svc-bkp-rmt-2-loc: migrate pull script to Python + add unit tests; lock down backup-provider ACLs

- Replace Bash pull-specific-host.sh with Python pull-specific-host.py (argparse, identical logic)
- Update role vars and runner template to call python script
- Add __init__.py files for test discovery/imports
- Add unittest: tests/unit/roles/svc-bkp-rmt-2-loc/files/test_pull_specific_host.py (mocks subprocess/os/time; covers success, no types, find-fail, retry-exhaustion)
- Backup provider SSH wrapper: align allowed ls path (backup-docker-to-local)
- Split user role tasks: 01_core (sudoers), 02_permissions_ssh (SSH keys + wrapper), 03_permissions_folders (ownership + default ACLs + depth-limited chown/chmod)
- Ensure default ACLs grant rwx to 'backup' and none to group/other; keep sudo rsync working

Ref: ChatGPT discussion (2025-10-14) — https://chatgpt.com/share/68ee920a-9b98-800f-8806-ddcfe0255149

roles/svc-bkp-rmt-2-loc/files/pull-specific-host.py

@@ -0,0 +1,132 @@
+#!/usr/bin/env python3
+import argparse
+import os
+import subprocess
+import time
+import sys
+
+
+def run_command(command, capture_output=True, check=False, shell=True):
+    """Run a shell command and return its output as string."""
+    try:
+        result = subprocess.run(
+            command,
+            capture_output=capture_output,
+            shell=shell,
+            text=True,
+            check=check
+        )
+        return result.stdout.strip()
+    except subprocess.CalledProcessError as e:
+        if capture_output:
+            print(e.stdout)
+            print(e.stderr)
+        raise
+
+
+def pull_backups(hostname: str):
+    print(f"pulling backups from: {hostname}")
+    errors = 0
+
+    print("loading meta data...")
+    remote_host = f"backup@{hostname}"
+    print(f"host address:         {remote_host}")
+
+    remote_machine_id = run_command(f'ssh "{remote_host}" sha256sum /etc/machine-id')[:64]
+    print(f"remote machine id:    {remote_machine_id}")
+
+    general_backup_machine_dir = f"/Backups/{remote_machine_id}/"
+    print(f"backup dir:           {general_backup_machine_dir}")
+
+    try:
+        remote_backup_types = run_command(
+            f'ssh "{remote_host}" "find {general_backup_machine_dir} -maxdepth 1 -type d -execdir basename {{}} ;"'
+        ).splitlines()
+        print(f"backup types:          {' '.join(remote_backup_types)}")
+    except subprocess.CalledProcessError:
+        sys.exit(1)
+
+    for backup_type in remote_backup_types:
+        if backup_type == remote_machine_id:
+            continue
+
+        print(f"backup type:              {backup_type}")
+
+        general_backup_type_dir = f"{general_backup_machine_dir}{backup_type}/"
+        general_versions_dir = general_backup_type_dir
+
+        # local previous version
+        try:
+            local_previous_version_dir = run_command(f"ls -d {general_versions_dir}* | tail -1")
+        except subprocess.CalledProcessError:
+            local_previous_version_dir = ""
+        print(f"last local backup:      {local_previous_version_dir}")
+
+        # remote versions
+        remote_backup_versions = run_command(
+            f'ssh "{remote_host}" "ls -d /Backups/{remote_machine_id}/backup-docker-to-local/*"'
+        ).splitlines()
+        print(f"remote backup versions:   {' '.join(remote_backup_versions)}")
+
+        remote_last_backup_dir = remote_backup_versions[-1] if remote_backup_versions else ""
+        print(f"last remote backup:       {remote_last_backup_dir}")
+
+        remote_source_path = f"{remote_host}:{remote_last_backup_dir}/"
+        print(f"source path:              {remote_source_path}")
+
+        local_backup_destination_path = remote_last_backup_dir
+        print(f"backup destination:       {local_backup_destination_path}")
+
+        print("creating local backup destination folder...")
+        os.makedirs(local_backup_destination_path, exist_ok=True)
+
+        rsync_command = (
+            f'rsync -abP --delete --delete-excluded --rsync-path="sudo rsync" '
+            f'--link-dest="{local_previous_version_dir}" "{remote_source_path}" "{local_backup_destination_path}"'
+        )
+        print("starting backup...")
+        print(f"executing:                {rsync_command}")
+
+        retry_count = 0
+        max_retries = 12
+        retry_delay = 300  # 5 minutes
+        last_retry_start = 0
+        max_retry_duration = 43200  # 12 hours
+
+        rsync_exit_code = 1
+        while retry_count < max_retries:
+            print(f"Retry attempt: {retry_count + 1}")
+            if retry_count > 0:
+                current_time = int(time.time())
+                last_retry_duration = current_time - last_retry_start
+                if last_retry_duration >= max_retry_duration:
+                    print("Last retry took more than 12 hours, increasing max retries to 12.")
+                    max_retries = 12
+            last_retry_start = int(time.time())
+            rsync_exit_code = os.system(rsync_command)
+            if rsync_exit_code == 0:
+                break
+            retry_count += 1
+            time.sleep(retry_delay)
+
+        if rsync_exit_code != 0:
+            print(f"Error: rsync failed after {max_retries} attempts")
+            errors += 1
+
+    sys.exit(errors)
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        description="Pull backups from a remote backup host via rsync."
+    )
+    parser.add_argument(
+        "hostname",
+        help="Hostname from which backup should be pulled"
+    )
+    args = parser.parse_args()
+    pull_backups(args.hostname)
+
+
+if __name__ == "__main__":
+    main()

roles/svc-bkp-rmt-2-loc/files/pull-specific-host.sh

@@ -1,85 +0,0 @@
-#!/bin/bash
-# @param $1 hostname from which backup should be pulled
-
-echo "pulling backups from: $1" &&
-
-# error counter
-errors=0 &&
-
-echo "loading meta data..." &&
-
-remote_host="backup@$1" &&
-echo "host address:         $remote_host" &&
-
-remote_machine_id="$( (ssh "$remote_host" sha256sum /etc/machine-id) | head -c 64 )" &&
-echo "remote machine id:    $remote_machine_id" &&
-
-general_backup_machine_dir="/Backups/$remote_machine_id/" &&
-echo "backup dir:           $general_backup_machine_dir" &&
-
-remote_backup_types="$(ssh "$remote_host" "find $general_backup_machine_dir -maxdepth 1 -type d -execdir basename {} ;")" &&
-echo "backup types:          $remote_backup_types" || exit 1
-
-for backup_type in $remote_backup_types; do
-  if [ "$backup_type" != "$remote_machine_id" ]; then
-    echo "backup type:              $backup_type" &&
-    
-    general_backup_type_dir="$general_backup_machine_dir""$backup_type/" &&
-    general_versions_dir="$general_backup_type_dir" &&
-    local_previous_version_dir="$(ls -d $general_versions_dir* | tail -1)" &&
-    echo "last local backup:      $local_previous_version_dir" &&
-
-    remote_backup_versions="$(ssh "$remote_host" ls -d "$general_backup_type_dir"\*)" &&
-    echo "remote backup versions:   $remote_backup_versions" &&
-
-
-    remote_last_backup_dir=$(echo "$remote_backup_versions" | tail -1) &&
-    echo "last remote backup:       $remote_last_backup_dir" &&
-
-    remote_source_path="$remote_host:$remote_last_backup_dir/" &&
-    echo "source path:              $remote_source_path" &&
-
-    local_backup_destination_path=$remote_last_backup_dir &&
-    echo "backup destination:       $local_backup_destination_path" &&
-
-    echo "creating local backup destination folder..." &&
-    mkdir -vp "$local_backup_destination_path" &&
-
-    echo "starting backup..."
-    rsync_command='rsync -abP --delete --delete-excluded --rsync-path="sudo rsync" --link-dest="'$local_previous_version_dir'" "'$remote_source_path'" "'$local_backup_destination_path'"'
-
-    echo "executing:                $rsync_command"
-
-    retry_count=0
-    max_retries=12
-    retry_delay=300  # Retry delay in seconds (5 minutes)
-    last_retry_start=0
-    max_retry_duration=43200  # Maximum duration for a single retry attempt (12 hours)
-
-    while [[ $retry_count -lt $max_retries ]]; do
-      echo "Retry attempt: $((retry_count + 1))"
-      if [[ $retry_count -gt 0 ]]; then
-        current_time=$(date +%s)
-        last_retry_duration=$((current_time - last_retry_start))
-        if [[ $last_retry_duration -ge $max_retry_duration ]]; then
-          echo "Last retry took more than 12 hours, increasing max retries to 12."
-          max_retries=12
-        fi
-      fi
-      last_retry_start=$(date +%s)
-      eval "$rsync_command"
-      rsync_exit_code=$?
-      if [[ $rsync_exit_code -eq 0 ]]; then
-        break
-      fi
-      retry_count=$((retry_count + 1))
-      sleep $retry_delay
-    done
-
-    if [[ $rsync_exit_code -ne 0 ]]; then
-      echo "Error: rsync failed after $max_retries attempts"
-      ((errors += 1))
-    fi
-  fi
-done
-exit $errors;

roles/svc-bkp-rmt-2-loc/templates/script.sh.j2

@@ -3,6 +3,6 @@
 hosts="{{ DOCKER_BACKUP_REMOTE_2_LOCAL_BACKUP_PROVIDERS | join(' ') }}";
 errors=0
 for host in $hosts; do
-  bash {{ DOCKER_BACKUP_REMOTE_2_LOCAL_SCRIPT }} $host || ((errors+=1));
+  python {{ DOCKER_BACKUP_REMOTE_2_LOCAL_SCRIPT }} $host || ((errors+=1));
 done;
 exit $errors;

roles/svc-bkp-rmt-2-loc/vars/main.yml

@@ -4,6 +4,6 @@ system_service_id:                              "{{ application_id }}"
 
 # Role Specific
 DOCKER_BACKUP_REMOTE_2_LOCAL_DIR:               '{{ PATH_ADMINISTRATOR_SCRIPTS }}{{ application_id }}/'
-DOCKER_BACKUP_REMOTE_2_LOCAL_FILE:              'pull-specific-host.sh'
+DOCKER_BACKUP_REMOTE_2_LOCAL_FILE:              'pull-specific-host.py'
 DOCKER_BACKUP_REMOTE_2_LOCAL_SCRIPT:            "{{ [ DOCKER_BACKUP_REMOTE_2_LOCAL_DIR , DOCKER_BACKUP_REMOTE_2_LOCAL_FILE ] | path_join }}"
 DOCKER_BACKUP_REMOTE_2_LOCAL_BACKUP_PROVIDERS:  "{{ applications | get_app_conf(application_id, 'backup_providers')  }}"

roles/sys-bkp-provider-user/files/ssh-wrapper.sh

@@ -13,7 +13,7 @@ get_backup_types="find /Backups/$hashed_machine_id/ -maxdepth 1 -type d -execdir
 
 
 # @todo This configuration is not scalable yet. If other backup services then sys-ctl-bkp-docker-2-loc are integrated, this logic needs to be optimized
-get_version_directories="ls -d /Backups/$hashed_machine_id/sys-ctl-bkp-docker-2-loc/*"
+get_version_directories="ls -d /Backups/$hashed_machine_id/backup-docker-to-local/*"
 last_version_directory="$($get_version_directories | tail -1)"
 rsync_command="sudo rsync --server --sender -blogDtpre.iLsfxCIvu . $last_version_directory/"
 

