Learn how Neon compares to Aurora Serverless v2 - TL;DR: faster cold starts, responsive autoscaling, 80% lower costs

PostgreSQL Backup

Summary: in this tutorial, you will learn how to backup the PostgreSQL databases using the pg_dump and pg_dumpall tools.

Introduction to PostgreSQL backup

A PostgreSQL backup is a copy of the data that you can use to recover the database later. Typically, a backup includes all or selected data, schema, and configuration settings necessary to restore the database to a desired state.

You need to back up PostgreSQL databases regularly to prevent data loss in case of human errors, hardware failures, disasters, or other unforeseen circumstances.

Before backing up the databases, you should consider the following types of backups:

  • Logical backups
  • Physical backups

PostgreSQL comes with pg_dump and pg_dumpall tools that help you perform logical backups effectively.

pg_dump

The pg_dump tool is a command-line utility that you can use to create a logical backup of a PostgreSQL database.

The pg_dump extracts a PostgreSQL database into a script file or other archive file. The file represents the snapshot of the database at the time pg_dump begins running.

Here’s the syntax of the pg_dump command:

pg_dump [connection_option] [option] [dbname]

The following table illustrates the options for the pg_dump command:

OptionDescription
-U, –username=USERNAMESpecifies the username to connect to the database.
-h, –host=HOSTNAMESpecifies the hostname of the server on which the PostgreSQL server is running.
-p, –port=PORTSpecifies the port number on which the PostgreSQL server is listening.
-d, –dbname=DBNAMESpecifies the name of the database to be dumped.
-n, –schema=SCHEMASpecifies the schema(s) to be dumped. Multiple schemas can be specified, separated by commas.
-t, –table=TABLESpecifies the table(s) to be dumped. Multiple tables can be specified, separated by commas.
-F, –format=FORMATSpecifies the output file format (e.g., plain, custom, directory).
-f, –file=FILENAMESpecifies the name of the output file.
-W, –passwordForces pg_dump to prompt for a password before connecting to the PostgreSQL server.
-w, –no-passwordSuppresses the password prompt, but the password might be supplied in other ways (e.g., .pgpass file).
-c, –cleanAdds SQL commands to clean (drop) database objects before recreating them.
-C, –createAdds SQL commands to create the database before restoring data into it.
-a, –data-onlyDumps only the data, not the schema (no schema-creating commands are included).
-s, –schema-onlyDumps only the schema, not the data.
-x, –no-privilegesExcludes access privileges (GRANT/REVOKE commands) from the dump.
-X, –no-ownerPrevents dumping of object ownership information (such as OWNER TO).
-h, –helpDisplays help and usage information.

pg_dumpall

The pg_dumpall tool is a command-line utility that you can use to create logical backups of the entire PostgreSQL cluster, including all databases, schemas, roles, and other cluster-wide objects.

Unlike the pg_dump tool which backups individual databases or objects, the pg_dumpall tool offers a convenient way to make a backup of all databases in an PostgreSQL cluster (instance) in a single operation.

Here’s the syntax for pg_dumpall command:

pg_dump [connection_option] [option]

The following table presents some common command-line options for pg_dumpall utility:

OptionDescription
-U, –username=USERNAMESpecifies the username to connect to the PostgreSQL server.
-h, –host=HOSTNAMESpecifies the hostname of the server on which the PostgreSQL server is running.
-p, –port=PORTSpecifies the port number on which the PostgreSQL server is listening.
-g, –globals-onlyDumps only global objects (roles and tablespaces), no database-specific objects.
-r, –roles-onlyDumps only role (user and group) definitions, no databases or tablespaces.
-t, –tablespaces-onlyDumps only the definitions of tablespaces, no databases or roles.
-c, –cleanAdds SQL commands to clean (drop) database objects before recreating them.
-C, –createAdds SQL commands to create the database before restoring data into it.
-x, –no-privilegesExcludes access privileges (GRANT/REVOKE commands) from the dump.
-X, –no-ownerPrevents dumping of object ownership information (such as OWNER TO).
-s, –schema-onlyDumps only the schema, no data.
-v, –verboseDisplays verbose output, including informational messages during the dump process.
-V, –versionDisplays the version of pg_dumpall and exits.
-?, –helpDisplays help and usage information.

PostgreSQL backup examples

Let’s take some examples of backing up PostgreSQL databases.

1) Backing up a single database using the pg_dump utility

First, open the Command Prompt on Windows or Terminal on Unix-like systems.

