PostgreSQL
Related articles
PostgreSQL is an open source, community driven, standard compliant object-relational database system.
This document describes how to set up PostgreSQL. It also describes how to configure PostgreSQL to be accessible from a remote client. Among other applications, PostgreSQL can be substituted for MySQL as part of the LAMP web stack.
Contents
Installing PostgreSQL
Install the postgresql package. Then set a password for the newly created postgres user.
Then, switch to the default PostgreSQL user postgres by executing the following command:
- If you have sudo and your username is in
sudoers
:
$ sudo -u postgres -i
- Otherwise:
$ su # su -l postgres
See su(1) or sudo(8) for their usage.
Before PostgreSQL can function correctly, the database cluster must be initialized:
[postgres]$ initdb --locale $LANG -E UTF8 -D '/var/lib/postgres/data'
Where:
- the
--locale
is the one defined in the file/etc/locale.conf
; - the
-E
is the default encoding of the database that will be created in the future; - and
-D
is the default location where the database cluster must be stored.
Many lines should now appear on the screen with several ending by ... ok
:
The files belonging to this database system will be owned by user "postgres". This user must also own the server process. The database cluster will be initialized with locale "en_GB.UTF-8". The default text search configuration will be set to "english". Data page checksums are disabled. fixing permissions on existing directory /var/lib/postgres/data ... ok creating subdirectories ... ok selecting default max_connections ... 100 selecting default shared_buffers ... 128MB selecting dynamic shared memory implementation ... posix creating configuration files ... ok creating template1 database in /var/lib/postgres/data/base/1 ... ok initializing pg_authid ... ok [...]
If these are the kind of lines you see, then the process succeeded. Return to the regular user using exit
.
As root, start and enable postgresql.service
.
Create your first database/user
Become the postgres user. Add a new database user using the createuser command:
[postgres]$ createuser --interactive
Create a new database over which the above user has read/write privileges using the createdb command (execute this command from your login shell if the database user has the same name as your Linux user, otherwise add -U database-username
to the following command):
$ createdb myDatabaseName
Familiarize with PostgreSQL
Access the database shell
Become the postgres user. Start the primary database shell, psql, where you can do all your creation of databases/tables, deletion, set permissions, and run raw SQL commands. Use the -d
option to connect to the database you created (without specifying a database, psql
will try to access a database that matches your username).
[postgres]$ psql -d myDatabaseName
Some helpful commands:
Get help:
=> \help
Connect to a particular database:
=> \c <database>
List all users and their permission levels:
=> \du
Show summary information about all tables in the current database:
=> \dt
Exit/quit the psql
shell:
=> \q or CTRL+d
There are of course many more meta-commands, but these should help you get started. To see all meta-commands run:
=> \?
Optional configuration
Configure PostgreSQL to be accessible from remote hosts
The PostgreSQL database server configuration file is postgresql.conf
. This file is located in the data directory of the server, typically /var/lib/postgres/data
. This folder also houses the other main configuration files, including the pg_hba.conf
.
Edit the file /var/lib/postgres/data/postgresql.conf
. In the connections and authentications section, add the listen_addresses
line to your needs:
listen_addresses = 'localhost,my_local_ip_address' #You can use '*' to listen on all local addresses
Take a careful look at the other lines.
Host-based authentication is configured in /var/lib/postgres/data/pg_hba.conf
. This file controls which hosts are allowed to connect. Note that the defaults allow any local user to connect as any database user, including the database superuser. Add a line like the following:
# IPv4 local connections: host all all my_remote_client_ip_address/32 md5
where my_remote_client_ip_address
is the IP address of the client.
See the documentation for pg_hba.conf.
After this you should restart postgresql.service
for the changes to take effect.
For troubleshooting take a look in the server log file:
$ journalctl -u postgresql
Configure PostgreSQL authenticate against PAM
PostgreSQL offers a number of authentication methods. If you would like to allow users to authenticate with their system password, additional steps are necessary. First you need to enable PAM for the connection.
