Skip to content

Does pgsql-pitr.yml hang for 24h when recovery target is beyond WAL archive on PG13+? #734

@rushsky518-jpg

Description

@rushsky518-jpg

When intentionally specifying a PITR target time that exceeds the available WAL archive on PostgreSQL 13+, does the playbook hang in the [recovery] stage for up to 24 hours before timing out?

#--------------------------------------------------------------#

PITR RECOVERY [recovery]

#--------------------------------------------------------------#

pgbackrest will restore the base data, while you'll need to

start postgres to restore them to the target location

  • name: start postgres recovery
    tags: recovery
    become_user: "{{ pg_dbsu|default('postgres') }}"
    become: true
    block:

    • name: start postgresql with /pg/tmp/recovery.log
      shell: |
      "{{ pg_bin_dir|default('/usr/pgsql/bin') }}/pg_ctl" -W -D {{ pg_data }} -o "-c port={{ pg_port }} -c logging_collector=on -c log_destination=stderr -c log_directory=/pg/tmp -c log_filename=recovery.log" start

    • name: wait for wal replay to recovery target
      tags: recovery
      wait_for:
      path: /pg/tmp/recovery.log
      search_regex: "consistent recovery state reached"
      timeout: 86400
      delay: 5

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions