Deploy Nebula Graph cluster¶

This topic describes how to manually deploy a Nebula Graph cluster.

NOTE: For now, Nebula Graph does not have an official deployment tool.

Prerequisites¶

Prepare hardware for deploying the cluster.

Step 1: Install Nebula Graph¶

Install Nebula Graph on each machine in the cluster. Available approaches of installation are as follows.

• Install Nebula Graph with RPM or DEB package
• Install Nebula Graph by compiling the source code

Step 2: Modify the configurations¶

To deploy Nebula Graph according to your requirements, you have to modify the configuration files. All the configuration files for Nebula Graph, including nebula-graphd.conf, nebula-metad.conf, and nebula-storaged.conf, are stored in the etc directory in the installation path.

You only need to modify the configuration for the corresponding service on the machines. For example, modify nebula-graphd.conf on the machines where you want to deploy the Graph Service.

For how to prepare the configuration files, see:

• Meta Service configurations
• Graph Service configurations
• Storage Service configurations

Step 3: Start the cluster¶

Start the corresponding service on each machine. The command to start the Nebula Graph services is as follows.

sudo /usr/local/nebula/scripts/nebula.service start <metad|graphd|storaged|all>

/usr/local/nebula is the default installation path for Nebula Graph. Use the actual path if you have customized the path.

For more information about how to start and stop the services, see Manage Nebula Graph services.

Connect to the cluster¶

Connect to the Graph Service with a Nebula Graph client, such as Nebula Console. For more information, see Connect to Nebula Graph.

Check the cluster status¶

After connecting to the Nebula Graph cluster, run SHOW HOSTS to check the cluster status.