acquire_data.yml

---
# Acquires the data from an archive database.
# See also, https://github.com/Connexions/cnx-archive/issues/197

# Note, this host, 'data_source', is specific to this playbook.
# It's this way on purpose. You'll be grabbing data from a live
# server.

# Please DO NOT use this during peak hours!

- name: gather info
  hosts: data_source
  vars_prompt:
    - name: "db_name"
      prompt: "What database are we extracting?"
      default: repository
    - name: "db_user"
      prompt: "What database user should we use?"
      default: postgres
  vars:
    keep_filenames:
      - "index.cnxml"
      - "index.cnxml.html"
      - "index_auto_generated.cnxml"
      - "ruleset.css"
      - "featured-cover.png"
      - "collection.xml"
    keep_filename_str: "{{ keep_filenames|map('tojson')|join(',')|regex_replace('\"', \"'\") }}"
  tasks:
    - name: dump database without files
      shell: "nice -n {{ nice_priority }} pg_dump -U {{ db_user }} --exclude-table-data=files --exclude-table-data=module_files {{ db_name }} | gzip > cnxarchive_dump_without_files.sql.gz"
    - name: dump database files table for index.cnxml, index.cnxml.html, index_auto_generated.cnxml
      shell: "nice -n {{ nice_priority }} psql -U {{ db_user }} {{ db_name }} -c \"copy ( SELECT fileid, md5, sha1, file, media_type FROM files WHERE fileid IN (SELECT fileid FROM module_files WHERE filename IN ({{ keep_filename_str }}) ) ) TO STDOUT\" | gzip > cnxarchive_index_files.txt.gz"
    - name: dump database files table for other files
      shell: "nice -n {{ nice_priority }} psql -U {{ db_user }} {{ db_name }} -c \"copy ( SELECT fileid, md5, sha1, 'dummy file', media_type FROM files WHERE fileid NOT IN (SELECT fileid FROM module_files WHERE filename IN ({{ keep_filename_str }}) ) ) TO STDOUT\" | gzip > cnxarchive_other_files.txt.gz"

    - name: dump database module files table for index.cnxml, index.cnxml.html, index_auto_generated.cnxml
      shell: "nice -n {{ nice_priority }} psql -U {{ db_user }} {{ db_name }} -c \"copy ( SELECT module_ident, fileid, filename FROM module_files ) TO STDOUT\" | gzip > cnxarchive_index_module_files.txt.gz"
    - name: produce dump filename
      shell: "echo \"cnxarchive_dump.$(date +'%Y-%m-%d').tar\""
      register: dump_filename
    - name: create dump .tar.gz file
      shell: "nice -n {{ nice_priority }} tar cf {{ dump_filename.stdout }} --remove-files cnxarchive_dump_without_files.sql.gz cnxarchive_index_files.txt.gz cnxarchive_other_files.txt.gz cnxarchive_index_module_files.txt.gz"
    - name: download dump .tar file
      fetch:
        src: "{{ dump_filename.stdout }}"
        dest: "cnxarchive_dump.tar"
        flat: yes