This cookbook includes recipes and providers to install and configure postgresql database. This cookbook was tested with Postgresql 9.1, 9.2, 9.3, 9.4, 9.5, 9.6 & 10.
Supported platforms:
- Debian 8
- Debian 9
- Ubuntu 14.04
- Ubuntu 16.04
- Ubuntu 18.04
Note: TravisCI tests for Ubuntu 18.04 are omitted now because they somehow hang. Local Vagrant & Docker-based tests are succesfull. This will be investigated further.
See CHANGELOG.md
The minimal recommended version of chef-client is 13.0.113
. It may still work on version 12.5.1
and older, but no tests are made starting from version 1.3.0
of this cookbook as Chef 12 is reaching its EOL in the April, 2018
- apt
- cron
- poise-python
This cookbook have server and client attribute files.
With client attributes(["postgresql"]["client"]) you can set only postgresql client and library version.
Server attributes are starting from ["postgresql"]["defaults"] and used as default attributes for postgresql provider. You should not override this defaults, you can pass your settings to provider instead.
- :create: creates postgresql cluster
- cluster_name: name attribute. Cluster name (e.g. main). Be aware, systemd (in Ubuntu 16.04 and Debian Jessie) not working with cluster names that containing dashes ('-').
- cluster_version: set cluster version
- cookbook: cookbook for templates. Skip this for default templates.
- cluster_create_options: options for pg_createcluster (only locale related options)
- configuration: Hash with configuration options for postgresql, see examples.
- hba_configuration: Array with hba configuration, see examples.
- ident_configuration: Array with ident configuration, see examples.
- replication: Hash with replication configuration. See replication example.
- replication_initial_copy: Boolean. If
true
pg_basebackup will be exec to make initial replication copy. Default isfalse
. - replication_start_slave: Boolean. If
true
slave cluster will be started after creation. Should be used with replication_initial_copy option. Defaultfalse
. - allow_restart_cluster: Can be
first
,always
ornone
. Specifies when cluster must restart instead of reload.first
– only first time after installation.always
– always restart, even if changes doesn't require restart.none
- never, use reload every time. Default isnone
.
postgresql_cloud_backup_helper.sh
helper can be found at /opt/wal-e/bin/
.
postgresql_cloud_backup_helper.sh <cluster_name> <cluster_version> last|count
cluster_name
– postgresql cluster name (ex. main)cluser_version
– postgresql cluser version (ex. 9.3)last
– shows last backup timecount
– shows total number of backups.
Example master database setup:
postgresql 'main' do
cluster_version '9.3'
cluster_create_options( locale: 'ru_RU.UTF-8' )
configuration(
listen_addresses: '192.168.0.2',
max_connections: 300,
ssl_renegotiation_limit: 0,
shared_buffers: '512MB',
maintenance_work_mem: '64MB',
work_mem: '8MB',
log_min_duration_statement: 200
)
hba_configuration(
[
{ type: 'host', database: 'all', user: 'all', address: '192.168.0.0/24', method: 'md5' },
{ type: 'host', database: 'replication', user: 'postgres', address: '192.168.0.3/32', method: 'trust' }
]
)
end
Example slave database setup:
postgresql 'main' do
cluster_version '9.3'
cluster_create_options( locale: 'ru_RU.UTF-8' )
configuration(
listen_addresses: '192.168.0.3',
max_connections: 300,
ssl_renegotiation_limit: 0,
shared_buffers: '512MB',
maintenance_work_mem: '64MB',
work_mem: '8MB',
log_min_duration_statement: 200
)
hba_configuration(
[
{ type: 'host', database: 'all', user: 'all', address: '192.168.0.0/24', method: 'md5' },
{ type: 'host', database: 'replication', user: 'postgres', address: '192.168.0.2/32', method: 'trust' }
]
)
replication(
standby_mode: 'on',
primary_conninfo: 'host=192.168.0.1',
trigger_file: '/tmp/pgtrigger'
)
replication_initial_copy true
replication_start_slave true
end
Example slave configuration with replication slots (PostgreSQL >= 9.4)
replication(
standby_mode: 'on',
primary_conninfo: 'host=192.168.0.1',
trigger_file: '/tmp/pgtrigger'
primary_slot_name: 'some_slot_on_master'
)
Don't forget to create slot on master server before:
# SELECT pg_create_physical_replication_slot('some_slot_on_master');
Example users and databases setup
postgresql_user 'user01' do
in_version '9.3'
in_cluster 'main'
unencrypted_password 'user01password'
end
postgresql_database 'database01' do
in_version '9.3'
in_cluster 'main'
owner 'user01'
end
Example full daily database backup
postgresql_cloud_backup 'main' do
utility 'wal-g'
in_version '9.3'
in_cluster 'main'
full_backup_time weekday: '*', month: '*', day: '*', hour: '3', minute: '0'
# Data bag item should contain following keys for S3 protocol:
# aws_access_key_id, aws_secret_access_key, wale_s3_prefix
parameters Chef::EncryptedDataBagItem.load('s3', 'secrets').to_hash.select {|i| i != "id"}
# Or just a hash, if you don't use data bags:
parameters { aws_access_key_id: 'access_key', aws_secret_access_key: 'secret_key', walg_s3_prefix: 's3_prefix' }
# In case you need to prepend wal-e with, for example, traffic limiter
# you can use following method:
command_prefix 'trickle -s -u 1024'
# It will be prepended to resulting wal-e execution in cron task
end
Example usage of cloud backup helper usage
$ /opt/wal-e/bin/postgresql_cloud_backup_helper.sh main 9.3 last
1428192159
$ /opt/wal-e/bin/postgresql_cloud_backup_helper.sh main 9.3 count
31
Example of how to install extensions from postgresql-contrib NOTE: schema and version are optional parameters, but others are required
postgresql_extension 'cube' do
in_version '9.4'
in_cluster 'main'
db 'test01'
schema 'public'
end
Example of how to install extensions from http://pgxn.org/ NOTE: schema is an optional parameter, but others are required
pgxn_extension 'pg_lambda' do
in_version '9.4'
in_cluster 'main'
db 'test01'
version '1.0.2'
stage 'stable'
end
Maintainer:: LLC Express 42 ([email protected]) Source:: https://github.com/express42/postgresql_lwrp Issues:: https://github.com/express42/postgresql_lwrp/issues
License:: MIT