For example, the same configuration as above, but with PAM enabled:
# IPv4 local connections: host all all my_remote_client_ip_address/32 pam
The PostgreSQL server is however running without root privileges and will not be able to access /etc/shadow
. We can work around that by allowing the postgres group to access this file:
setfacl -m g:postgres:r /etc/shadow
Change default data directory
The default directory where all your newly created databases will be stored is /var/lib/postgres/data
. To change this, follow these steps:
Create the new directory and make the postgres user its owner:
# mkdir -p /pathto/pgroot/data # chown -R postgres:postgres /pathto/pgroot
Become the postgres user, and initialize the new cluster:
[postgres]$ initdb -D /pathto/pgroot/data
Some variables need to be overridden in the configuration, as described in Systemd#Editing provided units. First, run:
# systemctl edit postgresql.service
Systemctl will open a drop-in configuration file in your editor. Write the following in that file:
[Service] Environment=PGROOT=/pathto/pgroot PIDFile=/pathto/pgroot/data/postmaster.pid
If you want to use /home
directory for default directory or for tablespaces, add one more line in this file:
ProtectHome=false
Change default encoding of new databases to UTF-8
When creating a new database (e.g. with createdb blog
) PostgreSQL actually copies a template database. There are two predefined templates: template0
is vanilla, while template1
is meant as an on-site template changeable by the administrator and is used by default. In order to change the encoding of a new database, one of the options is to change on-site template1
. To do this, log into PostgreSQL shell (psql
) and execute the following:
First, we need to drop template1
. Templates cannot be dropped, so we first modify it so it is an ordinary database:
UPDATE pg_database SET datistemplate = FALSE WHERE datname = 'template1';
Now we can drop it:
DROP DATABASE template1;
The next step is to create a new database from template0
, with a new default encoding:
CREATE DATABASE template1 WITH TEMPLATE = template0 ENCODING = 'UNICODE';
Now modify template1
so it is actually a template:
UPDATE pg_database SET datistemplate = TRUE WHERE datname = 'template1';
Optionally, if you do not want anyone connecting to this template, set datallowconn
to FALSE
:
UPDATE pg_database SET datallowconn = FALSE WHERE datname = 'template1';
Now you can create a new database:
[postgres]$ createdb blog
If you log back in to psql
and check the databases, you should see the proper encoding of your new database:
\l
List of databases Name | Owner | Encoding | Collation | Ctype | Access privileges -----------+----------+-----------+-----------+-------+---------------------- blog | postgres | UTF8 | C | C | postgres | postgres | SQL_ASCII | C | C | template0 | postgres | SQL_ASCII | C | C | =c/postgres : postgres=CTc/postgres template1 | postgres | UTF8 | C | C |
Administration tools
- phpPgAdmin — Web-based administration tool for PostgreSQL.
- pgAdmin — GUI-based administration tool for PostgreSQL.
Setup HHVM to work with PostgreSQL
$ git clone https://github.com/PocketRent/hhvm-pgsql.git $ cd hhvm-pgsql
If you do not use a nightly build, then run this command (verified on HHVM 3.6.1) to avoid compile errors:
$ git checkout tags/3.6.0
Then build the extension (if you do not need an improved support for Hack language, then remove -DHACK_FRIENDLY=ON):
$ hphpize $ cmake -DHACK_FRIENDLY=ON . $ make
Then copy the built extension:
# cp pgsql.so /etc/hhvm/
Add to /etc/hhvm/server.ini:
extension_dir = /etc/hhvm hhvm.extensions[pgsql] = pgsql.so
Upgrading PostgreSQL
Upgrading minor PostgreSQL versions (i.e. 9.2, 9.3, 9.4, 9.5, 9.6
) requires some extra maintenance.
Quick guide
If you had custom settings in configuration files like pg_hba.conf
and postgresql.conf
, merge them into the new ones. If you have set options such as data_directory
, hba_file
or ident_file
in your postgresql.conf
, please temporally unset them before the migration.
First step: dump your database and stop PostgreSQL
# su -l postgres [postgres]$ cd /var/lib/postgres [postgres]$ mkdir dump [postgres]$ pg_dumpall | gzip > dump/pgdumpall.gz
Stop postgresql.service
.
Second step: prepare the upgrade script
upgrade_pg.sh
# Usage: sh ./upgrade_pg.sh <x.y version to update from> FROM_VERSION="$1" su -l postgres -c "mv /var/lib/postgres/data /var/lib/postgres/data-${FROM_VERSION}" su -l postgres -c 'mkdir /var/lib/postgres/data' su -l postgres -c 'chmod 700 /var/lib/postgres/data' su -l postgres -c "initdb --locale $LANG -E UTF8 -D /var/lib/postgres/data" su -l postgres -c "pg_upgrade -b /opt/pgsql-${FROM_VERSION}/bin/ -B /usr/bin/ -d /var/lib/postgres/data-${FROM_VERSION} -D /var/lib/postgres/data"
Third step: update the packages and install postgresql-old-upgrade
# pacman -Syu postgresql-old-upgrade postgresql postgresql-libs
Fourth step: run the upgrade script
The first time, it is advisable to run one by one the commands of this script.
Fith step: start the database
Start postgresql.service
Troubleshooting
If the pg_upgrade
step fails with the following messages,
- cannot write to log file pg_upgrade_internal.log
- Failure, exiting
- Make sure you are in a directory that the postgres user has enough rights to write the log file to (
/tmp
for example), or usesu -l postgres
instead ofsudo -u postgres
.
- If you are in a directory that postgres user has enough rights to write the log file to however you still get this error then make sure
/var/lib/postgres
is owned by postgres
- LC_COLLATE error that says that old and new values are different
- Figure out what the old locale was,
C
oren_US.UTF-8
for example, and force it when callinginitdb
. -
sudo -u postgres LC_ALL=C initdb -D /var/lib/postgres/data
- There seems to be a postmaster servicing the old cluster.
- Please shutdown that postmaster and try again.
- Make sure postgres is not running. If you still get the error, then chances are there is an old PID file you need to clear out.
- Find the old pid in old pg data,
-
$ sudo -u postgres ls -l /var/lib/postgres/data-9.X total 88 -rw------- 1 postgres postgres 4 Mar 25 2012 PG_VERSION drwx------ 8 postgres postgres 4096 Jul 17 00:36 base drwx------ 2 postgres postgres 4096 Jul 17 00:38 global drwx------ 2 postgres postgres 4096 Mar 25 2012 pg_clog -rw------- 1 postgres postgres 4476 Mar 25 2012 pg_hba.conf -rw------- 1 postgres postgres 1636 Mar 25 2012 pg_ident.conf drwx------ 4 postgres postgres 4096 Mar 25 2012 pg_multixact drwx------ 2 postgres postgres 4096 Jul 17 00:05 pg_notify drwx------ 2 postgres postgres 4096 Mar 25 2012 pg_serial drwx------ 2 postgres postgres 4096 Jul 17 00:53 pg_stat_tmp drwx------ 2 postgres postgres 4096 Mar 25 2012 pg_subtrans drwx------ 2 postgres postgres 4096 Mar 25 2012 pg_tblspc drwx------ 2 postgres postgres 4096 Mar 25 2012 pg_twophase drwx------ 3 postgres postgres 4096 Mar 25 2012 pg_xlog -rw------- 1 postgres postgres 19169 Mar 25 2012 postgresql.conf -rw------- 1 postgres postgres 48 Jul 17 00:05 postmaster.opts -rw------- 1 postgres postgres 80 Jul 17 00:05 postmaster.pid # <-- This is the problem
- Move the old pid to temporary directory,
-
$ sudo -u postgres mv /var/lib/postgres/data-9.X/postmaster.pid /tmp
- ERROR: could not access file "$libdir/postgis-2.0": No such file or directory
- Retrieve
postgis-2.0.so
from postgis for version postgresql 9.X () and copy it to/opt/pgsql-9.X/lib
and make sure the privileges are readable by postgres user.
- Your installation references loadable libraries that are missing from the new installation.
- could not load library "$libdir/postgis-2.2":
- ERROR: incompatible library "/usr/lib/postgresql/postgis-2.2.so": version mismatch
- DETAIL: Server is version 9.6, library is version 9.5.
- You can manually compile the previous version of PostGIS against the new version of PostgreSQL, use it for the upgrade process, then install the new version back.
- In this example we upgrade from the old PostgreSQL 9.5 + PostGIS 2.2 to the new PostgreSQL 9.6 + PostGIS 2.3:
-
# Uninstall PostGIS 2.3. sudo pacman -R postgis # Download and compile PostGIS 2.2 against PostgreSQL 5.6. The latest sources can be found at http://postgis.net/source/ wget http://download.osgeo.org/postgis/source/postgis-2.2.3.tar.gz tar -xvf postgis-2.2.3.tar.gz cd postgis-2.2.3 ./configure make # Copy just the files we need to upgrade the database. sudo cp postgis/postgis-2.2.so /usr/lib/postgresql sudo cp raster/rt_pg/rtpostgis-2.2.so /usr/lib/postgresql sudo cp topology/postgis_topology-2.2.so /usr/lib/postgresql # Run pg_upgrade as described above. It should now finish successfully. [...] # Remove PostGIS 2.2. sudo rm /usr/lib/postgresql/postgis-2.2.so sudo rm /usr/lib/postgresql/rtpostgis-2.2.so sudo rm /usr/lib/postgresql/postgis_topology-2.2.so # Install PostGIS 2.3 back again. sudo pacman -S postgis
- Your installation references loadable libraries that are missing from the new installation.
- You can add these libraries to the new installation, or remove the functions using them from the old installation.
- A list of problem libraries is in the file
- loadable_libraries.txt
- Could not load library "$libdir/pg_upgrade_support"
- ERROR: could not access file "$libdir/pg_upgrade_support": No such file or directory
- It means you have leftovers from old failed pg_upgrade attempts that you never completed. So you'll need to start postgres with old data and
- in each database execute
DROP SCHEMA IF EXISTS binary_upgrade CASCADE;
Detailed instructions
It is recommended to add the following to your /etc/pacman.conf
file:
IgnorePkg = postgresql postgresql-libs
This will ensure you do not accidentally upgrade the database to an incompatible version. When an upgrade is available, pacman will notify you that it is skipping the upgrade because of the entry in pacman.conf
. Minor version upgrades (e.g. 9.0.3 to 9.0.4) are safe to perform. However, if you do an accidental upgrade to a different major version (e.g. 9.0.x to 9.1.x), you might not be able to access any of your data. Always check the PostgreSQL home page (http://www.postgresql.org/) to be sure of what steps are required for each upgrade. For a bit about why this is the case, see the versioning policy.
There are two main ways to upgrade your PostgreSQL database. Read the official documentation for details.
For those wishing to use pg_upgrade
, a postgresql-old-upgrade package is available that will always run one major version behind the real PostgreSQL package. This can be installed side-by-side with the new version of PostgreSQL.
When you are ready, upgrade the following packages: postgresql, postgresql-libs, and postgresql-old-upgrade. Note that the data directory does not change from version to version, so before running pg_upgrade
, it is necessary to rename your existing data directory and migrate into a new directory. The new database must be initialized, as described near the top of this page.
# systemctl stop postgresql.service # su -l postgres -c 'mv /var/lib/postgres/data /var/lib/postgres/olddata' # su -l postgres -c 'initdb --locale en_US.UTF-8 -E UTF8 -D /var/lib/postgres/data'
The upgrade invocation will likely look something like the following. Do not run this command blindly without understanding what it does! Reference the upstream pg_upgrade documentation for details.
# su -l postgres -c 'pg_upgrade -d /var/lib/postgres/olddata/ -D /var/lib/postgres/data/ -b /opt/pgsql-9.4/bin/ -B /usr/bin/'
Manual dump and reload
You could also do something like this (after the upgrade and install of postgresql-old-upgrade).
# systemctl stop postgresql.service # /opt/pgsql-9.4/bin/pg_ctl -D /var/lib/postgres/olddata/ start # /opt/pgsql-9.4/bin/pg_dumpall >> old_backup.sql # /opt/pgsql-9.4/bin/pg_ctl -D /var/lib/postgres/olddata/ stop # systemctl start postgresql.service # psql -f old_backup.sql postgres
Troubleshooting
Improve performance of small transactions
If you are using PostgresSQL on a local machine for development and it seems slow, you could try turning synchronous_commit off in the configuration. Beware of the caveats, however.
/var/lib/postgres/data/postgresql.conf
synchronous_commit = off
Prevent disk writes when idle
PostgreSQL periodically updates its internal "statistics" file. By default, this file is stored on disk, which prevents disks from spinning down on laptops and causes hard drive seek noise. It is simple and safe to relocate this file to a memory-only file system with the following configuration option:
/var/lib/postgres/data/postgresql.conf
stats_temp_directory = '/run/postgresql'
Cannot connect to database through pg_connect()
Install php-pgsql and edit the php.ini
file uncommenting the lines extension=pdo_pgsql.so
and extension=pgsql.so
, then restart httpd
.