Your Own Machines

Deploy on existing servers

info

This method is suitable for deployment on existing servers, whether with another cloud provider or in your own data center.

Prerequisites

You will need root access or a user with sudo privileges to access the servers via SSH. You can use your private SSH key (assuming the corresponding public key has already been added to the servers), or a username and password if password access is enabled on your servers.

See also the Requirements and Compatibility pages.

Console (UI)

Select 'Your Own Machines' in the deployment destination.

Specify the name and IP addresses of your servers.

note

Use private IP addresses so that the cluster does not listen a public IP.

Example:

warning

Please note that at least 3 servers are required to ensure high availability.

Select authentication method.

Optionally, specify an IP address to provide a single entry point for client access to databases in the cluster.

Example:

info

This must be an IP address that is currently unused on your network. The address will be assigned to the cluster after deployment.

warning

Not for cloud environments. Because VIP-based solutions like keepalived or vip-manager may not works in such environments.

Optionally, check the box for "HAProxy load balancer".

tip

This feature supports load balancing for read operations, facilitating effective scale-out strategies through the use of read-only replicas. See the details on the Architecture page

Choose which environment your database cluster belongs to.

Specify a name for your cluster.

Optionally, specify a description.

Select the PostgreSQL version to install.

Review the summary and click the "CREATE CLUSTER" button.

Wait until deployment is complete. This process takes about 10 minutes.

info

You can see the deployment log in the "Operations" section. To do this, select the relevant event with the "deploy" type and click "Show details" under the Actions tab.

tip

After a successful deployment, you can obtain the connection info on the cluster page. To do this, click on the name of your cluster on the "Clusters" page.

Example of a cluster page:

Command line

tip

We also support converting your existing PostgreSQL installation into a high-availability cluster. If you want to upgrade your current PostgreSQL setup to a clustered configuration, simply set postgresql_exists=true in the inventory file.

note

Please note that during the cluster setup process, your existing PostgreSQL service will be automatically restarted, leading to a brief period of database downtime. Plan this transition accordingly.

1. Edit the inventory file

Specify IP addresses and connection settings for your environment:

• ansible_user
• ansible_ssh_pass or ansible_ssh_private_key_file

nano inventory

2. Edit the variable file vars/main.yml

nano vars/main.yml

Minimum set of variables:

• proxy_env to download packages in environments without direct internet access (optional)
• patroni_cluster_name
• postgresql_version
• postgresql_data_dir
• cluster_vip to provide a single entry point for client access to databases in the cluster (optional)
• with_haproxy_load_balancing to enable load balancing (optional)
• dcs_type "etcd" (default) or "consul"

info

See the vars/main.yml, system.yml and (Debian.yml or RedHat.yml) files for more details.

3. Try to connect to hosts

ansible all -m ping

4. Run playbook:

ansible-playbook deploy_pgcluster.yml

Deploy Cluster with TimescaleDB:

Click here to expand...

To deploy a TimescaleDB HA cluster, set the enable_timescale variable:

ansible-playbook deploy_pgcluster.yml -e "enable_timescale=true"

How to start from scratch:

Click here to expand...

If you need to start from scratch, you can use the remove_cluster.yml playbook. Run the following command to remove the specified components:

ansible-playbook remove_cluster.yml -e "remove_postgres=true remove_etcd=true"

This command will delete the specified components, allowing you to start a new installation.

Available variables:

• remove_postgres: stop the PostgreSQL service and remove data.
• remove_etcd: stop the ETCD service and remove data.
• remove_consul: stop the Consul service and remove data.

warning

Caution: be careful when running this command in a production environment.