External PostgreSQL server operator for Kubernetes

Sponsors

Please consider sponsoring my work

Sponsor

Current Sponsors

None

Features

• Creates a database from a CR
• Creates a role with random username and password from a CR
• If the database exist, it will only create a role
• Multiple user roles can own one database
• Creates Kubernetes secret with postgres_uri in the same namespace as CR
• Support for AWS RDS and Azure Database for PostgresSQL
• Support for managing CRs in dynamically created namespaces
• Template secret values

Cloud specific configuration

AWS

In order for this operator to work correctly with AWS RDS, you need to set POSTGRES_CLOUD_PROVIDER to AWS either in the ext-postgres-operator kubernetes secret or directly in the deployment manifest (operator.yaml).

Azure Database for PostgreSQL (both Single Server and Flexible Server)

In order for this operator to work correctly with Azure managed PostgreSQL database, two env variables needs to be provided for the operator:

• POSTGRES_CLOUD_PROVIDER set to Azure
• POSTGRES_DEFAULT_DATABASE set to your default database, i.e. postgres

GCP

In order for this operator to work correctly with GCP, you need to set POSTGRES_CLOUD_PROVIDER to GCP

To have operator work with GCP properly you have to:

• use postgresql connection in secret
• manually create a Master role e.g. "devops-operators"
• use such role in database CR e.g. spec.masterRole: devops-operator

DropRole method will check for db owner and will skip master role dropping

General Configuration

These environment variables are embedded in deploy/operator.yaml, env section.

• WATCH_NAMESPACE - which namespace to watch. Defaults to empty string for all namespaces
• OPERATOR_NAME - name of the operator, defaults to ext-postgres-operator
• POSTGRES_INSTANCE - identity of operator, this matched with postgres.db.movetokube.com/instance in CRs. Default is empty
• KEEP_SECRET_NAME - use secret name as provided by user (disabled by default)

POSTGRES_INSTANCE is only available since version 1.2.0

While using KEEP_SECRET_NAME could be a convenient way to define secrets with predictable and explicit names, the default logic reduces risk of operator from entering the endless reconcile loop as secret is very unlikely to exist.

The administrator should ensure that the SecretName does not collide with other secrets in the same namespace. If the secret already exists, the operator will never stop reconciling the CR until either offending secret is deleted or CR is deleted or updated with another SecretName

Installation

This operator requires a Kubernetes Secret to be created in the same namespace as operator itself. Secret should contain these keys: POSTGRES_HOST, POSTGRES_USER, POSTGRES_PASS, POSTGRES_URI_ARGS, POSTGRES_CLOUD_PROVIDER, POSTGRES_DEFAULT_DATABASE. Example:

apiVersion: v1
kind: Secret
metadata:
  name: ext-postgres-operator
  namespace: operators
type: Opaque
data:
  POSTGRES_HOST: cG9zdGdyZXM=
  POSTGRES_USER: cG9zdGdyZXM=
  POSTGRES_PASS: YWRtaW4=
  POSTGRES_URI_ARGS: IA==
  POSTGRES_CLOUD_PROVIDER: QVdT
  POSTGRES_DEFAULT_DATABASE: cG9zdGdyZXM=

To install the operator using kustomize, follow the steps below.

1. Configure Postgres credentials for the operator in deploy/secret.yaml
2. Create namespace if needed with
   kubectl apply -f deploy/namespace.yaml
3. Apply the secret with
   kubectl apply -f deploy/secret.yaml
4. Create the operator with either
   kubectl kustomize deploy/ | apply -f -
   or by using kustomize directly
   kustomize build deploy/ | apply -f -

Alternatively you can install operator using Helm Chart located in the charts/ext-postgres-operator subdirectory. Sample installation commands provided below:

helm repo add ext-postgres-operator https://movetokube.github.io/postgres-operator/
helm install -n operators ext-postgres-operator  ext-postgres-operator/ext-postgres-operator

See values.yaml for the possible values to define.

CRs

Postgres

apiVersion: db.movetokube.com/v1alpha1
kind: Postgres
metadata:
  name: my-db
  namespace: app
  annotations:
    # OPTIONAL
    # use this to target which instance of operator should process this CR. See General config 
    postgres.db.movetokube.com/instance: POSTGRES_INSTANCE
spec:
  database: test-db # Name of database created in PostgreSQL
  dropOnDelete: false # Set to true if you want the operator to drop the database and role when this CR is deleted (optional)
  masterRole: test-db-group (optional)
  schemas: # List of schemas the operator should create in database (optional)
  - stores
  - customers
  extensions: # List of extensions that should be created in the database (optional)
  - fuzzystrmatch
  - pgcrypto

This creates a database called test-db and a role test-db-group that is set as the owner of the database. Reader and writer roles are also created. These roles have read and write permissions to all tables in the schemas created by the operator, if any.

PostgresUser

apiVersion: db.movetokube.com/v1alpha1
kind: PostgresUser
metadata:
  name: my-db-user
  namespace: app
  annotations:
    # OPTIONAL
    # use this to target which instance of operator should process this CR. See general config
    postgres.db.movetokube.com/instance: POSTGRES_INSTANCE
spec:
  role: username
  database: my-db       # This references the Postgres CR
  secretName: my-secret
  privileges: OWNER     # Can be OWNER/READ/WRITE
  annotations:          # Annotations to be propagated to the secrets metadata section (optional)
    foo: "bar"
  secretTemplate:       # Output secrets can be customized using standard Go templates
    PQ_URL: "host={{.Host}} user={{.Role}} password={{.Password}} dbname={{.Database}}"

This creates a user role username-<hash> and grants role test-db-group, test-db-writer or test-db-reader depending on privileges property. Its credentials are put in secret my-secret-my-db-user (unless KEEP_SECRET_NAME is enabled).

PostgresUser needs to reference a Postgres in the same namespace.

Two Postgres referencing the same database can exist in more than one namespace. The last CR referencing a database will drop the group role and transfer database ownership to the role used by the operator. Every PostgresUser has a generated Kubernetes secret attached to it, which contains the following data (i.e.):

Key	Comment
DATABASE_NAME	Name of the database, same as in Postgres CR, copied for convenience
HOST	PostgreSQL server host
PASSWORD	Autogenerated password for user
ROLE	Autogenerated role with login enabled (user)
LOGIN	Same as ROLE. In case POSTGRES_CLOUD_PROVIDER is set to "Azure", LOGIN it will be set to {role}@{serverName}, serverName is extracted from POSTGRES_USER from operator's config.
POSTGRES_URL	Connection string for Posgres, could be used for Go applications
POSTGRES_JDBC_URL	JDBC compatible Postgres URI, formatter as jdbc:postgresql://{POSTGRES_HOST}/{DATABASE_NAME}

Multiple operator support

Since version 1.2 it is possible to use many instances of postgres-operator to control different databases based on annotations in CRs. Follow the steps below to enable multi-operator support.

1. Add POSTGRES_INSTANCE

Annotations Use Case

With the help of annotations it is possible to create annotation-based copies of secrets in other namespaces.

For more information and an example, see kubernetes-replicator#pull-based-replication

Template Use Case

Users can specify the structure and content of secrets based on their unique requirements using standard Go templates. This flexibility allows for a more tailored approach to meeting the specific needs of different applications.

Available context:

Variable	Meaning
.Host	Database host
.Role	Generated user/role name
.Database	Referenced database name
.Password	Generated role password

Contribution

You can contribute to this project by opening a PR to merge to master, or one of the vX.X.X branches.

Branching

master branch contains the latest source code with all the features. vX.X.X contains code for the specific major versions. i.e. v0.4.x contains the latest code for 0.4 version of the operator. See compatibility matrix below.

Tests

Please write tests and fix any broken tests before you open a PR. Tests should cover at least 80% of your code.

Compatibility

Postgres operator uses Operator SDK, which uses kubernetes client. Kubernetes client compatibility with Kubernetes cluster can be found here

Postgres operator compatibility with Operator SDK version is in the table below

	Operator SDK version	apiextensions.k8s.io
postgres-operator 0.4.x	v0.17	v1beta1
postgres-operator 1.x.x	v0.18	v1
HEAD	v0.18	v1