Edit

Installing Using Docker

Considerations

In-memory vs. Persistent Cluster

When deploying a persistent Ignite cluster, you should always mount a persistent volume or local directory. If you do not use a persistent volume, Ignite will store the data in the container’s file system. It means that the data will be erased when you remove the container. To save the data in a permanent location, mount a persistent volume.

Networking

By default, Ignite Docker image exposes the following ports: 11211, 47100, 47500, 49112. You can expose more ports as needed by adding -p <port> to the docker run command. For example, to connect a thin client to the node running inside a docker container, open port 10800:

docker run -d -p 10800:10800 apacheignite/ignite

Downloading Ignite Docker Image

Assuming that you already have Docker installed on your machine, you can pull and run the Ignite Docker image using the following commands.

Open a command shell and use the following command to pull the Ignite Docker image.

# Pull latest version
sudo docker pull apacheignite/ignite

By default, the latest version is downloaded, but you can download a specific version too.

# Pull a specific Ignite version
sudo docker pull apacheignite/ignite:2.16.0

Running In-Memory Cluster

Run Ignite in a docker container using the docker run command.

# Run the latest version
sudo docker run -d apacheignite/ignite

This command will launch a single Ignite node.

To run a specific version of Ignite, use the following command:

# Run a specific Ignite version
sudo docker run -d apacheignite/ignite:2.16.0

Running Persistent Cluster

If you use Native Persistence, Ignite stores the user data under the default work directory ({IGNITE_HOME}/work) in the file system of the container. This directory will be erased if you restart the docker container. To avoid this, you can:

Use a persistent volume to store the data; or
Mount a local directory

These two options are described in the following sections.

Using Persistent Volume

To create a persistent volume, run the following command:

sudo docker volume create persistence-volume

We will mount this volume to a specific directory when running the Ignite docker image. This directory will have to be passed to Ignite. This can be done in two ways:

Using the IGNITE_WORK_DIR system property
In the node configuration file

The following command launches the Ignite Docker image and passes the work directory to Ignite via the system property:

docker run -d \
  -v storage-volume:/storage \
  -e IGNITE_WORK_DIR=/storage \
  apacheignite/ignite

Using Local Directory

Instead of creating a volume, you can mount a local directory to the container in which the Ignite image is running and use this directory to store persistent data. When restarting the container with the same command, Ignite will load the data from the directory.

mkdir work_dir

docker run -d \
  -v ${PWD}/work_dir:/storage \
  -e IGNITE_WORK_DIR=/storage \
  apacheignite/ignite

The -v option mounts a local directory under the /storage path in the container. The -e IGNITE_WORK_DIR=/storage option tells Ignite to use this folder as the work directory.

Providing Configuration File

When you run the image, it starts a node with the default configuration file. You can pass a custom configuration file by using the CONFIG_URI environment variable:

docker run -d \
  -e CONFIG_URI=http://myserver/config.xml  \
  apacheignite/ignite

You can also use a file from your local file system. You need to mount the file first under a specific path inside the container by using the -v option. Then, use this path in the CONFIG_URI option:

docker run -d \
  -v /local/dir/config.xml:/config-file.xml \
  -e CONFIG_URI=/config-file.xml \
  apacheignite/ignite

Deploying User Libraries

When starting, a node adds all libraries found in the {IGNITE_HOME}/libs directory to the classpath (ignoring the "optional" directory). If you want to deploy user libraries, you can mount a directory from your local machine to a path in the /opt/ignite/apache-ignite/libs/ in the container by using the -v option.

The following command mounts a directory on your machine to libs/user_libs in the container. All files located in the directory are added to the classpath of the node.

docker run -v /local_path/to/dir_with_libs/:/opt/ignite/apache-ignite/libs/user_libs apacheignite/ignite

Another option is to use the EXTERNAL_LIBS variable if your libraries are available via an URL.

docker run -e "EXTERNAL_LIBS=http://url_to_your_jar" apacheignite/ignite

Enabling Modules

To enable specific modules, specify their names in the "OPTION_LIBS" system variable as follows:

sudo docker run -d \
  -e "OPTION_LIBS=ignite-rest-http" \
  apacheignite/ignite

By default, the Ignite Docker image starts with the following modules enabled:

ignite-log4j2,
ignite-spring,
ignite-indexing.

Environment Variables

The following parameters can be passed as environment variables in the docker container:

Parameter Name Description Default

Parameter Name	Description	Default
`CONFIG_URI`	URL to the Ignite configuration file (can also be relative to the META-INF folder on the class path). The downloaded config file is saved to ./ignite-config.xml	N/A
`OPTION_LIBS`	A list of modules that will be enabled for the node.	ignite-log4j2, ignite-spring, ignite-indexing
`JVM_OPTS`	JVM arguments passed to the Ignite instance.	N/A
`EXTERNAL_LIBS`	A list of URL’s to external libraries. Refer to Deploying User Libraries.	N/A

CONFIG_URI

URL to the Ignite configuration file (can also be relative to the META-INF folder on the class path). The downloaded config file is saved to ./ignite-config.xml

N/A

OPTION_LIBS

A list of modules that will be enabled for the node.

ignite-log4j2, ignite-spring, ignite-indexing

JVM_OPTS

JVM arguments passed to the Ignite instance.

N/A

EXTERNAL_LIBS

A list of URL’s to external libraries. Refer to Deploying User Libraries.

N/A

© 2024 The Apache Software Foundation.
Apache, Apache Ignite, the Apache feather and the Apache Ignite logo are either registered trademarks or trademarks of The Apache Software Foundation.