refactor: increase performance of XSD deployment by creating deduplicated...

refactor: increase performance of XSD deployment by creating deduplicated directories and using ansible.posix.synchronize to copy files

tasks/rosetta/install_format_library_xsds.yml

 ---
 - name: prepare XSD directories
   ansible.builtin.file:
-    path: "/{{ item.dir }}"
+    path: "/{{ item }}"
     owner: "{{ vault_rosetta_user }}"
     group: "{{ vault_rosetta_group }}"
     state: directory
     mode: "0775"
-  loop: "{{ local_jhove_xsds }}"
+  loop: "{{ ( local_jhove_xsds | selectattr('dir', 'defined') | map(attribute='dir') | unique | list ) }}"
 
 - name: install XSDs used in Rosetta format library XML metadata extractor plugin configuration
-  ansible.builtin.copy:
-    src: "{{ item.dir }}{{ item.filename }}"
-    dest: "/{{ item.dir }}{{ item.filename }}"
-    owner: "{{ vault_rosetta_user }}"
-    group: "{{ vault_rosetta_group }}"
-    mode: "0664"
-    checksum: "{{ item.sha1 }}"
+  # 'ansible.builtin.copy' doesn't scale well for hundreds of files, so we use
+  # 'ansible.posix.synchronize' instead
+  # from "man 1 rsync":
+  #  Note that rsync always verifies that each transferred file was correctly
+  #  reconstructed on the receiving side by checking a whole-file checksum that
+  #  is generated as the file is transferred, but that automatic
+  #  after-the-transfer verification has nothing to do with this option's
+  #  before-the-transfer "Does this file need to be updated?" check.
+  ansible.posix.synchronize:
+    src: "{{ item.dir }}{{ item.filename}}"
+    dest: "/{{ item.dir }}{{ item.filename}}"
+    rsync_opts:
+      - "--chmod=F664"
+      - "--chown={{ vault_rosetta_user }}:{{ vault_rosetta_group }}"
   loop: "{{ local_jhove_xsds }}"