bitnami/postgresql-repmgr
Bitnami container image for PostgreSQL HA
50M+
This PostgreSQL cluster solution includes the PostgreSQL replication manager, an open-source tool for managing replication and failover on PostgreSQL clusters.
Overview of PostgreSQL HA Trademarks: This software listing is packaged by Bitnami. The respective trademarks mentioned in the offering are owned by the respective companies, and use of them does not imply any affiliation or endorsement.
docker run --name postgresql-repmgr bitnami/postgresql-repmgr:latest
You can find the default credentials and available configuration options in the Environment Variables section.
Looking to use PostgreSQL HA in production? Try VMware Tanzu Application Catalog, the commercial edition of the Bitnami catalog.
Non-root container images add an extra layer of security and are generally recommended for production environments. However, because they run as a non-root user, privileged tasks are typically off-limits. Learn more about non-root containers in our docs.
Deploying Bitnami applications as Helm Charts is the easiest way to get started with our applications on Kubernetes. Read more about the installation in the Bitnami PostgreSQL HA Chart GitHub repository.
Bitnami containers can be used with Kubeapps for deployment and management of Helm Charts in clusters.
Starting December 10th 2024, only the latest stable branch of any container will receive updates in the free Bitnami catalog. To access up-to-date releases for all upstream-supported branches, consider upgrading to Bitnami Premium. Previous versions already released will not be deleted. They are still available to pull from DockerHub.
Please check the Bitnami Premium page in our partner Arrow Electronics for more information.
Dockerfile
linksLearn more about the Bitnami tagging policy and the difference between rolling tags and immutable tags in our documentation page.
You can see the equivalence between the different tags by taking a look at the tags-info.yaml
file present in the branch folder, i.e bitnami/ASSET/BRANCH/DISTRO/tags-info.yaml
.
Subscribe to project updates by watching the bitnami/containers GitHub repo.
The recommended way to get the Bitnami PostgreSQL HA Docker Image is to pull the prebuilt image from the Docker Hub Registry.
docker pull bitnami/postgresql-repmgr:latest
To use a specific version, you can pull a versioned tag. You can view the list of available versions in the Docker Hub Registry.
docker pull bitnami/postgresql-repmgr:[TAG]
If you wish, you can also build the image yourself by cloning the repository, changing to the directory containing the Dockerfile and executing the docker build
command. Remember to replace the APP
, VERSION
and OPERATING-SYSTEM
path placeholders in the example command below with the correct values.
git clone https://github.com/bitnami/containers.git
cd bitnami/APP/VERSION/OPERATING-SYSTEM
docker build -t bitnami/APP:latest .
If you remove the container all your data will be lost, and the next time you run the image the database will be reinitialized. To avoid this loss of data, you should mount a volume that will persist even after the container is removed.
For persistence you should mount a directory at the /bitnami/postgresql
path. If the mounted directory is empty, it will be initialized on the first run.
docker run \
-v /path/to/postgresql-repmgr-persistence:/bitnami/postgresql \
bitnami/postgresql-repmgr:latest
The docker-compose.yml
file present in this repository already configures persistence.
NOTE: As this is a non-root container, the mounted files and directories must have the proper permissions for the UID
1001
.
Using Docker container networking, a PostgreSQL server running inside a container can easily be accessed by your application containers and vice-versa.
Containers attached to the same network can communicate with each other using the container name as the hostname.
In this example, we will create a PostgreSQL client instance that will connect to the server instance that is running on the same docker network as the client.
Step 1: Create a network
docker network create my-network --driver bridge
Step 2: Launch the postgresql-repmgr container within your network
Use the --network <NETWORK>
argument to the docker run
command to attach the container to the my-network
network.
docker run --detach --rm --name pg-0 \
--network my-network \
--env REPMGR_PARTNER_NODES=pg-0 \
--env REPMGR_NODE_NAME=pg-0 \
--env REPMGR_NODE_NETWORK_NAME=pg-0 \
--env REPMGR_PRIMARY_HOST=pg-0 \
--env REPMGR_PASSWORD=repmgrpass \
--env POSTGRESQL_PASSWORD=secretpass \
bitnami/postgresql-repmgr:latest
Step 3: Launch your PostgreSQL client instance
Finally we create a new container instance to launch the PostgreSQL client and connect to the server created in the previous step:
docker run -it --rm \
--network my-network \
bitnami/postgresql:10 \
psql -h pg-0 -U postgres
When not specified, Docker Compose automatically sets up a new network and attaches all deployed services to that network. However, we will explicitly define a new bridge
network named my-network
. In this example we assume that you want to connect to the PostgreSQL server from your own custom application image which is identified in the following snippet by the service name myapp
.
version: '2'
networks:
my-network:
driver: bridge
services:
pg-0:
image: 'bitnami/postgresql-repmgr:latest'
networks:
- my-network
environment:
- POSTGRESQL_PASSWORD=custompassword
- REPMGR_PASSWORD=repmgrpassword
- REPMGR_PRIMARY_HOST=pg-0
- REPMGR_NODE_NETWORK_NAME=pg-0
- REPMGR_NODE_NAME=pg-0
- REPMGR_PARTNER_NODES=pg-0
myapp:
image: 'YOUR_APPLICATION_IMAGE'
networks:
- my-network
IMPORTANT:
- Please update the YOUR_APPLICATION_IMAGE_ placeholder in the above snippet with your application image
- In your application container, use the hostname
pg-0
to connect to the PostgreSQL server
Launch the containers using:
docker-compose up -d
When the container is executed for the first time, it will execute the files with extensions .sh
, .sql
and .sql.gz
located at /docker-entrypoint-initdb.d
.
In order to have your custom files inside the docker image you can mount them as a volume.
In the above commands you may have noticed the use of the POSTGRESQL_PASSWORD
and REPMGR_PASSWORD
environment variables. Passing the POSTGRESQL_PASSWORD
environment variable when running the image for the first time will set the password of the postgres
user to the value of POSTGRESQL_PASSWORD
(or the content of the file specified in POSTGRESQL_PASSWORD_FILE
). In the same way, passing the REPMGR_PASSWORD
environment variable sets the password of the repmgr
user to the value of REPMGR_PASSWORD
(or the content of the file specified in REPMGR_PASSWORD_FILE
).
docker run --name pg-0 --env REPMGR_PASSWORD=repmgrpass --env POSTGRESQL_PASSWORD=secretpass bitnami/postgresql-repmgr:latest
or by modifying the docker-compose.yml
file present in this repository:
...
services:
pg-0:
...
environment:
- - POSTGRESQL_PASSWORD=adminpassword
+ - POSTGRESQL_PASSWORD=password123
- - REPMGR_PASSWORD=repmgrpassword
+ - REPMGR_PASSWORD=password123
...
pg-1:
...
environment:
- - POSTGRESQL_PASSWORD=adminpassword
+ - POSTGRESQL_PASSWORD=password123
- - REPMGR_PASSWORD=repmgrpassword
+ - REPMGR_PASSWORD=password123
...
Note!
Both postgres
and repmgr
users are superusers and have full administrative access to the PostgreSQL database.
Refer to Creating a database user on first run if you want to set an unprivileged user and a password for the postgres
user.
By passing the POSTGRESQL_DATABASE
environment variable when running the image for the first time, a database will be created. This is useful if your application requires that a database already exists, saving you from having to manually create the database using the PostgreSQL client.
docker run --name pg-0 --env POSTGRESQL_DATABASE=my_database bitnami/postgresql-repmgr:latest
You can also create a restricted database user that only has permissions for the database created with the POSTGRESQL_DATABASE
environment variable. To do this, provide the POSTGRESQL_USERNAME
environment variable.
docker run --name pg-0 --env POSTGRESQL_USERNAME=my_user --env POSTGRESQL_PASSWORD=password123 --env POSTGRESQL_DATABASE=my_database bitnami/postgresql-repmgr:latest
The docker-compose.yml
file present in this repository already configures this setup.
Note!
When POSTGRESQL_USERNAME
is specified, the postgres
user is not assigned a password and as a result you cannot login remotely to the PostgreSQL server as the postgres
user. If you still want to have access with the user postgres
, please set the POSTGRESQL_POSTGRES_PASSWORD
environment variable (or the content of the file specified in POSTGRESQL_POSTGRES_PASSWORD_FILE
).
A HA PostgreSQL cluster with Streaming replication and repmgr can easily be setup with the Bitnami PostgreSQL HA Docker Image using the following environment variables:
POSTGRESQL_PASSWORD
: Password for postgres
user. No defaults.POSTGRESQL_PASSWORD_FILE
: Path to a file that contains the postgres
user password. This will override the value specified in POSTGRESQL_PASSWORD
. No defaults.REPMGR_USERNAME
: Username for repmgr
user. Defaults to repmgr
.REPMGR_PASSWORD_FILE
: Path to a file that contains the repmgr
user password. This will override the value specified in REPMGR_PASSWORD
. No defaults.REPMGR_PASSWORD
: Password for repmgr
user. No defaults.REPMGR_USE_PASSFILE
: Configure repmgr to use passfile
and PGPASSFILE
instead of plain-text password in its configuration.REPMGR_PASSFILE_PATH
: Location of the passfile, if it doesn't exist it will be created using REPMGR credentials.REPMGR_PRIMARY_HOST
: Hostname of the initial primary node. No defaults.REPMGR_PARTNER_NODES
: Comma separated list of partner nodes in the cluster. No defaults.REPMGR_NODE_NAME
: Node name. No defaults.REPMGR_NODE_TYPE
: Node type. Defaults to data
. Allowed values: data
for data nodes (master or replicas), witness
for witness nodes.REPMGR_NODE_NETWORK_NAME
: Node hostname. No defaults.REPMGR_PGHBA_TRUST_ALL
: This will set the auth-method in the generated pg_hba.conf. Set it to yes
only if you are using pgpool with LDAP authentication. Default to no
.In a HA PostgreSQL cluster you can have one primary and zero or more standby nodes. The primary node is in read-write mode, while the standby nodes are in read-only mode. For best performance its advisable to limit the reads to the standby nodes.
NOTE: REPMGR_USE_PASSFILE and REPMGR_PASSFILE_PATH will be ignored for Postgresql prior to version 9.6.
When mounting an external passfile using REPMGR_PASSFILE_PATH, it is necessary to also configure REPMGR_PASSWORD and REPMGR_USERNAME accordingly.
Step 1: Create a network and the initial primary node
The first step is to start the initial primary node:
docker network create my-network --driver bridge
docker run --detach --name pg-0 \
--network my-network \
--env REPMGR_PARTNER_NODES=pg-0,pg-1 \
--env REPMGR_NODE_NAME=pg-0 \
--env REPMGR_NODE_NETWORK_NAME=pg-0 \
--env REPMGR_PRIMARY_HOST=pg-0 \
--env REPMGR_PASSWORD=repmgrpass \
--env POSTGRESQL_PASSWORD=secretpass \
bitnami/postgresql-repmgr:latest
Step 2: Create a standby node
Next we start a standby node:
docker run --detach --name pg-1 \
--network my-network \
--env REPMGR_PARTNER_NODES=pg-0,pg-1 \
--env REPMGR_NODE_NAME=pg-1 \
--env REPMGR_NODE_NETWORK_NAME=pg-1 \
--env REPMGR_PRIMARY_HOST=pg-0 \
--env REPMGR_PASSWORD=repmgrpass \
--env POSTGRESQL_PASSWORD=secretpass \
bitnami/postgresql-repmgr:latest
With these three commands you now have a two node PostgreSQL primary-standby streaming replication cluster up and running. You can scale the cluster by adding/removing standby nodes without incurring any downtime.
Note: The cluster replicates the primary in its entirety, which includes all users and databases.
If the master goes down, repmgr will ensure any of the standby nodes takes the primary role, guaranteeing high availability.
Note: The configuration of the other nodes in the cluster needs to be updated so that they are aware of them. This would require you to restart the old nodes adapting the
REPMGR_PARTNER_NODES
environment variable.
With Docker Compose the HA PostgreSQL cluster can be setup using the docker-compose.yml
file present in this repository:
curl -sSL https://raw.githubusercontent.com/bitnami/containers/main/bitnami/postgresql-repmgr/docker-compose.yml > docker-compose.yml
docker-compose up -d
PostgreSQL supports the encryption of connections using the SSL/TLS protocol. Should you desire to enable this optional feature, you may use the following environment variables to configure the application:
POSTGRESQL_ENABLE_TLS
: Whether to enable TLS for traffic or not. Defaults to no
.POSTGRESQL_TLS_CERT_FILE
: File containing the certificate file for the TLS traffic. No defaults.POSTGRESQL_TLS_KEY_FILE
: File containing the key for certificate. No defaults.POSTGRESQL_TLS_CA_FILE
: File containing the CA of the certificate. If provided, PostgreSQL will authenticate TLS/SSL clients by requesting them a certificate (see ref). No defaults.POSTGRESQL_TLS_CRL_FILE
: File containing a Certificate Revocation List. No defaults.POSTGRESQL_TLS_PREFER_SERVER_CIPHERS
: Whether to use the server's TLS cipher preferences rather than the client's. Defaults to yes
.When enabling TLS, PostgreSQL will support both standard and encrypted traffic by default, but prefer the latter. Below there are some examples on how to quickly set up TLS traffic:
Using docker run
$ docker run \
-v /path/to/certs:/opt/bitnami/postgresql/certs \
-e POSTGRESQL_ENABLE_TLS=yes \
-e POSTGRESQL_TLS_CERT_FILE=/opt/bitnami/postgresql/certs/postgres.crt \
-e POSTGRESQL_TLS_KEY_FILE=/opt/bitnami/postgresql/certs/postgres.key \
bitnami/postgresql-repmgr:latest
Modifying the docker-compose.yml
file present in this repository:
services:
pg-0:
...
environment:
...
- POSTGRESQL_ENABLE_TLS=yes
- POSTGRESQL_TLS_CERT_FILE=/opt/bitnami/postgresql/certs/postgres.crt
- POSTGRESQL_TLS_KEY_FILE=/opt/bitnami/postgresql/certs/postgres.key
...
volumes:
...
- /path/to/certs:/opt/bitnami/postgresql/certs
...
Alternatively, you may also provide this configuration in your custom configuration file.
The image looks for the repmgr.conf
, postgresql.conf
and pg_hba.conf
files in /opt/bitnami/repmgr/conf/
and /opt/bitnami/postgresql/conf/
. You can mount a volume at /bitnami/repmgr/conf/
and copy/edit the configuration files in the /path/to/custom-conf/
. The default configurations will be populated to the conf/
directories if /bitnami/repmgr/conf/
is empty.
/path/to/custom-conf/
└── postgresql.conf
As the PostgreSQL with Replication manager image is non-root, you need to set the proper permissions to the mounted directory in your host:
sudo chgrp -R root /path/to/custom-conf/
sudo chmod -R g+rwX /path/to/custom-conf/
Step 1: Run the PostgreSQL image
Run the PostgreSQL image, mounting a directory from your host.
docker run --name pg-0 \
-v /path/to/custom-conf/:/bitnami/repmgr/conf/ \
bitnami/postgresql-repmgr:latest
or using Docker Compose:
version: '2'
services:
pg-0:
image: bitnami/postgresql-repmgr:latest
ports:
- '5432:5432'
volumes:
- /path/to/custom-conf/:/bitnami/repmgr/conf/
pg-1:
image: bitnami/postgresql-repmgr:latest
ports:
- '5432:5432'
volumes:
- /path/to/custom-conf/:/bitnami/repmgr/conf/
Step 2: Edit the configuration
Edit the configuration on your host using your favorite editor.
vi /path/to/custom-conf/postgresql.conf
Step 3: Restart PostgreSQL
After changing the configuration, restart your PostgreSQL container for changes to take effect.
docker restart pg-0
or using Docker Compose:
docker-compose restart pg-0
docker-compose restart pg-1
Refer to the server configuration manual for the complete list of configuration options.
