check-ostree.yaml

---
- hosts: ostree_guest
  become: no
  vars:
    fdo_credential: "false"
    embedded_container: "false"
    sysroot_ro: "false"
    total_counter: "0"
    failed_counter: "0"
    firewall_feature: "false"
    ignition: "false"
    test_custom_dirs_files: "false"

  tasks:
    # current target host's IP address
    - debug: var=ansible_all_ipv4_addresses
    - debug: var=ansible_facts['distribution_version']
    - debug: var=ansible_facts['distribution']
    - debug: var=ansible_facts['architecture']

    # check BIOS or UEFI
    - name: check bios or uefi
      stat:
        path: /sys/firmware/efi

    # check secure boot status if it's enabled
    - name: check secure boot status
      command: mokutil --sb-state
      ignore_errors: yes

    # check tpm device
    - name: check tpm device
      stat:
        path: /dev/tpm0
      ignore_errors: yes
      when: fdo_credential == "true"

    - name: check partition size
      command: df -h
      ignore_errors: yes
      become: yes

    - name: check disk partition table
      command: fdisk -l
      ignore_errors: yes
      become: yes

    - name: check rpm-ostree status
      command: rpm-ostree status
      ignore_errors: yes

    # case: check rt kernel installed after upgrade
    - name: check installed kernel
      command: uname -r
      register: result_kernel

    - name: check rt kernel installed
      block:
        - assert:
            that:
              - "'rt' in result_kernel.stdout"
            fail_msg: "rt kernel not installed, ostree upgrade might be failed"
            success_msg: "rt kernel installed in ostree upgrade"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: "'rt' in result_kernel.stdout"

    # first installed or upgraded
    - name: determine which stage the checking is running on
      shell: rpm-ostree status --json | jq '.deployments | length'
      register: result_stage

    - set_fact:
        checking_stage: "{{ result_stage.stdout }}"

    # case: check fdo onboarding status
    # after fdo onboarding finished, /boot/device-credentials will be moved to /etc/device-credentials
    - name: check if fdo onboarding completed successfully
      block:
        - name: wait until the file /etc/device-credentials is present before continuing
          wait_for:
            path: /etc/device-credentials
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: fdo_credential == "true"

    # case: check ostree commit correctly updated
    - name: get deployed ostree commit
      shell: rpm-ostree status --json | jq -r '.deployments[0].checksum'
      register: result_commit

    - name: make a json result
      set_fact:
        deploy_commit: "{{ result_commit.stdout }}"

    - name: check commit deployed and built
      block:
        - assert:
            that:
              - deploy_commit == ostree_commit
            fail_msg: "deployed ostree commit is not commit built by osbuild-composer"
            success_msg: "successful building and deployment"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check ostree ref
    - name: check ostree ref
      shell: rpm-ostree status --json | jq -r '.deployments[0].origin'
      register: result_ref

    - name: check ostree ref deployed
      block:
        - assert:
            that:
              - result_ref.stdout == ostree_ref
            fail_msg: "deployed ostree ref failed"
            success_msg: "ostree ref successful building and deployment"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case from bug: https://bugzilla.redhat.com/show_bug.cgi?id=1848453
    - name: check ostree-remount status
      command: systemctl is-active ostree-remount.service
      register: result_remount

    - name: ostree-remount should be started
      block:
        - assert:
            that:
              - result_remount.stdout == "active"
            fail_msg: "ostree-remount is not started by default"
            success_msg: "starting ostree-remount successful"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    - name: set mount point device name
      command: findmnt -r -o SOURCE -n /sysroot
      register: result_sysroot_source

    - set_fact:
        device_name: "{{ result_sysroot_source.stdout }}"

    # case: check pv format
    - name: check pv format
      shell: pvs --reportformat json | jq .report[0].pv[0].pv_fmt -r
      become: yes
      register: result_pv_fmt
      when: "'/dev/mapper/rootvg-rootlv' in result_sysroot_source.stdout"

    - name: "pv format should be lvm2"
      block:
        - assert:
            that:
              - result_pv_fmt.stdout == "lvm2"
            fail_msg: "pv format is not lvm2"
            success_msg: "pv format is lvm2"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: "'/dev/mapper/rootvg-rootlv' in result_sysroot_source.stdout"

    # case: check pv size
    - name: check pv size
      shell: pvs --reportformat json | jq .report[0].pv[0].pv_size -r
      become: yes
      register: result_pv_size
      when: "'/dev/mapper/rootvg-rootlv' in result_sysroot_source.stdout"

    # simplified installer uses coreos-installer to grow fs to 19G
    - name: "pv size should bigger than 19G"
      block:
        - assert:
            that:
              - "'19' in result_pv_size.stdout"
            fail_msg: "pv size is not bigger than 19G"
            success_msg: "pv size is bigger than 19G"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - "'/dev/mapper/rootvg-rootlv' in result_sysroot_source.stdout"
        - fdo_credential == "true"
        - ansible_facts['distribution'] == 'RedHat'

    # on fedora, it grows to 18.49G
    - name: "pv size should bigger than 18G"
      block:
        - assert:
            that:
              - "'18' in result_pv_size.stdout"
            fail_msg: "pv size is not bigger than 18G"
            success_msg: "pv size is bigger than 18G"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - "'/dev/mapper/rootvg-rootlv' in result_sysroot_source.stdout"
        - fdo_credential == "true"
        - ansible_facts['distribution'] == "Fedora"

    # raw image does not have coreos-installer to grow fs to 19G
    - name: "pv size should keep at 9G for raw image"
      block:
        - assert:
            that:
              - "'9' in result_pv_size.stdout"
            fail_msg: "pv size does not keep at 9G"
            success_msg: "pv size keeps at 9G"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - "'/dev/mapper/rootvg-rootlv' in result_sysroot_source.stdout"
        - fdo_credential == "false"

    # case: check /sysroot lv size
    - name: check sysroot lv size
      shell: df -h | grep sysroot
      register: result_sysroot_lv_size
      when: "'/dev/mapper/rootvg-rootlv' in result_sysroot_source.stdout"

    - name: "/sysroot lv size should be 9G"
      block:
        - assert:
            that:
              - "'9.0G' in result_sysroot_lv_size.stdout"
            fail_msg: "pv size is not 9G"
            success_msg: "pv size is 9G"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - "'/dev/mapper/rootvg-rootlv' in result_sysroot_source.stdout"
        - ansible_facts['distribution'] == 'RedHat'

    # lv size on fedora is 7.8G, /~https://github.com/osbuild/osbuild-composer/issues/3529
    - name: "/sysroot lv size should be 7.8G on fedora"
      block:
        - assert:
            that:
              - "'7.8G' in result_sysroot_lv_size.stdout"
            fail_msg: "lv size is not 7.8G"
            success_msg: "lv size is 7.8G"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - "'/dev/mapper/rootvg-rootlv' in result_sysroot_source.stdout"
        - ansible_facts['distribution'] == "Fedora"

    # case: check /sysroot mount status
    - name: check /sysroot mount status
      shell: findmnt -r -o OPTIONS -n /sysroot | awk -F "," '{print $1}'
      register: result_sysroot_mount_status

    - name: /sysroot should be mount with rw permission
      block:
        - assert:
            that:
              - result_sysroot_mount_status.stdout == "rw"
            fail_msg: "/sysroot is not mounted with rw permission"
            success_msg: "/sysroot is mounted with rw permission"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: sysroot_ro == "false"

    # https://fedoraproject.org/wiki/Changes/Silverblue_Kinoite_readonly_sysroot
    - name: /sysroot should be mount with ro permission on RHEL 9.2 and Fedora 37 above
      block:
        - assert:
            that:
              - result_sysroot_mount_status.stdout == "ro"
            fail_msg: "/sysroot is not mounted with ro permission"
            success_msg: "/sysroot is mounted with ro permission"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: sysroot_ro == "true"

    # case: check /var mount point
    - name: check /var mount point
      command: findmnt -r -o SOURCE -n /var
      register: result_var_mount_point

    - name: "/var should be mounted on {{ device_name }}[/ostree/deploy/{{ os_name }}/var]"
      block:
        - assert:
            that:
              - result_var_mount_point.stdout == "{{ device_name }}[/ostree/deploy/{{ os_name }}/var]"
            fail_msg: "/var does not mount on {{ device_name }}[/ostree/deploy/{{ os_name }}/var]"
            success_msg: "/var mounts on {{ device_name }}[/ostree/deploy/{{ os_name }}/var]"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check /var mount status
    - name: check /var mount status
      shell: findmnt -r -o OPTIONS -n /var | awk -F "," '{print $1}'
      register: result_var_mount_status

    - name: /var should be mount with rw permission
      block:
        - assert:
            that:
              - result_var_mount_status.stdout == "rw"
            fail_msg: "/var is not mounted with rw permission"
            success_msg: "/var is mounted with rw permission"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check /usr mount point
    - name: check /usr mount point
      command: findmnt -r -o SOURCE -n /usr
      register: result_usr_mount_point

    - name: "/usr should be mounted on {{ device_name }}[/ostree/deploy/{{ os_name }}/deploy/{{ deploy_commit }}.0/usr]"
      block:
        - assert:
            that:
              - result_usr_mount_point.stdout == "{{ device_name }}[/ostree/deploy/{{ os_name }}/deploy/{{ deploy_commit }}.0/usr]"
            fail_msg: "/usr does not mount on {{ device_name }}[/ostree/deploy/{{ os_name }}/deploy/{{ deploy_commit }}.0/usr]"
            success_msg: "/usr mounts on {{ device_name }}[/ostree/deploy/{{ os_name }}/deploy/{{ deploy_commit }}.0/usr]"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check /usr mount status
    - name: check /usr mount status
      shell: findmnt -r -o OPTIONS -n /usr | awk -F "," '{print $1}'
      register: result_usr_mount_status

    - name: /usr should be mount with rw permission
      block:
        - assert:
            that:
              - result_usr_mount_status.stdout == "ro"
            fail_msg: "/usr is not mounted with ro permission"
            success_msg: "/usr is mounted with ro permission"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    - name: get the first 10 chars in commit hash
      set_fact:
        commit_log: "{{ deploy_commit[:11] }}"

    # Ignition test only
    # Ignition test running by user core, otherwise by user admin
    - name: check hello.service added by ignition
      command: systemctl list-unit-files hello.service
      register: result_hello_service
      changed_when: no
      ignore_errors: yes
      when: ignition == "true"

    - name: hello.service should be added by ignition
      block:
        - assert:
            that:
              - result_hello_service.rc == 0
            fail_msg: "hello.service does not exist"
            success_msg: "hello.service has been added by ignition"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: ignition == "true"

    # Ignition test only
    # Ignition test running by user core, otherwise by user admin
    - name: check log_trace.conf for fdo-client-linuxapp.service updated by ignition
      command: grep -Fxq "Environment=LOG_LEVEL=trace" /etc/systemd/system/fdo-client-linuxapp.service.d/log_trace.conf
      register: result_log_trace
      changed_when: no
      become: yes
      ignore_errors: yes
      when: ignition == "true"

    - name: Environment=LOG_LEVEL=trace should be added by ignition
      block:
        - assert:
            that:
              - result_log_trace.rc == 0
            fail_msg: "Config Environment=LOG_LEVEL=trace does not exist"
            success_msg: "Environment=LOG_LEVEL=trace has been added by ignition"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: ignition == "true"

    # Ignition test only
    # hello.service just run a shell script for echo "Hello, World!"
    - name: check systemd service correctly started on firstboot
      block:

        - name: check hello.service logs
          command: journalctl -b -0 -u hello.service
          register: result_hello_service_log
          become: yes

        - assert:
            that:
              - "'Hello, World!' in result_hello_service_log.stdout"
            fail_msg: "hello.service doesn't have the correct log"
            success_msg: "hello.service started and working"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: ignition == "true"

    # case: check wget installed after upgrade
    - name: check installed package
      shell: rpm -qa | sort
      register: result_packages

    - name: check wget installed
      block:
        - assert:
            that:
              - "'wget' in result_packages.stdout"
            fail_msg: "wget not installed, ostree upgrade might be failed"
            success_msg: "wget installed in ostree upgrade"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: checking_stage == "2"

    # case: check dmesg error and failed log
    - name: check dmesg output
      command: dmesg
      become: yes

    # Ignore ignition first attempt error log, it'll be successful at the second attempt
    # e.g. ignition[630]: GET error: Get "http://192.168.100.1/ignition/config.ign": dial tcp 192.168.100.1:80: connect: network is unreachable
    # Workaround for bug https://issues.redhat.com/browse/RHEL-43587, ignore NetworkManager[614] NetworkManager[625]
    - name: check dmesg error and fail log
      shell: dmesg --notime | grep -i "error\|fail" | grep -v "skipped" | grep -v "failover" | grep -v "ignition" | grep -v "BAR 13":" failed to assign"  | grep -v "failed to assign" | grep -v "NetworkManager" || true
      register: result_dmesg_error
      become: yes

    - name: no more error or failed log
      block:
        - assert:
            that:
              # Work around for issue /~https://github.com/virt-s1/rhel-edge/issues/1402
              - result_dmesg_error.stdout_lines | length <= 5
              - "'pcieport 0000:00:01.6: Failed to check link status' in result_dmesg_error.stdout or 'Error: Driver \\'pcspkr\\' is already registered, aborting' in result_dmesg_error.stdout or 'KD_FONT_OP_GET failed while trying to get the font metadata: Invalid argument' in result_dmesg_error.stdout or 'sys-module-fuse.device: Failed to enqueue SYSTEMD_WANTS= job, ignoring: Unit sys-fs-fuse-connections.mount not found' in result_dmesg_error.stdout or '/usr/lib/systemd/system-generators/podman-system-generator failed with exit status 1' in result_dmesg_error.stdout or result_dmesg_error.stdout == ''"
            fail_msg: "more or less error and failed log"
            success_msg: "everything works as expected"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - ansible_facts['distribution'] == "RedHat" or ansible_facts['distribution'] == 'CentOS'
        - "'rt' not in result_kernel.stdout"
        - ansible_facts['architecture'] == "x86_64"

    - name: no more error or failed log on aarch64
      block:
        - assert:
            that:
              # Fedora has 5 failed logs:
              # NUMA: Failed to initialise from firmware
              # 2 * systemd[1]: bpf-lsm: Failed to link program; assuming BPF LSM is not available
              # systemd-gpt-auto-generator[581]: Failed to dissect: Permission denied
              # systemd[572]: /usr/lib/systemd/system-generators/systemd-gpt-auto-generator failed with exit status 1.
              - result_dmesg_error.stdout_lines | length <= 10
              - "'NUMA: Failed to initialise from firmware' in result_dmesg_error.stdout or 'systemd[1]: bpf-lsm: Failed to link program; assuming BPF LSM is not available' in result_dmesg_error.stdout or result_dmesg_error.stdout == ''"
            fail_msg: "more or less error and failed log"
            success_msg: "everything works as expected"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - ansible_facts['architecture'] == "aarch64"

    - name: no any error or fail log
      block:
        - assert:
            that:
              # Work around for issue /~https://github.com/virt-s1/rhel-edge/issues/1402
              # Work around for issue /~https://github.com/virt-s1/rhel-edge/issues/1893
              # Word around for issue /~https://github.com/virt-s1/rhel-edge/issues/1943
              - result_dmesg_error.stdout_lines | length <= 5
            fail_msg: "no any error fail log or only have one log"
            success_msg: "everything works as expected"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - "'rt' in result_kernel.stdout"
        - ansible_facts['architecture'] == "x86_64"

    - name: no more error or failed log
      block:
        - assert:
            that:
              - result_dmesg_error.stdout_lines | length <= 5
              # - "'pcieport 0000:00:01.6: pciehp: Failed to check link status' in result_dmesg_error.stdout"
              - "'RAS: Correctable Errors collector initialized' in result_dmesg_error.stdout or 'systemd[1]: bpf-lsm: Failed to load BPF object: No such process' in result_dmesg_error.stdout or 'Failed to dissect: Permission denied' in result_dmesg_error.stdout or '/usr/lib/systemd/system-generators/systemd-gpt-auto-generator failed with exit status 1' in result_dmesg_error.stdout"
            fail_msg: "more or less error and failed log"
            success_msg: "everything works as expected"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - ansible_facts['distribution'] == "Fedora"
        - "'rt' not in result_kernel.stdout"
        - ansible_facts['architecture'] == "x86_64"

    - name: check embedded container image with podman
      command: podman images
      become: yes
      register: result_podman_images
      when:
        - embedded_container == "true"

    - name: embedded container should be listed by podman images
      block:
        - assert:
            that:
              - "'quay.io/fedora/fedora' in result_podman_images.stdout"
            fail_msg: "fedora image is not built in image"
            success_msg: "fedora image is built in image"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - embedded_container == "true"

    - name: embedded fedora-minimal container image should be listed by podman images
      block:
        - assert:
            that:
              - "'localhost/fedora-minimal' in result_podman_images.stdout"
            fail_msg: "fedora-minimal image is not built in image"
            success_msg: "fedora-minimal image is built in image"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - embedded_container == "true"
        - ansible_architecture == "x86_64"
        - (ansible_facts['distribution'] == 'RedHat' and ansible_facts ['distribution_version'] is version('9.2', '>='))

    - name: embedded fedora-aarch64 container image should be listed by podman images
      block:
        - assert:
            that:
              - "'localhost/fedora-aarch64' in result_podman_images.stdout"
            fail_msg: "fedora-aarch64 image is not built in image"
            success_msg: "fedora-aarch64 image is built in image"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - embedded_container == "true"
        - ansible_architecture == "aarch64"
        - (ansible_facts['distribution'] == 'RedHat' and ansible_facts ['distribution_version'] is version('9.2', '>='))

    - name: run ubi8 image with root
      command: podman run ubi8-minimal:latest cat /etc/redhat-release
      register: podman_result
      become: yes
      retries: 30
      delay: 2
      until: podman_result is success
      ignore_errors: yes

    - name: run container test
      block:
        - assert:
            that:
              - podman_result is succeeded
              - "'Red Hat Enterprise Linux release' in podman_result.stdout"
            fail_msg: "failed run container with podman (root)"
            success_msg: "running container with podman (root) succeeded"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check running container with podman (non-root)
    # workaround on RHEL 8.6 and 9.0 for issue https://bugzilla.redhat.com/show_bug.cgi?id=2078937
    # because the fix is not available on 8.6 and 9.0 yet
    # bug https://bugzilla.redhat.com/show_bug.cgi?id=2123611 fix is not landed on CS8
    - name: run ubi8 image with non-root
      command: podman run ubi8:latest cat /etc/redhat-release
      register: podman_result
      retries: 30  # due to /~https://github.com/osbuild/osbuild-composer/issues/2492
      delay: 2
      until: podman_result is success
      ignore_errors: yes

    # bug https://bugzilla.redhat.com/show_bug.cgi?id=2123611 fix is not landed on CS8
    - name: run container test
      block:
        - assert:
            that:
              - podman_result is succeeded
              - "'Red Hat Enterprise Linux release' in podman_result.stdout"
            fail_msg: "failed run container with podman (non-root)"
            success_msg: "running container with podman (non-root) succeeded"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check fedora-minimal container with podman
    - name: run fedora-minimal image
      command: podman run localhost/fedora-minimal@sha256:4d76a7480ce1861c95975945633dc9d03807ffb45c64b664ef22e673798d414b cat /etc/os-release
      register: fminimal_container_result
      become: yes
      retries: 30  # due to /~https://github.com/osbuild/osbuild-composer/issues/2492
      delay: 2
      until: fminimal_container_result is success
      ignore_errors: yes  # due to https://bugzilla.redhat.com/show_bug.cgi?id=1903983
      when:
        - embedded_container == "true"
        - ansible_architecture == "x86_64"
        - (ansible_facts['distribution'] == 'RedHat' and ansible_facts ['distribution_version'] is version('9.2', '>='))

    - name: run fedora-minimal container test
      block:
        - assert:
            that:
              - fminimal_container_result is succeeded
              - "'Fedora Linux 36 (Container Image)' in fminimal_container_result.stdout"
              - "'Trying to pull' not in fminimal_container_result.stdout"
            fail_msg: "failed run fedora-minimal container with podman"
            success_msg: "running fedora-minimal container with podman succeeded"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - embedded_container == "true"
        - ansible_architecture == "x86_64"
        - (ansible_facts['distribution'] == 'RedHat' and ansible_facts ['distribution_version'] is version('9.2', '>='))

    # case: check fedora-aarch64 container with podman
    - name: run fedora-aarch64 image
      command: podman run localhost/fedora-aarch64@sha256:8fd6ac4c552bbec7910df7b0625310561d56513ecbcc418825a2f5635efecfab cat /etc/os-release
      register: faarch64_container_result
      become: yes
      retries: 30  # due to /~https://github.com/osbuild/osbuild-composer/issues/2492
      delay: 2
      until: faarch64_container_result is success
      ignore_errors: yes  # due to https://bugzilla.redhat.com/show_bug.cgi?id=1903983
      when:
        - embedded_container == "true"
        - ansible_architecture == "aarch64"
        - (ansible_facts['distribution'] == 'RedHat' and ansible_facts ['distribution_version'] is version('9.2', '>='))

    - name: run fedora-aarch64 container test
      block:
        - assert:
            that:
              - faarch64_container_result is succeeded
              - "'Trying to pull' not in faarch64_container_result.stdout"
            fail_msg: "failed run fedora-aarch64 container with podman"
            success_msg: "running fedora-aarch64 container with podman succeeded"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - embedded_container == "true"
        - ansible_architecture == "aarch64"
        - (ansible_facts['distribution'] == 'RedHat' and ansible_facts ['distribution_version'] is version('9.2', '>='))


    # case: check dnf package and it should not be installed
    # /~https://github.com/osbuild/osbuild-composer/blob/master/internal/distro/rhel8/distro.go#L642
    - name: dnf should not be installed
      block:
        - name: dnf should not be installed
          shell: rpm -qa | grep dnf || echo -n PASS
          register: result_dnf

        - assert:
            that:
              - result_dnf.stdout == "PASS"
            fail_msg: "dnf is installed"
            success_msg: "No dnf installed"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: ansible_facts['distribution'] == 'RedHat' and ansible_facts ['distribution_version'] is version('9.0', '<')

    # case: check installed greenboot packages
    - name: greenboot should be installed
      block:
        - name: greenboot should be installed
          shell: rpm -qa | grep greenboot
          register: result_greenboot_packages

        - assert:
            that:
              - "'greenboot-0' in result_greenboot_packages.stdout"
              - "'greenboot-default-health-checks' in result_greenboot_packages.stdout"
            fail_msg: "greenboot is not installed"
            success_msg: "greenboot is installed"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check greenboot* services
    - name: a list of greenboot* service should be enabled
      block:
        - name: a list of greenboot* service should be enabled
          command: systemctl is-enabled greenboot-grub2-set-counter greenboot-grub2-set-success greenboot-healthcheck greenboot-rpm-ostree-grub2-check-fallback greenboot-status greenboot-task-runner redboot-auto-reboot redboot-task-runner
          register: result_greenboot_service