Second, execute the following pg_dump to back up the dvdrental database on the local server:

pg_dump -U postgres -d dvdrental -F tar -f d:\backup\dvdrental.tar

In this command:

  • -U postgres:  specifies the user (postgres) that connects to the PostgreSQL database server.
  • -d dvdrental: specifies the database name that you want to back up.
  • -F tar : specifies tar as the output format of the archive file.
  • -f d:\backup\dvdrental.tar: This is the file path of the output backup file. Note that the D:\backup directory must exist. Also, it is a good practice to place a backup file in a server that is not the same as the one the PostgreSQL server is running.

After you run the command, pg_dump will prompt you to enter a password for the postgres user:

password:;

After you enter a valid password, the pg_dump will create the backup file dvdrental.tar in the d:\backup directory.

If you want to automate the backup regularly, you need to use a password file (.pgpass on Unix-like systems and pgpass.conf on Windows) so that when the command executes, it doesn’t ask for the password.

2) Backing up all databases using the pg_dumpall utility

First, open the Command Prompt on Windows or Terminal on Unix-like systems.

Second, run the pg_dumpall command to back up all databases in the local PostgreSQL server into the all_databases.sql file:

pg_dumpall -U postgres > D:\backup\all_databases.sql

This command will prompt you to enter a password for each database in the PostgreSQL server.

To suppress the password prompt, you can use the -w option:

pg_dumpall -U postgres > D:\backup\all_databases.sql

However, you need to use a password file with the following entry:

hostname:port:database:username:password

For example:

localhost:5432:*:postgres:SecurePass1

The asterisk (*) means all databases.

3) Backing up database objects using the pg_dumpall utility

Sometimes, you want to backup only database object definitions, not the data. This is helpful in the testing phase, in which you do not want to move test data to the live system.

To back up objects in all databases, including roles, tablespaces, databases, schemas, tables, indexes, triggers, functions, constraints, views, ownerships, and privileges, you use the following command:

pg_dumpall --schema-only > d:\backup\schemas.sql

If you want to back up role definition only, use the following command:

pg_dumpall --roles-only > D:\backup\roles.sql

If you want to back up tablespaces definition, use the following command:

pg_dumpall --tablespaces-only > d:\backup\tablespaces.sql

Creating backup script on Windows

The following backup script assumes that you have a password file setup properly.

First, open a text editor such a Notepad.

Second, write a batch script to generate backup file with a date and timestamp:

@echo off
REM Set variables for database connection
set PGUSER=your_username
set PGDATABASE=your_database_name

REM Set the path where you want to store the backup files
set BACKUP_DIR=C:\path\to\backup\directory

REM Get current date and time
for /f "tokens=2-4 delims=/ " %%a in ('date /t') do (set "datestamp=%%c-%%a-%%b")
for /f "tokens=1-2 delims=: " %%a in ('time /t') do (set "timestamp=%%a%%b")

REM Execute pg_dump command to dump the database
pg_dump -U %PGUSER% -d %PGDATABASE% -f "%BACKUP_DIR%\%PGDATABASE%_%datestamp%_%timestamp%.sql"

In the script, you need to replace your_username, your_database_name, and the C:\path\to\backup\directory with your actual ones that you want to use.

Third, save the script to a file with the .bat extension such as backup.bat.

Finally, run the script by double-clicking it.

If you want to make a regular backup, you can use the Task Scheduler on Windows to run the backup.bat file periodically.

Creating backup script on Linux or macOS

The following backup script assumes that you have a password file setup properly.

First, open a text editor.

Second, write a shell script to generate a backup file with a date and timestamp:

#!/bin/bash

# Set variables for database connection
PGUSER=your_username
PGDATABASE=your_database_name

# Set the path where you want to store the backup files
BACKUP_DIR=/path/to/backup/directory

# Get current date and time
datestamp=$(date +'%Y-%m-%d')
timestamp=$(date +'%H%M')

# Execute pg_dump command to dump the database
pg_dump -U "$PGUSER" -d "$PGDATABASE" > "$BACKUP_DIR/$PGDATABASE"_"$datestamp"_"$timestamp".sql

In this script, you need to replace your_username, your_database_name, and /path/to/backup/directory with your actual ones that you want to use.

Third, save the script with a .sh extension such as backup.sh.

Fourth, make the script executable by executing the following command on your terminal:

chmod +x backup_script.sh

Fifth, run the script by executing it in your terminal:

./backup.sh

This will execute the pg_dump command to dump your database into a backup file with extension .sql with the current date and timestamp appended to its name.

If you want to backup a database regularly, you can create a cron job to run the script periodically.

Summary

  • Back up PostgreSQL databases regularly for recovery later.
  • Use the pg_dump and pg_dumpall tools to perform logical backups.
  • Use a password file to automate the backup processes.

Last updated on

Was this page helpful?