Allow settings to be loaded from files other than the default postgresql.conf
Apart of using a custom repmgr.conf
, postgresql.conf
or pg_hba.conf
, you can include files ending in .conf
from the conf.d
directory in the volume at /bitnami/postgresql/conf/
.
For this purpose, the default postgresql.conf
contains the following section:
##------------------------------------------------------------------------------
## CONFIG FILE INCLUDES
##------------------------------------------------------------------------------
## These options allow settings to be loaded from files other than the
## default postgresql.conf.
include_dir = 'conf.d' # Include files ending in '.conf' from directory 'conf.d'
If you are using your custom postgresql.conf
, you should create (or uncomment) the above section in your config file, in this case the structure should be something like
/path/to/custom-conf/
└── postgresql.conf
/path/to/extra-custom-conf/
└── extended.conf
Remember to set the proper permissions to the mounted directory in your host:
sudo chgrp -R root /path/to/extra-custom-conf/
sudo chmod -R g+rwX /path/to/extra-custom-conf/
Step 1: Deploy the PostgreSQL image
Run the PostgreSQL image, mounting a directory from your host.
docker run --name pg-0 \
-v /path/to/extra-custom-conf/:/bitnami/postgresql/conf/conf.d/ \
-v /path/to/custom-conf/:/bitnami/repmgr/conf/ \
bitnami/postgresql-repmgr:latest
or using Docker Compose:
version: '2'
services:
pg-0:
image: bitnami/postgresql-repmgr:latest
ports:
- '5432:5432'
volumes:
- /path/to/extra-custom-conf/:/bitnami/postgresql/conf/conf.d/
- /path/to/custom-conf/:/bitnami/repmgr/conf/
pg-1:
image: bitnami/postgresql-repmgr:latest
ports:
- '5432:5432'
volumes:
- /path/to/extra-custom-conf/:/bitnami/postgresql/conf/conf.d/
- /path/to/custom-conf/:/bitnami/repmgr/conf/
Step 2: Edit the configuration file
Edit the configuration on your host using your favorite editor.
vi /path/to/extra-custom-conf/extended.conf
Step 3: Restart the PostgreSQL container
After changing the configuration, restart your PostgreSQL container for changes to take effect.
docker restart pg-0
or using Docker Compose:
docker-compose restart pg-0
docker-compose restart pg-1
It is possible to add extra services to the provided docker-compose.yaml
file, like a witness node. When adding the new service, please take into account the cluster set up process relays on the REPMGR_NODE_ID_START_SEED
environment variable plus the service ID in the name (if present, or zero (0
) by default) to assign cluster's ID to each service/node involved on it. In the case of docker-compose based clusters, this may lead to collisions in the internal IDs in case two or more services share the same ID in their names, making the service initialization process to fail. This isn't an issue on Kubernetes environments, as the Kubernetes controller enumerates the pods with different ID numbers by default.
We recommend setting a different value for the REPMGR_NODE_ID_START_SEED
in those nodes, or ensuring no services names use repeated numbers. Find below a sample service for a witness service:
pg-0:
(...)
pg-1:
(...)
pgw-0:
image: bitnami/postgresql-repmgr:latest
ports:
- 6439:5432
volumes:
- /docker/local/database_repmgr2/pgw-0:/bitnami/postgresql
environment:
(...)
- REPMGR_PRIMARY_HOST=pg-0
- REPMGR_PRIMARY_PORT=5432
- REPMGR_PARTNER_NODES=pg-0:5432,pg-1:5432,pgw-0:5432
- REPMGR_NODE_NAME=pgw-0
- REPMGR_NODE_NETWORK_NAME=pgw-0
- REPMGR_PORT_NUMBER=5432
- REPMGR_NODE_TYPE=witness
# Avoid naming collision with 'pg-0' service
- REPMGR_NODE_ID_START_SEED=2000
(...)
Refer to issues/27124 for further details on this.
Customizable environment variables
| Name | Description | Default Value
Note: the README for this container is longer than the DockerHub length limit of 25000, so it has been trimmed. The full README can be found at https://github.com/bitnami/containers/blob/main/bitnami/postgresql-repmgr/README.md