        - assert:
            that:
              - result_greenboot_service.stdout == 'enabled\nenabled\nenabled\nenabled\nenabled\nenabled\nenabled\nenabled'
            fail_msg: "Some of greenboot* services are not enabled"
            success_msg: "All greenboot* services are enabled"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check greenboot* services log
    - name: all greenboot* service should run without error
      block:
        # check service/target status instead of string matching messages
        - name: check boot-complete.target
          # will fail if the target was not reached
          command: systemctl --no-pager status boot-complete.target
          register: result
          retries: 10
          delay: 60
          until: "'inactive' not in result.stdout"

        - name: all greenboot* service should run without error
          command: journalctl -b -0 -u greenboot -u greenboot-healthcheck -u greenboot-rpm-ostree-grub2-check-fallback -u greenboot-grub2-set-counter -u greenboot-grub2-set-success -u greenboot-status -u redboot -u redboot-auto-reboot -u redboot.target
          become: yes
          register: result_greenboot_log

        - assert:
            that:
              - "'Script \\'00_required_scripts_start.sh\\' SUCCESS' in result_greenboot_log.stdout"
              - "'Script \\'00_wanted_scripts_start.sh\\' SUCCESS' in result_greenboot_log.stdout"
              - "'greenboot Health Checks Runner' in result_greenboot_log.stdout"
              - "'Mark boot as successful in grubenv' in result_greenboot_log.stdout"
              - "'Boot Status is GREEN - Health Check SUCCESS' in result_greenboot_log.stdout"
              - "'greenboot MotD Generator' in result_greenboot_log.stdout"
            fail_msg: "Some errors happened in service boot"
            success_msg: "All greenboot services booted success"

      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check grubenv variables
    - name: grubenv variables should contain boot_success=1
      block:
        - name: grubenv variables should contain boot_success=1
          command: grub2-editenv list
          register: result_grubenv
          become: yes

        - assert:
            that:
              - "'boot_success=1' in result_grubenv.stdout"
            fail_msg: "Not found boot_success=1"
            success_msg: "Found boot_success=1"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # Check FDO status and task status
    - name: check fdo-client-linuxapp logs
      command: journalctl -u fdo-client-linuxapp
      register: result_fdo_client_linuxapp_journalctl
      become: yes
      when:
        - fdo_credential == "true"

    # Check FDO client avc log
    - name: check FDO client avc logs
      command: ausearch -m avc -m user_avc -m selinux_err -i
      ignore_errors: yes
      become: yes
      when:
        - fdo_credential == "true"

    # case: check rollback function if boot error found
    - name: install sanely failing health check unit to test red boot status behavior
      block:
        - name: install sanely failing health check unit to test red boot status behavior
          command: rpm-ostree install --cache-only https://kite-webhook-prod.s3.amazonaws.com/greenboot-failing-unit-1.0-1.el8.noarch.rpm --reboot
          become: yes
          ignore_errors: yes
          ignore_unreachable: yes

        - name: delay 60 seconds before reboot to make system stable
          pause:
            seconds: 60
          delegate_to: 127.0.0.1

        - name: wait for connection to become reachable/usable
          wait_for_connection:
            delay: 30

        - name: waits until instance is reachable
          wait_for:
            host: "{{ ansible_all_ipv4_addresses[0] }}"
            port: 22
            search_regex: OpenSSH
            delay: 10
          register: result_rollback
          until: result_rollback is success
          retries: 6
          delay: 10

        - assert:
            that:
              - result_rollback is succeeded
            fail_msg: "Rollback failed"
            success_msg: "Rollback success"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"

    # case: check persistent log in Edge system
    - name: check journald has persistent logging
      block:
        - name: list boots
          shell: journalctl --list-boots
          register: result_list_boots
          become: yes

        - assert:
            that:
              - result_list_boots.stdout_lines | length > 1
            fail_msg: "journald hasn't persistent logging"
            success_msg: "journald has persistent logging"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - result_rollback is succeeded
        # Disable rhel9.5 check due to bug https://issues.redhat.com/projects/RHEL/issues/RHEL-45151
        - ((ansible_facts['distribution'] == 'RedHat' and ansible_facts ['distribution_version'] is version('9.2', '>=')) and ansible_facts ['distribution_version'] is version('9.5', '!=')) or
          ((ansible_facts['distribution'] == 'RedHat' and ansible_facts ['distribution_version'] is version('8.8', '>=')) and ansible_facts ['distribution_version'] is version('9.0', '!=') and ansible_facts ['distribution_version'] is version('9.1', '!=')) or
          (ansible_facts['distribution'] == 'CentOS') or
          (ansible_facts['distribution'] == 'Fedora')

    # case: check ostree commit after rollback
    - name: check ostree commit after rollback
      block:
        - name: check ostree commit after rollback
          shell: rpm-ostree status --json | jq -r '.deployments[0].checksum'
          register: result_commit

        - assert:
            that:
              - deploy_commit == ostree_commit
            fail_msg: "Not rollbackto last commit"
            success_msg: "Rollback success"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: result_rollback is succeeded

    # case: check greenboot* services log again
    - name: fallback log should be found here
      block:
        # check service/target status instead of string matching messages
        - name: check boot-complete.target
          # will fail if the target was not reached
          command: systemctl --no-pager status boot-complete.target
          become: yes
          register: result
          retries: 10
          delay: 60
          until: "'inactive' not in result.stdout"

        - name: fallback log should be found here
          command: journalctl -b -0 -u greenboot -u greenboot-healthcheck -u greenboot-rpm-ostree-grub2-check-fallback -u greenboot-grub2-set-counter -u greenboot-grub2-set-success -u greenboot-status -u redboot -u redboot-auto-reboot -u redboot.target
          become: yes
          register: result_greenboot_log

        - assert:
            that:
              - "'FALLBACK BOOT DETECTED! Default rpm-ostree deployment has been rolled back' in result_greenboot_log.stdout"
              - "'Script \\'00_required_scripts_start.sh\\' SUCCESS' in result_greenboot_log.stdout"
              - "'Script \\'00_wanted_scripts_start.sh\\' SUCCESS' in result_greenboot_log.stdout"
              - "'greenboot Health Checks Runner' in result_greenboot_log.stdout"
              - "'Mark boot as successful in grubenv' in result_greenboot_log.stdout"
              - "'Boot Status is GREEN - Health Check SUCCESS' in result_greenboot_log.stdout"
              - "'greenboot MotD Generator' in result_greenboot_log.stdout"
            fail_msg: "Fallback log not found"
            success_msg: "Found fallback log"

      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: result_rollback is succeeded

    # case: check grubenv variables again
    - name: grubenv variables should contain boot_success=1
      block:
        - name: grubenv variables should contain boot_success=1
          command: grub2-editenv list
          register: result_grubenv
          become: yes

        - assert:
            that:
              - "'boot_success=1' in result_grubenv.stdout"
            fail_msg: "Not found boot_success=1"
            success_msg: "Found boot_success=1"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: result_rollback is succeeded

    # Check FDO status and task status
    - name: check fdo-client-linuxapp logs
      command: journalctl -u fdo-client-linuxapp
      register: result_fdo_client_linuxapp_journalctl
      become: yes
      when:
        - fdo_credential == "true"

    # Check FDO client avc log
    - name: check FDO client avc logs
      command: ausearch -m avc -m user_avc -m selinux_err -i
      ignore_errors: yes
      become: yes
      when:
        - fdo_credential == "true"

    # Check re-encryption status on x86_64
    - name: wait for FDO re-encryption
      block:
        - shell: cryptsetup luksDump /dev/vda4
          register: result
          until: not result.stdout_lines is search("cipher_null-ecb")
          retries: 30
          delay: 60
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - fdo_credential == "true"
        - ansible_facts['architecture'] == "x86_64"
        - ansible_facts['distribution'] == 'RedHat'

    # Check re-encryption status on aarch64
    - name: wait for FDO re-encryption
      block:
        - shell: cryptsetup luksDump /dev/vda3
          register: result
          become: yes
          until: not result.stdout_lines is search("cipher_null-ecb")
          retries: 30
          delay: 60
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - fdo_credential == "true"
        - ansible_facts['architecture'] == "aarch64"

    # Check re-encryption status on fedora
    - name: wait for FDO re-encryption
      block:
        - shell: cryptsetup luksDump /dev/vda3
          register: result
          until: not result.stdout_lines is search("cipher_null-ecb")
          retries: 30
          delay: 60
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - fdo_credential == "true"
        - ansible_facts['distribution'] == "Fedora"

    # Check FDO status and task status
    - name: check fdo-client-linuxapp logs
      command: journalctl -u fdo-client-linuxapp
      register: result_fdo_client_linuxapp_journalctl
      become: yes
      when:
        - fdo_credential == "true"

    # Check FDO client avc log
    - name: check FDO client avc logs
      command: ausearch -m avc -m user_avc -m selinux_err -i
      ignore_errors: yes
      become: yes
      when:
        - fdo_credential == "true"

    # case: checking firewall customizations
    - name: Check applied firewall customizations
      block:
        - name: Ensure firewall customizations applied from blueprint in trusted zone
          command: firewall-cmd --info-zone=trusted
          register: result_trusted_zone
          become: yes
        - name: Ensure firewall customizations applied from blueprint in work zone
          command: firewall-cmd --info-zone=work
          register: result_work_zone
          become: yes

        - assert:
            that:
              - "'192.168.100.51' in result_trusted_zone.stdout"
              - "'192.168.100.52' in result_work_zone.stdout"
            fail_msg: "No firewall customizations found"
            success_msg: "Firewall customizations added from blueprint"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: firewall_feature == "true"

    # case: check sos package
    - name: Check SOS Package
      block:
        - name: Generate sos report
          command: sos report --no-report --batch -o podman
          become: "true"
          register: result_sos_report

        - assert:
            that:
              - result_sos_report is succeeded
            fail_msg: "No usable sos package present"
            success_msg: "sos package present and usable"
      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when:
        - ansible_facts['distribution'] != 'Fedora'

# case: checking files and directories customizations
    - name: Check that custom files and directories are present
      block:
        # Test basic custom files and directories creation
        - name: Check that /etc/custom_dir exists
          stat:
            path: /etc/custom_dir

        - name: Check that /etc/custom_dir/dir1 exists
          stat:
            path: /etc/custom_dir/dir1
          register: result_custom_dir

        - name: Check that /etc/custom_dir/dir1 is owned by user ID `1020`
          assert:
            that:
              - result_custom_dir.stat.uid == 1020
            fail_msg: "Directory /etc/custom_dir/dir1 is not owned by user ID '1020'"
            success_msg: "Directory /etc/custom_dir/dir1 is owned by user ID '1020'"

        - name: Check that /etc/custom_dir/dir1 is owned by group ID `1020`
          assert:
            that:
              - result_custom_dir.stat.gid == 1020
            fail_msg: "Directory /etc/custom_dir/dir1 is not owned by group ID '1020'"
            success_msg: "Directory /etc/custom_dir/dir1 is owned by group ID '1020'"

        - name: Check that /etc/custom_dir/dir1 has 0770 permissions
          assert:
            that:
              - result_custom_dir.stat.mode == '0770'
            fail_msg: "Directory /etc/custom_dir/dir1 has wrong permissions"
            success_msg: "Directory /etc/custom_dir/dir1 has correct permissions"

        # Test the use of custom files for systemd units
        - name: Check status of 'custom.service'
          systemd:
            name: custom.service
          register: result_custom_service

        - name: Check that 'custom.service' is started and enabled
          assert:
            that:
              - result_custom_service.status['LoadState'] == 'loaded'
              - result_custom_service.status['ActiveState'] == 'active'
              - result_custom_service.status['SubState'] == 'exited'
              - result_custom_service.status['UnitFileState'] == 'enabled'
            fail_msg: "Service 'custom.service' is not started or enabled"
            success_msg: "Service 'custom.service' is started and enabled"

        # The DropInPaths on Fedora 38+ is different with Fedora 37 and other OS
        - name: Check that 'custom.service' was overridden by drop-in for Fedora 38+
          assert:
            that:
              - result_custom_service.status['DropInPaths'] == "/usr/lib/systemd/system/service.d/10-timeout-abort.conf /etc/systemd/system/custom.service.d/override.conf"
            fail_msg: "Service 'custom.service' was not overridden by drop-in"
            success_msg: "Service 'custom.service' was overridden by drop-in"
          when:
            - (ansible_facts['distribution'] == 'Fedora') and (ansible_facts ['distribution_version'] is version('38', '>='))

        - name: Check that 'custom.service' was overridden by drop-in for OS rather than Fedora 38+
          assert:
            that:
              - result_custom_service.status['DropInPaths'] == "/etc/systemd/system/custom.service.d/override.conf"
            fail_msg: "Service 'custom.service' was not overridden by drop-in"
            success_msg: "Service 'custom.service' was overridden by drop-in"
          when:
            - not (ansible_facts['distribution'] == 'Fedora' and ansible_facts ['distribution_version'] is version('38', '>='))

        - name: Check output of 'custom.service' in the journal
          command: journalctl -b -u custom.service
          register: custom_service_journal
          become: yes

        - name: Check that 'image builder is the best' message is present in journal
          assert:
            that:
              - "'image builder is the best' in custom_service_journal.stdout"
            fail_msg: "Message 'image builder is the best' is not present in journal"
            success_msg: "Message 'image builder is the best' is present in journal"

      always:
        - set_fact:
            total_counter: "{{ total_counter | int + 1 }}"
      rescue:
        - name: failed count + 1
          set_fact:
            failed_counter: "{{ failed_counter | int + 1 }}"
      when: test_custom_dirs_files == "true"

    - assert:
        that:
          - failed_counter == "0"
        fail_msg: "Run {{ total_counter }} tests, but {{ failed_counter }} of them failed"
        success_msg: "Totally {{ total_counter }} test passed"