bacula

ansible-role-bacula

This Role installs any component of the backup system "bacula" AND configures everything (including agent configs like filesets for all systems) so you can use bacula gitops-style (aka do everything from ansible).

quick bacula introduction

Component	function	variable (ansible-tag)
bacula-director (dir)	central management, coordination on other components	bacula_dir_role (bacula-dir)
bacula-storage daemon (sd)	controls the writing device(s)	bacula_sd_role (bacula-sd)
bacula-console	management console, connects to bacula-dir	bacula_console_role (bacula-console)
bacula-file daemon (fd)	agent on target system	bacula_fd_role (bacula-fd)

see the notes on firewalling: https://www.bacula.org/9.6.x-manuals/en/problems/Dealing_with_Firewalls.html

Requirements

• installed database-server (mysql or pgsql) on director

Limitations

Currently only tested on mysql / mariadb and Ubuntu/Debian in LTS-Versions.

Quick start

customize (at least) this vars:

(see example playbooks).

bacula_dir_fqdn: "bacula-dir.DOMAIN.TLD" bacula_sd_fqdn: "bacula-sd.DOMAIN.TLD"

bacula_dir_restore_path: "/var/bacula-restores" (Default restore path) bacula_sd_archive_device: "/var/bacula-volumes" (Default volume path)

-> DB engine (can be 'pgsql' or 'mysql') bacula_dir_db_engine: mysql

Define the director/SD via (host_vars / group_vars):

host_vars/bacula-dir.DOMAIN.TLD.yml

---
bacula_console_role: True
bacula_dir_role: True
bacula_sd_role: True
bacula_fd_role: True

Variables

defaults

defaults are set here:

• defaults/main.yml
• vars/{{ ansible_os_family }}.yml

vaulted passwords

my recommendation: set the following variables in your vault:

bacula_console_password: ""
bacula_dir_dbpassword: ""
bacula_mon_password: ""
bacula_sd_password: ""

Attention: empty passwords are substituted with random values - on very run. Useful for testing - but not in production. Exception are the FD-passwords, they will be preserved when bacula_fd_auto_psk is True (default).

customize storage device (tape, Filestorage etc.)

Example: Ringbuffer

• oldest data is recycled/overwritten when the storage is full
• size is bacula_sd_pool_max_volumes x bacula_sd_pool_max_volume_byte Example: 10G x 100 Volumes means maximum capacity is ~1TB backupspace, depending on the amount

see template from var "bacula_sd_device_template" (default is storage_device_include_ringbuffer.j2) and optional more options from var "bacula_sd_device_extra_options".

Storage Definitions

bacula_sd_device_name: "RingBuffer" bacula_sd_media_type: "File"

where to store the volumes:

bacula_sd_archive_device: "/srv/bacula" bacula_dir_pool_name: "Files" bacula_dir_pool_label_format: "Bacula-Vol-" bacula_dir_pool_recycle: "yes"

bacula_dir_pool_file_retention: "60 days" # Default: 60 days

bacula_dir_pool_job_retention: "24 months" # Default: 6 months

Prune expired Jobs/Files:

bacula_dir_pool_autoprune: "yes" bacula_dir_pool_volume_rention: "3650 days" bacula_dir_pool_max_volume_bytes: "10G" bacula_dir_pool_max_volumes: 100

Host overrides

Example: host_vars/host1.DOMAIN.TLD.yml

own fileset:

bacula_fd_fileset_name: "fileset-of-host1"
bacula_fd_fileset_includes: |+
File = "/files/only/on/host1"
File = "/etc"
File = "/root"
bacula_fd_fileset_excludes: |+
File = "/files/EXCLUDED/on/host1"
File = "/proc"

define jobs that the client should run before (recommended to put this into group_vars or hosts_vars !):

bacula_fd_client_run_before_job: command
bacula_fd_client_run_after_job: command

Example:

bacula_fd_client_run_before_job: "/path/to/my/backupskript create"
bacula_fd_client_run_after_job: "/path/to/my/backupskript cleanup"

additonal note: for jobs running on director/server-side the vars "bacula_fd_run_before_job", "bacula_fd_run_after_job" or "bacula_fd_run_after_failed_job" are existing.

additional options in job-ressource:

bacula_fd_extra_job_options: |+
Option1 = Example1
Option2 = Example2

Example: add more (than the predefined "bacula_fd_client_run_before_job") "Client Run Before Job" via bacula_fd_extra_job_options:

bacula_fd_extra_job_options: |+
    Client Run Before Job = "additional before job" 
    Client Run After Job = "additional after job"

same goes for client and jobdef-ressources:

bacula_fd_extra_client_options bacula_fd_extra_jobdefs_options

different FQDN for connection to fd:

bacula_fd_connect_address: "different.fqdn.to.connect.to"

set source IP (default: the kernel will choose the best address according to the routing table):

bacula_fd_source_address: "IP_or_FQDN"

specific schedule for a host:

• set a (valid!) name for var bacula_fd_schedule (i.e. in host_vars/$hostname.yml with content "bacula_fd_schedule: MoreFrequentSchedule").

Extend global schedules:

• see "bacula-dir-fileset-default.conf.j2" for a example.
• create your own config snippet (jinja2 template) with additional Schedules (i.e.: $ansibleroot/templates/bacula_dir_schedules_extra.j2)
• define the variable "bacula_dir_schedule_extra_template" with the filename of your template ("bacula_dir_schedule_extra.j2"), so it is included by the standard template bacula-dir-schedules.j2 : example: add in "group_vars/bacula_dir.yml" this line: bacula_dir_schedule_extra_template: "bacula_dir_schedules_extra.j2

Extend global filesets:

same as schedules but the var is "bacula_dir_fileset_extra"