Add a playbook and role to set the boot mode (bios, uefi) of Dell servers

ansible/dell-compute-node-boot-mode.yml

@@ -0,0 +1,44 @@
+---
+# Set the boot mode (BIOS, UEFI) of Dell compute nodes.
+
+- name: Ensure compute nodes are present in the Ansible inventory
+  hosts: config-mgmt
+  gather_facts: no
+  vars:
+    # Set this to a colon-separated list of compute node hostnames on which to
+    # set the boot mode. If unset, all compute nodes will be set.
+    compute_node_limit: ""
+    compute_node_limit_list: "{{ compute_node_limit.split(':') }}"
+  tasks:
+    - name: Add hosts for the compute nodes
+      add_host:
+        name: "{{ item.key }}"
+        groups: compute
+      with_dict: "{{ idrac_network_ips }}"
+      # Don't add hosts that already exist.
+      when:
+        - "{{ item.key not in groups['all'] }}"
+        - "{{ item.key | replace('-idrac', '') not in groups['all'] }}"
+        - "{{ not compute_node_limit or item.key | replace('-idrac', '') in compute_node_limit_list }}"
+      run_once: True
+
+- name: Ensure compute nodes boot mode is set
+  hosts: compute
+  gather_facts: no
+  vars:
+    # Set this to the required boot mode. One of 'bios' or 'uefi'.
+    drac_boot_mode: "bios"
+  tasks:
+    - name: Set a fact containing the compute node IPMI address
+      set_fact:
+        ansible_host: "{{ idrac_network_ips[inventory_hostname] }}"
+        ansible_user: "{{ ipmi_username }}"
+        ansible_ssh_pass: "{{ ipmi_password }}"
+
+    - name: Ensure the compute node's BMC is added as an SSH known host
+      include_role:
+        role: ssh-known-host
+
+    - name: Ensure the compute node's boot mode is set
+      include_role:
+        role: drac-boot-mode

ansible/roles/drac-boot-mode/README.md

@@ -0,0 +1,35 @@
+DRAC Boot Mode
+==============
+
+Ansible role to set the BIOS boot mode for a Dell server with a DRAC.
+
+Requirements
+------------
+
+None
+
+Role Variables
+--------------
+
+`drac_boot_mode`: Set this to the required boot mode.
+
+`drac_retries`: Number of times to attempt to perform write operations.
+
+`drac_delay`: Delay between successive write operations.
+
+Example Playbook
+----------------
+
+The following playbook sets the boot mode of a host to 'BIOS'.
+
+    ---
+    - name: Ensure DRAC boot mode is BIOS
+      hosts: dracs
+      roles:
+        - role: drac-boot-mode
+          drac_boot_mode: bios
+
+Author Information
+------------------
+
+- Mark Goddard (<mark@stackhpc.com>)

ansible/roles/drac-boot-mode/defaults/main.yml

@@ -0,0 +1,12 @@
+---
+# Set this to the required boot mode.
+drac_boot_mode: "bios"
+
+# Timeout when waiting for boot mode configuration to be applied.
+drac_boot_mode_timeout: 600
+
+# Number of times to attempt to perform write operations.
+drac_boot_mode_retries: 5
+
+# Interval between successive write operations.
+drac_boot_mode_interval: 5

ansible/roles/drac-boot-mode/tasks/main.yml

@@ -0,0 +1,64 @@
+---
+- name: Fail if the requested boot mode is unexpected
+  fail:
+    msg: >
+      Unexpected requested boot mode {{ drac_boot_mode }}. Expected one of
+      {{ drac_boot_mode_valid_modes | join(', ') }}.
+  when: "{{ drac_boot_mode | lower not in drac_boot_mode_valid_modes }}"
+
+- name: Check the boot mode
+  raw: "racadm get BIOS.BiosBootSettings.BootMode"
+  register: result
+  failed_when: "'ERROR' in result.stdout"
+  changed_when: False
+
+- name: Set a fact containing the current boot mode
+  set_fact:
+    # Format of the last line is:
+    # BootMode=<boot mode>[ (Pending Value=<pending boot mode>)]
+    current_boot_mode: "{{ result.stdout_lines[-1].partition('=')[2] | lower }}"
+
+- name: Fail if the current boot mode is unexpected
+  fail:
+    msg: >
+      Unexpected current boot mode {{ current_boot_mode }}. Expected one of
+      {{ drac_boot_mode_valid_modes | join(', ') }}.
+  when: "{{ current_boot_mode not in drac_boot_mode_valid_modes }}"
+
+- block:
+    - name: Set the boot mode
+      raw: "racadm set BIOS.BiosBootSettings.BootMode {{ drac_boot_mode | lower | capitalize }}"
+      register: result
+      failed_when: "'ERROR' in result.stdout"
+      until: "{{ drac_boot_mode_busy_message not in result.stdout }}"
+      retries: "{{ drac_boot_mode_retries }}"
+      delay: "{{ drac_boot_mode_interval }}"
+
+    - name: Ensure BIOS configuration job is created
+      raw: "racadm jobqueue create BIOS.Setup.1-1 -s TIME_NOW"
+      register: result
+      failed_when: "'ERROR' in result.stdout"
+      until: "{{ drac_boot_mode_busy_message not in result.stdout }}"
+      retries: "{{ drac_boot_mode_retries }}"
+      delay: "{{ drac_boot_mode_interval }}"
+
+    - name: Set a fact containing the BIOS configuration job ID
+      set_fact:
+        # Format of the last line is:
+        # JOB_ID = <job ID>
+        drac_boot_mode_bios_job_id: "{{ result.stdout_lines[-1].split()[-1] }}"
+
+    - name: Ensure server is rebooted
+      raw: "racadm serveraction powercycle"
+      register: result
+      failed_when: "'ERROR' in result.stdout"
+
+    - name: Wait for the BIOS configuration job to complete
+      raw: "racadm jobqueue view -i {{ drac_boot_mode_bios_job_id }}"
+      register: result
+      failed_when: "'ERROR' in result.stdout"
+      until: "{{ 'Status=Completed' in result.stdout }}"
+      retries: "{{ drac_boot_mode_timeout // drac_boot_mode_interval }}"
+      delay: "{{ drac_boot_mode_interval }}"
+
+  when: "{{ current_boot_mode != drac_boot_mode }}"

ansible/roles/drac-boot-mode/vars/main.yml

@@ -0,0 +1,8 @@
+---
+# List of valid DRAC boot modes.
+drac_boot_mode_valid_modes:
+  - "bios"
+  - "uefi"
+
+# Message emitted by the DRAC which can be retried.
+drac_boot_mode_busy_message: "ERROR: Lifecycle Controller is currently in use."