses Storm version 2.7.0. If necessary, adjust the commands for your version of Storm.

   storm-nimbus, storm-super-1, and storm-super-2

   wget https://dlcdn.apache.org/storm/apache-storm-2.7.0/apache-storm-2.7.0.tar.gz

2. Once the Storm tarballs are present, unpack them:

   storm-nimbus, storm-super-1, and storm-super-2

   tar -zxvf apache-storm-2.7.0.tar.gz

3. Open your .bashrc file:

   storm-nimbus, storm-super-1, and storm-super-2

   nano ~/.bashrc

   Append the following lines the end of the file to add Storm’s bin directory to your PATH. Save your changes when complete:

   File: ~/.bashrc

   1 2	#Set PATH for Storm export PATH="$HOME/apache-storm-2.7.0/bin:$PATH"

   Apply the changes to .bashrc:

   storm-nimbus, storm-super-1, and storm-super-2

   source ~/.bashrc

4. Verify the Storm installation by checking the version:

   storm-nimbus, storm-super-1, and storm-super-2

   storm version

   Running: java -client -Ddaemon.name= -Dstorm.options= -Dstorm.home=/home/USERNAME/apache-storm-2.7.0 -Dstorm.log.dir=/home/USERNAME/apache-storm-2.7.0/logs -Djava.library.path=/usr/local/lib:/opt/local/lib:/usr/lib:/usr/lib64 -Dstorm.conf.file= -cp /home/USERNAME/apache-storm-2.7.0/*:/home/USERNAME/apache-storm-2.7.0/lib/*:/home/USERNAME/apache-storm-2.7.0/extlib/*:/home/USERNAME/apache-storm-2.7.0/extlib-daemon/*:/home/USERNAME/apache-storm-2.7.0/conf org.apache.storm.utils.VersionInfo
   Storm 2.7.0
   URL https://TOKEN@github.com/apache/storm.git -r b95a7f25a114ae7fb9c23cbc2979d3cfff09fa73
   Branch v2.7.0
   Compiled by rui on 2024-10-11T17:28Z
   From source with checksum dcefb62616ea3f989d583d962257084

Edit the Storm Configuration on Nimbus and Supervisor Instances

The storm.yaml configuration file specifies the local directory for Storm’s operation and defines the connection settings for ZooKeeper and Nimbus.

1. Open the storm.yaml file on the storm-nimbus, storm-super-1, and storm-super-2 instances:

   storm-nimbus, storm-super-1, and storm-super-2

   nano ~/apache-storm-2.7.0/conf/storm.yaml

   Add the following lines to the end of the file. Replace storm-zoo_IP_ADDRESS with the IP address for your storm-zoo instance and storm-nimbus_IP_ADDRESS with the IP address for your storm-nimbus instance.

   File: conf/storm.yaml

   1 2 3 4

   storm.zookeeper.servers: - "storm-zoo_IP_ADDRESS" nimbus.seeds: ["storm-nimbus_IP_ADDRESS"] storm.local.dir: "~/apache-data"

   This defines the IP addresses of the ZooKeeper and Nimbus instances and sets the /var/storm directory for Storm’s state and temporary files. Refer to the Storm Setup Guide for more details.

   When done, save your changes.

2. On your storm-zoo instance, create the ~/storm-data directory for Storm’s application data:

   storm-zoo

   mkdir ~/storm-data

Start Storm

1. On the storm-super-1 and storm-super-2 instances, execute the following command to run Storm as a Supervisor daemon:

   storm-super-1 & storm-super-2

   storm supervisor &

2. On the storm-nimbus instance, execute the following commands to run Storm as a nimbus daemon and run the Storm UI web server:

   storm-nimbus

   storm nimbus &
   storm ui &

3. Open a web browser on your local machine and navigate to port 8080 of the storm-nimbus instance’s public IP address:

   Local machine

   http://storm-nimbus_IP_ADDRESS:8080

   The web site at that address shows the status of the cluster:

Create the Message Stream

The site above shows no topologies, and therefore has no message streams. The storm-starter sample, located in the apache-storm-2.7.0/examples/ directory, can solve this. It contains a variety of Storm topologies, including one titled WordCount that is used as an example below.

1. On the storm-nimbus instance, use wget to download the latest version of Apache Maven. This guide uses Maven version 3.9.9.

   storm-nimbus

   wget https://dlcdn.apache.org/maven/maven-3/3.9.9/binaries/apache-maven-3.9.9-bin.tar.gz

2. Unpack the downloaded archive file:

   storm-nimbus

   tar xzvf apache-maven-3.9.9-bin.tar.gz

3. Open your .bashrc file:

   storm-nimbus

   nano ~/.bashrc

   Add Maven’s bin directory to your PATH, and save your changes:

   File: ~/.bashrc

   1 2	#Set PATH for Maven export PATH="$HOME/apache-maven-3.9.9/bin:$PATH"

4. Apply the changes to .bashrc:

   storm-nimbus

   source ~/.bashrc

5. Verify the Maven installation by checking the version:

   storm-nimbus

   mvn -v

   Apache Maven 3.9.9 (8e8579a9e76f7d015ee5ec7bfcdc97d260186937)
   Maven home: /home/USERNAME/apache-maven-3.9.9
   Java version: 21.0.4, vendor: Ubuntu, runtime: /usr/lib/jvm/java-21-openjdk-amd64
   Default locale: en_US, platform encoding: UTF-8
   OS name: "linux", version: "6.8.0-47-generic", arch: "amd64", family: "unix"

6. Issue the following commands to change into the ~/apache-storm-2.7.0/examples/storm-starter directory and build a Storm “uber JAR”:

   storm-nimbus

   cd ~/apache-storm-2.7.0/examples/storm-starter
   mvn package

   After a few minutes, you should see a success message towards the bottom of the output:

   ...
   [INFO] BUILD SUCCESS
   ...

   The JAR file you built is located at ~/apache-storm-2.7.0/examples/storm-starter/target/storm-starter-2.7.0.jar.

Submit the Example Topology

The storm jar command allows you to choose a topology from the “uber JAR” and submit that to your cluster.

1. On the storm-nimbus instance, use the following command to submit the WordCount topology:

   storm-nimbus

   storm jar ~/apache-storm-2.7.0/examples/storm-starter/target/storm-starter-2.7.0.jar org.apache.storm.starter.WordCountTopology WordCount

2. On your local machine, return to your web browser and refresh the page located at port 8080 of the storm-nimbus instance:

   Local machine

   http://storm-nimbus_IP_ADDRESS:8080

   The Storm UI page should now show WordCount under the Topology Summary section: