More info is here
- Automate the process of enrolling baremetals into MaaS
- Onboarding is done by triggering Ansible playbooks.
Download the script and prerequisite files into the same directory.
Script directory structure:
.
├── cloud-init_tempalte.yaml
├── machines_tempalte.csv
├── vars_template.j2
├── main_script.py
├── lv_config_template.json
├── modules
│ ├── maasHelper.py ---> add and deploy machines to MAAS
│ ├── onboard.py ---> onboard machines to PCD
├── pcd_ansible-pcd_develop
│ ├── .....
│ ├── setup-local.sh ---> Set up the environment for PCD onboarding-
Maas cli login
maas login <maas_user> http://<maas_ip>:5240/MAAS/ $(sudo maas apikey --generate --username=<maas_user>) -
Clouds.yaml created in /{home}/.config/openstack/clouds.yaml
clouds: titan: auth: auth_url: https://<DU-name>.app.staging-pcd.platform9.com/keystone/v3 project_name: service username: abc@ACME.com password: ********* user_domain_name: default project_domain_name: default region_name: titan interface: public identity_api_version: 3 compute_api_version: 2 volume_api_version: 3 image_api_version: 2 identity_interface: public volume_interface: public -
untar the prerequisites file
tar -xzvf prerequisites.tar.gz
-
Set up the environment for PCD onboarding by running the script setup-local.sh from inside pcd_ansible-pcd_develop directory.
sudo bash setup-local.sh
-
You need to create/edit the following files according to your environment before running the script You can use the templates provided in the prerequisites folder:
-
machines_template.csv
hostname,architecture,mac_addresses,power_type,power_user,power_pass,power_driver,power_address,cipher_suite_id,power_boot_type,privilege_level,k_g,ip,storage_ip,cloud_init pf9-test001,amd64/generic,3c:fd:fe:b5:1a:8d,ipmi,admin,password,LAN_2_0,172.25.1.11,3,auto,ADMIN,,192.168.125.167,192.168.125.165,cloud_init.yaml pf9-test002,amd64/generic,3c:fd:fe:b5:1a:8d,ipmi,admin,password,LAN_2_0,172.25.1.12,3,auto,ADMIN,,192.168.125.168,192.168.125.166,cloud_init2.yaml
optional fields:
- storage_ip
- cloud_init : In case you want to provide a different cloud-init template per server, you can add it in the CSV file instead of passing it as an argument via command line.
-
cloud-init-template.yaml
- The only requirement for this file is to have the IP and the storage_ip as a placeholder to be filled in dynamically for each machine
#cloud-config write_files: - path: /etc/cloud/cloud.cfg.d/99-disable-network-config.cfg permissions: "0644" content: | network: {config: disabled} - path: /etc/netplan/99-custom.yaml permissions: "0644" content: | network: version: 2 renderer: networkd ethernets: enp94s0f0: dhcp4: false enp24s0f1: dhcp4: false bonds: bond0: interfaces: - enp24s0f1 - enp94s0f0 parameters: mode: active-backup primary: enp24s0f1 mii-monitor-interval: 100 dhcp4: false addresses: - $ip/24 --> placeholder for the IP routes: - to: default via: 192.168.125.254 nameservers: addresses: - 8.8.8.8 vlans: vlan.301: id: 301 link: bond0 addresses: - $storage_ip/18 --> placeholder for the storage IP routes: - to: default via: 10.10.0.1 nameservers: addresses: - 8.8.8.8 - 1.1.1.1
-
lv_config_template.json (in case the --storage_layout flag is set to yes)
{ "vg_group": "rootvg", # the name for the vg group "boot_efi_size": "1G", "boot_size": "3G", "volumes": [ {"name": "home", "size": "10G" , "fs_type": "xfs" , "mount_point": "/home"}, {"name": "opt", "size": "10G" , "fs_type": "xfs" , "mount_point": "/opt"}, {"name": "opt-pf9", "size": "10G" , "fs_type": "xfs" , "mount_point": "/opt/pf9"}, {"name": "swap", "size": "3G" , "fs_type": "" , "mount_point": ""} , {"name": "tmp", "size": "10G" , "fs_type": "xfs" , "mount_point": "/tmp"}, {"name": "usr-openv", "size": "30G" , "fs_type": "xfs" , "mount_point": "/usr/openv"}, {"name": "var", "size": "20G" , "fs_type": "xfs" , "mount_point": "/var"}, {"name": "var-log", "size": "20G" , "fs_type": "xfs" , "mount_point": "/var/log"}, {"name": "var-log-audit", "size": "20G" , "fs_type": "xfs" , "mount_point": "/var/log/audit"}, {"name": "var-tmp", "size": "5G" , "fs_type": "xfs" , "mount_point": "/var/tmp"}, {"name": "root", "size": "20G" , "fs_type": "xfs" , "mount_point": "/"} ] }
-
-
Ensure the SSH key for the MAAS server user (used to connect to deployed machines during onboarding) is added in the MAAS UI.
python3 main_script.py \
--maas_user admin \
--csv_filename machines.csv \
--cloud_init_template cloud-init.yaml \
--portal exalt-pcd \
--region jrs \
--environment stage \
--url https://exalt-pcd-jrs.app.qa-pcd.platform9.com/ \
--max_workers 5 \
--ssh_user ubuntu \
--setup_env yesWhat the options mean:
--maas_user: MAAS admin username.--csv_filename: CSV file path.--cloud_init_template: Cloud-init template YAML path. If not provided, it will look for it in the CSV file.--max_workers: Maximum number of concurrent threads for provisioning.--ssh_user: SSH user for Ansible.--setup_env: By default, it's no, but for the first run on the maas machine, it should be yes to set up the necessary directories
There is an optional argument:
--preserve_cloud_init: By default, it's no, but when set to yes, it will keep all generated cloud-init files under/{script directory}/maas-cloud-init/cloud-init-{hostname}.yaml--storage_layout: By default, it's no, but if set to yes will allow the storage script to run and create custome storage layout--storage_layout_template: Required if --storage_layout is set to yes.--onprem: If the setup is on-prem, set this to yes.--controller_ip: Required if --onprem is set to yes.
Script directory structure after running the script:
.
├── cloud-init_tempalte.yaml
├── maas-cloud-init
│ └── cloud-init-{hostname}.yaml ---> generated cloud-init files for each machine
├── deploy_logs
│ └── maas_deployment.log ---> generated logs for MAAS
├── machines_tempalte.csv
├── {your CSV file name}_updated.csv ---> updated csv with the status of the deployment
├── vars_template.j2
├── vars.yaml ---> yaml file that will be used by the onboarding Ansible playbooks
├── lv_config_template.json
├── main_script.py
├── modules
│ ├── maasHelper.py
│ ├── onboard.py
├── pcd_ansible-pcd_develop
│ ├── .....
│ ├── logs
│ │ └── pcd_installer_logs
│ ├── setup-local.sh When in ready state, it will generate a cloud-init file for each machine with the IP specified for each one from the CSV file,it will be generated in the tmp directory,and then deploy the OS.
/maas-cloud-init/cloud-init-{hostname}.yamlThose files will be deleted once the deployment is done, unless the flag preserve_cloud_init is set to yes
After deployment, a new CSV file will be generated that contains all the previous info along with the deployment status to ensure that undeployed or uncommissioned machines don't go through the onboarding phase.
{your CSV file name}_updated.csv5. Will start by generating a vars.yaml file, which contains all the necessary information about each host, using Jinja2
- Loads a template (vars_template.j2) and fills it with the extracted data.
- Saves the rendered YAML to vars.yaml.
sample vars.yaml:
url: https://exalt-pcd-jrs.app.qa-pcd.platform9.com/
cloud: jrs
environment: stage
hosts:
192.168.125.165:
ansible_ssh_user: "ubuntu"
ansible_ssh_private_key_file: "/home/exalt/.ssh/id_rsa"
roles:
- "node_onboard"It will copy the vars file to user_resource_examples/templates/host_onboard_data.yaml.j2, where the Ansible playbooks will use this file