Skip to content

Latest commit

 

History

History
246 lines (181 loc) · 7.31 KB

README.md

File metadata and controls

246 lines (181 loc) · 7.31 KB
Raw

setup_pgd

This role helps in configuration of PGD Node cluster.

Requirements

Following are the requirements of this role.

  1. Ansible
  2. edb_devops.edb_postgres -> setup_repo role for setting the repository on the systems.
  3. edb_devops.edb_postgres -> install_dbserver role for installing the necessary packages on at least the lead primary.
  4. edb_devops.edb_postgres -> init_dbserver role for initializing the database cluster on the lead primary.

Role Variables

edb_enable_tde

Flag to enable TDE in the cluster. If true, then additional TDE variables must be defined. See init_dbserver role for additional TDE variables. Default: false.

pgd_version

Version of PGD to install and configure. Default: 5

install_pgd

Flag for installing PGD packages. Default: false.

pgd_cluster_database

Name of the PGD cluster database.

pgd_cluster_database_owner

Username of PGD database owner.

pgd_cluster_database_owner_password

Password for PDG cluster database owner. If none is provided, it will be generated during cluster initialization.

pgd_replication_user

Username of PGD replication user.

pgd_replication_user_password

Password for PGD replication user. If none is provided, it will be generated during cluster initialization.

pgd_cluster_name

Name of PGD cluster to be initialized.

pgd_local_node_name

Hostname of local node.

setup_pgd_cli

When true, the PGD CLI will be installed and configured. Default: false.

pgd_commit_scopes

Sets either CAMO or group commit PGD commit scopes to be configured. For more information on PGD Commit Scopes and the caveats you need to consider before implementing, please refer to the documentation..

For each commit scope, you need to set the:

  • cs_name - name of the commit scope
  • cs_type - type of commit scope. Options are CAMO or GROUP_COMMIT.
  • parent_group - name of the top level group the member nodes belong to, usually pgd_cluster_name.
  • cs_origin_node_group - name of the subgroup to be created that each node in member_nodes will join.
  • member_nodes - list of hostnames of nodes that will be part of the commit scope.
  • default_group_cs - flag to set the commit scope as the default commit scope for the cs_origin_node_group.
  • cs_rule - the commit scope rule to be applied to the commit scope. This must be correctly formatted and valid.

The commit scope configuration requirements are:

  1. CAMO and group commit are incompatible with each another. Only one can be enabled per cluster.
  2. When cs_type: CAMO, only two member data nodes, also referred to as a CAMO pair, can be present.
  3. Each host can only be a part of at most one commit scope.
  4. The cs_rule parameter must be: valid, correctly formatted, and adhere to the correct syntax
# CAMO example
pgd_commit_scopes:
  - cs_name: 'camo_scope_1'
    cs_type: 'CAMO'
    parent_group: 'pgd_cluster'
    cs_origin_node_group: 'pgd_camo_group_1'
    member_nodes: ['pgd1', 'pgd2']
    default_group_cs: true
    cs_rule: "ALL (pgd_camo_group_1) ON visible CAMO DEGRADE ON (timeout=500s) TO ASYNC"
# group commit example
pgd_commit_scopes:
  - cs_name: 'groupcommit_scope_1'
    cs_type: 'GROUP_COMMIT'
    parent_group: 'pgd_cluster'
    cs_origin_node_group: 'pgd_gc_group_1'
    member_nodes: ['pgd1', 'pgd2', 'pgd3', 'pgd4']
    default_group_cs: true
    cs_rule: "ANY 2 (pgd_gc_group_1) GROUP COMMIT"

The rest of the variables can be configured and are available in the:

Host Variables

node_kind

The role of node to initialize within PGD. Options are data, subscribe-only, witness and standby.

use_physical_backup

Set to true if you want to initialize non-lead-primary nodes from physical backup.

clustername

Name of cluster node belongs to.

lead_primary

Set to true if node is lead primary.

location

HA location of node.

Dependencies

This role does not have any dependencies, but package repositories should have been configured beforehand with the setup_repo role. At least one lead primary must exist and a database cluster must be initialized on that node.

Example Playbook

Inventory file content

Content of the inventory.yml file:

---
all:
  children:
    primary:
      hosts:
        edb-primary1:
          ansible_host: xxx.xxx.xxx.xxx
          private_ip: xxx.xxx.xxx.xxx
          location: PGD_DC1
          pgd:
            node_kind: data
            lead_primary: true
            clustername: pgdcluster
        edb-primary2:
          ansible_host: xxx.xxx.xxx.xxx
          private_ip: xxx.xxx.xxx.xxx
          location: PGD_DC1
          pgd:
            node_kind: data
            lead_primary: false
            clustername: pgdcluster
            use_physical_backup: true
            upstream_node_private_ip: xxx.xxx.xxx.xxx
        edb-primary3:
          ansible_host: xxx.xxx.xxx.xxx
          private_ip: xxx.xxx.xxx.xxx
          location: PGD_DC1
          pgd:
            node_kind: data
            lead_primary: false
            clustername: pgdcluster
            use_physical_backup: false

How to include the setup_pgd role in your Playbook

Below is an example of how to include the setup_pgd role:

- hosts: all
  name: Postgres deployment playbook
  become: true
  gather_facts: true
  any_errors_fatal: true
  max_fail_percentage: 0

  collections:
    - edb_devops.edb_postgres

  pre_tasks:
    - name: Initialize the user defined variables
      ansible.builtin.set_fact:
        pg_version: 15            # Change the version of Postgres you want to install
        pg_type: "EPAS"           # Change the pg_type to EPAS if EDB Advanced Server is needed
        repo_token: "XXXXXXXXX"   # EDB repo 2.0 token
        install_pgd: true         # Install PGD flag for installing PGD
        pgd_version: 5            # Postgres Distributed version
        enable_pgdg_repo: true    # flag/variable for enabling PGD repo

  roles:
    - role: setup_repo
      when: "'setup_repo' in lookup('edb_devops.edb_postgres.supported_roles', wantlist=True)"
    - role: install_dbserver
      when: "'install_dbserver' in lookup('edb_devops.edb_postgres.supported_roles', wantlist=True)"
    - role: init_dbserver
      when: "'init_dbserver' in lookup('edb_devops.edb_postgres.supported_roles', wantlist=True)"
    - role: setup_pgd
      when: "'setup_pgd' in lookup('edb_devops.edb_postgres.supported_roles', wantlist=True)"

Defining and adding variables is done in the set_fact of the pre_tasks.

All the variables are available at:

License

BSD

Author Information

Author: