Ansible Role: restic

Beta: This role is in beta status.

Description

Restic is a versatile Go based backup solution which supports multiple backends, deduplication and incremental backups.

This role installs restic on a client, configures the backup repositories and optionally sets systemd timer or cronjobs to run the backups. Aditionally, it will setup executable scripts to run a Backup manually.

This Project borrowed heavily from the donat-b/ansible-restic and the https://github.com/arillso/ansible.restic ansible role. We try to make this role more easy to anderstand and modern by using systemd timer, /etc/crontab to define backup paths, more absolute paths and less options. (no S3 Storage, No Windows...)

Backup Scripts

This role will create a backup script and a file with credentials usable with the source command on linux for each backup in the restic_script_dir. These executable scripts can be used to manually trigger a backup action, but are also used for automated backups if you have set restic_create_schedule variable to true. Make sure to not change the files manually, as this can interfere with your backups quite a bit.

on Linux, if you want to take a manual snapshot, you can run the backup like this:

$ /path/to/backup/script/backup-example.sh

by default, such a snapshot will be given the tag manual, so you can distinguish them from automatically created snapshots. You can also append more tags by simply appending them:

$ /path/to/backup/script/backup-example.sh --tag deployment

CRON / Scheduled Tasks

In order to make use of defined backups, they can be automatically setup as scheduled tasks. You have to be aware of the fact that (on linux systems at least) you need to have administrator permissions for configuring such an action.

If you cannot use the automatic creation of the tasks, you can still make use of the generated scripts. If you are for example on a shared hosting server and can define a cronjob via a webinterface, simply add each backup file to be executed. Make sure to prefix the command with CRON=true to imply that the snapshot was created via a scheduled task:

CRON=true /path/to/backup/script/backup-example.sh

Installation

There are multiple ways to install the role. Either clone or download it directly from the github repository. Or Install it via ansible galaxy:

ansible-galaxy install roles-ansible.restic

Requirements

• bzip2
• jmespath

Role Variables

Name	Default	Description
restic_url	undefined	The URL to download restic from. Use this variable to overwrite the default
restic_version	'0.15.1'	The version of Restic to install
restic_download_path	'/opt/restic'	Download location for the restic binary
restic_install_path	'/usr/local/bin'	Install location for the restic binary
restic_script_dir	'/opt/restic'	Location of the generated backup scripts
restic_log_dir	'{{ restic_script_dir }}/log'	Location of the logs of the backup scripts
restic_repos	{}	A dictionary of repositories where snapshots are stored. (More Info: Repos)
restic_backups	{} (or [])	A list of dictionaries specifying the files and directories to be backed up (More Infos: Backups)
restic_create_schedule	false	Should we schedule each backup? Either via cronjob or via systemd timer.
restic_backup_now	false	Whether or not the backup script should be run immediately
restic_schedule_type	systemd	Here you can define if we create a cronjob or a systemd timer. If it fails to create a systemd timer, a cronjob will be created.
restic_dir_owner	'{{ansible_user}}'	The owner of all created dirs
restic_dir_group	'{{ansible_user}}'	The group of all created dirs
restic_no_log	true	Set to false to see hidden ansible logs
restic_do_not_cleanup_cron	false	We changed the cron location and clean up the old one. You can skip the cleanup here
restic__cache_config	false	Configure custom cache directory
restic__cache_dir	'~/.cache/restic'	Define custom cache directory
submodules_versioncheck	false	If you set this variable to true, the role will run a simple versionscheck to prevent running older versions of this role.
restic__limit_cpu_usage	false	Should CPU usage be limited?
restic__max_cpus	1	Maximum number of CPUs that can be used simultaneously

Repos

Restic stores data in repositories. You have to specify at least one repository to be able to use this role. A repository can be local or remote (see the official documentation).

Using an SFTP repository

For using an SFTP backend, the user needs passwordless access to the host. Please make sure to distribute ssh keys accordingly, as this is outside of the scope of this role.

Available variables:

Name	Required	Description
location	yes	The location of the Backend. Currently, Local, SFTP, S3, Azure Blob and B2 are supported
password	yes	The password used to secure this repository
init	no	Describes if the repository should be initialized or not. Use false if you are backuping to an already existing repo.

Example:

restic_repos:
  local:
    location: /srv/restic-repo
    password: securepassword1
    init: true
  sftp:
    location: sftp:user@host:/srv/restic-repo
    password: securepassword2
    init: true
  aws:
    location: s3:s3.amazonaws.com/bucket_name
    password: securepassword3
    init: true
    aws_access_key: accesskey
    aws_secret_access_key: secretaccesskey
    aws_default_region: eu-west-1
  azure:
    location: azure:container:/
    password: securepassword4
    init: true
    azure_account_name: storageaccountname
    # Only one of the following are required
    azure_account_key: somekey
    azure_account_sas: sasurl
    # Optional
    azure_endpoint_suffix: core.windows.net
  b2:
    location: b2:bucketname:path/to/repo
    password: securepassword5
    init: true
    b2_account_id: accountid
    b2_account_key: accountkey

Backups

A backup specifies a directory or file to be backuped. A backup is written to a Repository defined in restic_repos.