roles/sys-bkp-provider-user/tasks/01_core.yml

@@ -3,30 +3,6 @@
     name: backup
     create_home: yes
 
-- name: create .ssh directory
-  file:
-    path: /home/backup/.ssh
-    state: directory
-    owner: backup
-    group: backup
-    mode: '0700'
-
-- name: create /home/backup/.ssh/authorized_keys
-  template:
-    src: "authorized_keys.j2"
-    dest: /home/backup/.ssh/authorized_keys
-    owner: backup
-    group: backup
-    mode: '0644'
-
-- name: create /home/backup/ssh-wrapper.sh
-  copy:
-    src: "ssh-wrapper.sh"
-    dest: /home/backup/ssh-wrapper.sh
-    owner: backup
-    group: backup
-    mode: '0700'
-
 - name: grant backup sudo rights
   copy:
     src: "backup"
@@ -35,3 +11,9 @@
     owner: root
     group: root
   notify: sshd restart
+
+- include_tasks: 02_permissions_ssh.yml
+
+- include_tasks: 03_permissions_folders.yml
+
+- include_tasks: utils/run_once.yml

roles/sys-bkp-provider-user/tasks/02_permissions_ssh.yml

@@ -0,0 +1,23 @@
+- name: create .ssh directory
+  file:
+    path: /home/backup/.ssh
+    state: directory
+    owner: backup
+    group: backup
+    mode: '0700'
+
+- name: create /home/backup/.ssh/authorized_keys
+  template:
+    src: "authorized_keys.j2"
+    dest: /home/backup/.ssh/authorized_keys
+    owner: backup
+    group: backup
+    mode: '0644'
+
+- name: create /home/backup/ssh-wrapper.sh
+  copy:
+    src: "ssh-wrapper.sh"
+    dest: /home/backup/ssh-wrapper.sh
+    owner: backup
+    group: backup
+    mode: '0700'

roles/sys-bkp-provider-user/tasks/03_permissions_folders.yml

@@ -0,0 +1,64 @@
+# Ensure the backups root exists and is owned by backup
+- name: Ensure backups root exists and owned by backup
+  file:
+    path: "{{ BACKUPS_FOLDER_PATH }}"
+    state: directory
+    owner: backup
+    group: backup
+    mode: "0700"
+
+# Explicit ACL so 'backup' has rwx, others none
+- name: Grant ACL rwx on backups root to backup user
+  ansible.posix.acl:
+    path: "{{ BACKUPS_FOLDER_PATH }}"
+    entity: backup
+    etype: user
+    permissions: rwx
+    state: present
+
+# Set default ACLs so new entries inherit rwx for backup and nothing for others
+- name: Set default ACL (inherit) for backup user under backups root
+  ansible.posix.acl:
+    path: "{{ BACKUPS_FOLDER_PATH }}"
+    entity: backup
+    etype: user
+    permissions: rwx
+    default: true
+    state: present
+
+# Remove default ACLs for group/others (defensive hardening)
+# Default ACLs so new entries inherit only backup's rwx
+- name: Default ACL for backup user (inherit)
+  ansible.posix.acl:
+    path: "{{ BACKUPS_FOLDER_PATH }}"
+    etype: user
+    entity: backup
+    permissions: rwx
+    default: true
+    state: present
+
+# Explicitly set default group/other to no permissions (instead of absent)
+- name: Default ACL for group -> none
+  ansible.posix.acl:
+    path: "{{ BACKUPS_FOLDER_PATH }}"
+    etype: group
+    permissions: '---'
+    default: true
+    state: present
+
+- name: Default ACL for other -> none
+  ansible.posix.acl:
+    path: "{{ BACKUPS_FOLDER_PATH }}"
+    etype: other
+    permissions: '---'
+    default: true
+    state: present
+
+- name: Fix ownership level 0..2 directories to backup:backup
+  ansible.builtin.shell: >
+    find "{{ BACKUPS_FOLDER_PATH }}" -mindepth 0 -maxdepth 2 -xdev -type d -exec chown backup:backup {} +
+
+- name: Fix perms level 0..2 directories to 0700
+  ansible.builtin.shell: >
+    find "{{ BACKUPS_FOLDER_PATH }}" -mindepth 0 -maxdepth 2 -xdev -type d -exec chmod 700 {} +
+

roles/sys-bkp-provider-user/tasks/main.yml

@@ -1,4 +1,2 @@
-- block:
-    - include_tasks: 01_core.yml
-    - include_tasks: utils/run_once.yml
+- include_tasks: 01_core.yml
   when: run_once_sys_bkp_provider_user is not defined