ansible-role-collect-logs/roles/collect_logs/tasks/collect.yml

---
- become: true
  ignore_errors: true
  block:
    - name: Ensure required rpms for logging are installed
      package:
        state: present
        name: "{{ artcl_collect_pkg_list }}"

    - name: Prepare directory with extra logs
      file:
        dest: /var/log/extra
        state: directory
        mode: 0755

    - name: Create rsync filter file
      template:
        src: "odl_extra_logs.j2"
        dest: "/tmp/odl_extra_logs.sh"
        mode: 0644

    - name: Determine commands to run
      run_once: true
      vars:
        combined_cmds: "{{ artcl_commands | combine(artcl_commands_extras, recursive=True) }}"
      # combines default dictionary with user defined one
      # keeps only commands from groups mentioned in collect_log_types
      flatten_nested_dict:
        data: "{{ combined_cmds | dict2items|selectattr('key', 'in', collect_log_types) | list | items2dict }}"
      register: artcl_commands_flatten

    - name: install setools
      ansible.builtin.package:
        name:
          - setools
          - setroubleshoot
        state: present

    - name: install custom consolidation script
      ansible.builtin.copy:
        dest: /usr/local/bin/consolidate-avc.pl
        src: consolidate-avc.pl
        mode: 0555

    - name: Run artcl_commands
      # noqa 305
      # noqa 102 :: No Jinja2 in when
      vars:
        capture_file: "{{ item.capture_file | default( item.name + '.txt') }}"
      shell:
        # redirection of output to log file, see https://ops.tips/gists/redirect-all-outputs-of-a-bash-script-to-a-file/
        cmd: |
          {% if not item.capture_disable | default(False) %}
          exec >{% if not capture_file.startswith('/') %}/var/log/extra/{% endif %}{{ capture_file }} 2>&1
          {% endif %}
          {# do not put anything after the command #}
          {{ item.cmd }}          
        warn: false
      args:
        chdir: /var/log/extra
        executable: /bin/bash
      changed_when: false
      when: item.when | default(true)
      loop: "{{ artcl_commands_flatten.data }}"
      loop_control:
        label: "{{ item.name }}"

# Change the collect_log_types if you don't want to collect
# some specific logs
- import_tasks: collect/container.yml
  when: "'container' in collect_log_types"

- import_tasks: collect/system.yml
  when: "'system' in collect_log_types"

- import_tasks: collect/network.yml
  when: "'network' in collect_log_types"

- import_tasks: collect/monitoring.yml
  when: "'monitoring' in collect_log_types"

- name: Set default collect list
  set_fact:
    collect_list: "{{ artcl_collect_list }} + {{ artcl_collect_list_append|default([]) }}"

- name: Override collect list
  set_fact:
    collect_list: "{{ artcl_collect_override[inventory_hostname] }}"
  when:
    - artcl_collect_override is defined
    - artcl_collect_override[inventory_hostname] is defined

- name: Set default exclude list
  set_fact:
    artcl_exclude_list: "{{ artcl_exclude_list|default([]) }} + {{ artcl_exclude_list_append|default([]) }}"

- name: Create temp directory before gathering logs
  file:
    dest: "/tmp/{{ inventory_hostname }}"
    state: directory
    mode: 0755

- name: Create rsync filter file
  template:
    src: "rsync-filter.j2"
    dest: "/tmp/{{ inventory_hostname }}-rsync-filter"
    mode: 0644
  when: artcl_rsync_collect_list|bool

# This task needs to be finished before generating find list of files
# to collect (Create find list file task) otherwise not all the container
# log files may be found and thus not collected later
- name: Wait for container logs collection if not finished yet
  become: true
  async_status:
    jid: "{{ container_collection.ansible_job_id }}"
  register: container_collection_result
  until: container_collection_result.finished
  delay: 10
  retries: "{{ ((artcl_container_collect_timeout|int) / 10)|int }}"
  when: "'container' in collect_log_types"

- name: Create find list file
  become: true
  shell: >
    find {{ collect_list|join(' ') }}
    -maxdepth {{ artcl_find_maxdepth }}
    -type f \
    -size -{{ artcl_find_max_size }}M
    {% if artcl_exclude_list is defined %}
    -not -path {{ artcl_exclude_list|map('quote')|join(' -not -path ') }}
    {% endif %}
    -print0 > /tmp/{{ inventory_hostname }}-rsync-list    
  failed_when: false
  when: not artcl_rsync_collect_list|bool

- name: Gather the logs to /tmp
  become: true
  shell: >
    set -o pipefail &&
    rsync --quiet --recursive --copy-links --prune-empty-dirs --ignore-errors
    {% if artcl_rsync_collect_list|bool %}
    --filter '. /tmp/{{ inventory_hostname }}-rsync-filter'
    {% else %}
    --from0 --files-from=/tmp/{{ inventory_hostname }}-rsync-list
    {% endif %}
    / /tmp/{{ inventory_hostname }};
    find /tmp/{{ inventory_hostname }} -type d -print0 | xargs -0 chmod 755;
    find /tmp/{{ inventory_hostname }} -type f -print0 | xargs -0 chmod 644;
    find /tmp/{{ inventory_hostname }} -not -type f -not -type d -delete;
    {# chown can fail with: chown: invalid spec: '0:' #}
    chown -R {{ ansible_user | default(ansible_effective_user_id) }}: /tmp/{{ inventory_hostname }} || true;    
  args:
    executable: /bin/bash
  changed_when: true

# See README section 'Sanitizing Log Strings'
- name: Sanitize logs to remove sensitive details
  include_tasks: sanitize_log_strings.yaml
  loop: "{{ sanitize_lines }}"
  loop_control:
    loop_var: outer_item
  when: sanitize_lines is defined and sanitize_lines|length

# it makes sense to compress the logs prior
# to sending them over the wire to the
# node where they are collected by infra.
# Regardless of the file size.
- name: Compress the collected files if configured
  when: artcl_gzip | bool
  shell: gzip -r ./{{ inventory_hostname }}
  args:
    chdir: /tmp
    warn: false
  changed_when: true
  tags:
    - skip_ansible_lint

- name: Create tar archive of logs for faster copying  # noqa: command-instead-of-module
  shell:
    cmd: tar cf {{ inventory_hostname }}.tar {{ inventory_hostname }};
    chdir: /tmp
  changed_when: true

- name: Fetch log archive (tar)
  fetch:
    src: "/tmp/{{ inventory_hostname }}.tar"
    dest: "{{ artcl_collect_dir }}/{{ inventory_hostname }}.tar"
    flat: true
    validate_checksum: false

- name: Delete temporary log directory after collection
  file:
    path: "/tmp/{{ inventory_hostname }}"
    state: absent
  ignore_errors: true  # noqa ignore-errors

- name: Extract the logs archive
  unarchive:
    src: "{{ artcl_collect_dir }}/{{ inventory_hostname }}.tar"
    dest: "{{ artcl_collect_dir }}"
    remote_src: true
  delegate_to: localhost

- name: Remove logs archive
  file:
    path: "{{ artcl_collect_dir }}/{{ inventory_hostname }}.tar"
    state: absent
  delegate_to: localhost