postgresql_pg_hba – Add, remove or modify a rule in a pg_hba file

New in version 2.8.

Synopsis

  • The fundamental function of the module is to create, or delete lines in pg_hba files.
  • The lines in the file should be in a typical pg_hba form and lines should be unique per key (type, databases, users, source). If they are not unique and the SID is ‘the one to change’, only one for state=present or none for state=absent of the SID’s will remain.

Requirements

The below requirements are needed on the host that executes this module.

  • ipaddress

Parameters

Parameter Choices/Defaults Comments
address
string
Default:
"samehost"
The source address/net where the connections could come from.
Will not be used for entries of type=local.
You can also use keywords all, samehost, and samenet.

aliases: source, src
attributes
string
added in 2.3
The attributes the resulting file or directory should have.
To get supported flags look at the man page for chattr on the target system.
This string should contain the attributes in the same order as the one displayed by lsattr.
The = operator is assumed as default, otherwise + or - operators need to be included in the string.

aliases: attr
backup
boolean
    Choices:
  • no ←
  • yes
If set, create a backup of the pg_hba file before it is modified. The location of the backup is returned in the (backup) variable by this module.
backup_file
string
Write backup to a specific backupfile rather than a temp file.
contype
string
    Choices:
  • local
  • host
  • hostnossl
  • hostssl
Type of the rule. If not set, postgresql_pg_hba will only return contents.
create
boolean
    Choices:
  • no ←
  • yes
Create an pg_hba file if none exists.
When set to false, an error is raised when the pg_hba file doesn't exist.
databases
string
Default:
"all"
Databases this line applies to.
dest
path / required
Path to pg_hba file to modify.
group
string
Name of the group that should own the file/directory, as would be fed to chown.
method
string
    Choices:
  • cert
  • gss
  • ident
  • krb5
  • ldap
  • md5 ←
  • pam
  • password
  • peer
  • radius
  • reject
  • scram-sha-256
  • sspi
  • trust
Authentication method to be used.
mode
string
The permissions the resulting file or directory should have.
For those used to /usr/bin/chmod remember that modes are actually octal numbers. You must either add a leading zero so that Ansible's YAML parser knows it is an octal number (like 0644 or 01777) or quote it (like '644' or '1777') so Ansible receives a string and can do its own conversion from string into number.
Giving Ansible a number without following one of these rules will end up with a decimal number which will have unexpected results.
As of Ansible 1.8, the mode may be specified as a symbolic mode (for example, u+rwx or u=rw,g=r,o=r).
As of Ansible 2.6, the mode may also be the special string preserve.
When set to preserve the file will be given the same permissions as the source file.
netmask
string
The netmask of the source address.
options
string
Additional options for the authentication method.
order
-
    Choices:
  • sdu ←
  • sud
  • dsu
  • dus
  • usd
  • uds
The entries will be written out in a specific order. With this option you can control by which field they are ordered first, second and last. s=source, d=databases, u=users.
owner
string
Name of the user that should own the file/directory, as would be fed to chown.
selevel
string
Default:
"s0"
The level part of the SELinux file context.
This is the MLS/MCS attribute, sometimes known as the range.
When set to _default, it will use the level portion of the policy if available.
serole
string
The role part of the SELinux file context.
When set to _default, it will use the role portion of the policy if available.
setype
string
The type part of the SELinux file context.
When set to _default, it will use the type portion of the policy if available.
seuser
string
The user part of the SELinux file context.
By default it uses the system policy, where applicable.
When set to _default, it will use the user portion of the policy if available.
state
-
    Choices:
  • absent
  • present ←
The lines will be added/modified when state=present and removed when state=absent.
unsafe_writes
boolean
added in 2.2
    Choices:
  • no ←
  • yes
Influence when to use atomic operation to prevent data corruption or inconsistent reads from the target file.
By default this module uses atomic operations to prevent data corruption or inconsistent reads from the target files, but sometimes systems are configured or just broken in ways that prevent this. One example is docker mounted files, which cannot be updated atomically from inside the container and can only be written in an unsafe manner.
This option allows Ansible to fall back to unsafe methods of updating files when atomic operations fail (however, it doesn't force Ansible to perform unsafe writes).
IMPORTANT! Unsafe writes are subject to race conditions and can lead to data corruption.
users
-
Default:
"all"
Users this line applies to.

Notes

Note

  • The default authentication assumes that on the host, you are either logging in as or sudo’ing to an account with appropriate permissions to read and modify the file.
  • This module also returns the pg_hba info. You can use this module to only retrieve it by only specifying dest. The info can be found in the returned data under key pg_hba, being a list, containing a dict per rule.
  • This module will sort resulting pg_hba files if a rule change is required. This could give unexpected results with manual created hba files, if it was improperly sorted. For example a rule was created for a net first and for a ip in that net range next. In that situation, the ‘ip specific rule’ will never hit, it is in the pg_hba file obsolete. After the pg_hba file is rewritten by the postgresql_pg_hba module, the ip specific rule will be sorted above the range rule. And then it will hit, which will give unexpected results.
  • With the ‘order’ parameter you can control which field is used to sort first, next and last.
  • The module supports a check mode and a diff mode.

Examples

- name: Grant users joe and simon access to databases sales and logistics from ipv6 localhost ::1/128 using peer authentication.
  postgresql_pg_hba:
    dest: /var/lib/postgres/data/pg_hba.conf
    contype: host
    users: joe,simon
    source: ::1
    databases: sales,logistics
    method: peer
    create: true

- name: Grant user replication from network 192.168.0.100/24 access for replication with client cert authentication.
  postgresql_pg_hba:
    dest: /var/lib/postgres/data/pg_hba.conf
    contype: host
    users: replication
    source: 192.168.0.100/24
    databases: replication
    method: cert

- name: Revoke access from local user mary on database mydb.
  postgresql_pg_hba:
    dest: /var/lib/postgres/data/pg_hba.conf
    contype: local
    users: mary
    databases: mydb
    state: absent

Return Values

Common return values are documented here, the following are the fields unique to this module:

Key Returned Description
backup_file
string
changed
File that the original pg_hba file was backed up to

Sample:
/tmp/pg_hba_jxobj_p
msgs
list
always
List of textual messages what was done

Sample:
{'msgs': ['Removing', 'Changed', 'Writing']}
pg_hba
list
always
List of the pg_hba rules as they are configured in the specified hba file

Sample:
{'pg_hba': [{'db': 'all', 'method': 'md5', 'src': 'samehost', 'type': 'host', 'usr': 'all'}]}


Status

Authors

  • Sebastiaan Mannem (@sebasmannem)

Hint

If you notice any issues in this documentation you can edit this document to improve it.