Run make_cluster_node.pyc

Use the make_cluster_node.pyc script to configure an installed instance into a node of a cluster. This script stores the bulk of required configuration information from the PostgreSQL database.

Before running make_cluster_node, make sure that all the required services are working, either as external services or as a Shared Services server. Additionally, make sure that the required ports and endpoints are opened in your firewall. See ports and endpoints.

Collect the required information

You need this information to answer prompts for make_cluster_node.

• IP addresses or hostnames for:
  • PostgreSQL server
  • HAProxy server and the port that the HAProxy server uses to accept HTTPS connections
  • GlusterFS server

• User names, passwords, tokens, or SSH key information for:
  • pgbouncer PostgreSQL database user
  • postgres PostgreSQL database user
  • login password for the HAProxy server, unless it uses an ssh key
  • username and password for the install being converted

Not all SSH key formats are accepted by make_cluster_node.pyc. You can use keys generated with the ssh-keygen -m PEM -t rsa -b 4096 command.

Create a node

Once you have either a Shared Services server or external services established, you convert installations of into cluster nodes.

You must first change to the directory where is installed.

1. Change to the home directory.

   cd <phantom_install_dir>/bin/

2. Run make_cluster_node.pyc using python.

   phenv python ./make_cluster_node.pyc --responses /path/to/mcn_responses.json

   You don't have to use mcn_responses.json. However, if you don't supply an alternate JSON file, the script prompts you for the information needed to create mcn_responses.json. The mcn_responses.json file contains secrets such as usernames and passwords in plain text. Store it in a secure location or delete it after the cluster configuration is complete.