Available variables:

Name	Required (Default)	Description
name	yes	The name of this backup. Used together with pruning and scheduling and needs to be unique.
repo	yes	The name of the repository to backup to.
src	yes (unless stdin == true)	The source directory or file
stdin	no	Is this backup created from a stdin?
stdin_cmd	no (yes if stdin == true)	The command to produce the stdin.
stdin_filename	no	The filename used in the repository.
tags	no	Array of default tags
keep_last	no	If set, only keeps the last n snapshots.
keep_hourly	no	If set, only keeps the last n hourly snapshots.
keep_daily	no	If set, only keeps the last n daily snapshots.
keep_weekly	no	If set, only keeps the last n weekly snapshots.
keep_monthly	no	If set, only keeps the last n monthly snapshots.
keep_yearly	no	If set, only keeps the last n yearly snapshots.
keep_within	no	If set, only keeps snapshots in this time period.
keep_tag	no	If set, keep snapshots with this tags. Make sure to specify a list.
prune	no (false)	If true, the restic forget command in the script has the --prune option appended.
scheduled	no (false)	If restic_create_schedule is set to true, this backup is scheduled and tries to create a systemd timer unit. If it fails, it is creating a cronjob.
schedule_oncalendar	'*-*-* 02:00:00'	The time for the systemd timer. Please notice the randomDelaySec option. By Default the backup is done every night at 2 am (+0-4h). But only if scheduled is true.
schedule_minute	no (*)	Minute when the job is run. ( 0-59, *, */2, etc )
schedule_hour	no (2)	Hour when the job is run. ( 0-23, *, */2, etc )
schedule_weekday	no (*)	Weekday when the job is run. ( 0-6 for Sunday-Saturday, *, etc )
schedule_month	no (*)	Month when the job is run. ( 1-12, *, */2, etc )
exclude	no ({})	Allows you to specify files to exclude. See Exclude for reference.
disable_logging	no	Optionally disable logging
log_to_journald	no	Optionally switch logging to journald with the name of the backup job as the tag
mail_on_error	no	Optionally send a mail if the backupjob will fail (mailx is required)
mail_address	if mail_on_error is true	The mail addressto recive mails if you enabled mail_on_error.
monitoring_call	no	A command that will be called if the backup is successful. Useful for heartbeat monitoring systems that warn when no heartbeat is received. Use the full command, you need to run. Example: curl https://monitoring.example.com/api/push/E9Wzm4lJ2O?status=up&msg=OK&ping=
niceness	no	If set, runs any scheduled backup with given niceness-value. On Linux -20 is highest priority, 0 default and 19 is lowest priority. 10 is a common low priority assigned to backup routines on production systems.

Example:

restic_backups:
  data:
    name: data
    repo: remote
    src: /path/to/data
    scheduled: true
    schedule_oncalendar: '*-*-* 01:00:00'
  database:
    name: database
    repo: remote
    stdin: true
    stdin_cmd: pg_dump -Ubackup db_name
    stdin_filename: db_name_dump.sql
    scheduled: true
    schedule_oncalendar: '*-*-* 01:30:00'
    niceness: 10

You can also specify restic_backups as an array, which is a legacy feature and might be deprecated in the future. currently, the name key is used for namint the access and backup files

Exclude

the exclude key on a backup allows you to specify multiple files to exclude or files to look for filenames to be excluded. You can specify the following keys:

exclude:
    exclude_caches: true
    exclude:
        - /path/to/file
    iexclude:
        - /path/to/file
    exclude_file:
        - /path/to/file
    exclude_if_present:
        - /path/to/file

Please refer to the use of the specific keys to the documentation.

Dependencies

This role does not have any other ansible role as dependencie.

Example Playbook

- name: backup your homefolders to /mnt/backup everyday night
  hosts: localhost
  roles:
    - {role: do1jlr.restic, tags: restic}
  vars:
    restic_create_schedule: true
    restic_repos:
      local:
        location: '/mnt/backup'
        password: 'ChangM3'
        init: true
    restic_backups:
      home:
        name: home
        repo: local
        src: /home/
        scheduled: true
        schedule_oncalendar: '*-*-* 01:00:00'

License

This project is under the MIT License. See the LICENSE file for the full license text.