Using ACM to deploy SNO

Using ACM to deploy SNO

In the previous post, we talked about the integration of ACM into DCI. Now is the time to get some hands-on experience installing an SNO instance in the DCI way.

Running a deployment

The recommended method to interact with DCI is using pipelines. Some configurations need an inventory that provides details for the target assets. The following steps will help to generate the required configuration files:

1. Create the inventory file

The YAML file contains information about the SNO instance, the target hosts, and how to connect to its BMC.

inventory.yml

all:
    hosts:
        jumphost:
            ansible_connection: local
        provisioner:
            ansible_connection: local
            ansible_user: dciteam
vars:
    cluster: serverX-sno
    dci_disconnected: true
    acm_force_deploy: true
    acm_cluster_name: sno1
    acm_base_domain: sno.mydomain.lab
    acm_bmc_address: 192.168.10.NN
    acm_boot_mac_address: 3c:fd:fe:c2:11:11
    acm_machine_cidr: 192.168.NNN.0/24
    acm_bmc_user: REDACTED
    acm_bmc_pass: REDACTED
    provision_cache_store: "/opt/cache"

2. Create the Hub pipeline

The file provides parameters for DCI. It sets the target OCP version for the ACM Hub, and the outputs that can be available for the next jobs. Please note one of the important settings is enable_acm: true. This instructs the agent to deploy and configure the ACM as a post-deployment task.

ocp-4.10-acm-hub-pipeline.yml

---
- name: openshift-acm-hub
stage: acm-hub
ansible_playbook: /usr/share/dci-openshift-agent/dci-openshift-agent.yml
ansible_cfg: /var/lib/dci/pipelines/ansible.cfg
ansible_inventory: /var/lib/dci/inventories/<lab>/8nodes/inventory.yml
dci_credentials: /etc/dci-openshift-agent/<dci_credentials.yml>
configuration: "@QUEUE"
pipeline_user: /etc/dci-openshift-agent/pipeline_user.yml
ansible_extravars:
    dci_config_dirs:
    - /var/lib/dci/<lab>-config/dci-openshift-agent
    dci_local_log_dir: /var/lib/dci-pipeline/upload-errors
    dci_gits_to_components:
    - /var/lib/dci/<lab>-config/dci-openshift-agent
    - /var/lib/dci/inventories
    - /var/lib/dci/pipelines
    dci_tags: []
    dci_cache_dir: /var/lib/dci-pipeline
    dci_base_ip: "{{ ansible_default_ipv4.address }}"
    dci_baseurl: "http://{{ dci_base_ip }}"
    enable_acm: true
    dci_teardown_on_success: false
topic: OCP-4.10
components:
    - ocp
outputs:
    kubeconfig: "kubeconfig"
success_tag: ocp-acm-hub-4.10-ok

3. Create the SNO pipeline

The following pipeline definition sets the install_type as ACM. In this case, it is mandatory that the previous stage is of type acm-hub. It should not be an issue if you are sure that the cluster is already running ACM. The file sets the target version for the SNO. This also requires the installation of the performance-addon-operator.

ocp-4.10-acm-sno-pipeline.yml

---
- name: openshift-acm-sno
type: ocp
prev_stages: [acm-hub]
ansible_playbook: /usr/share/dci-openshift-agent/dci-openshift-agent.yml
ansible_cfg: /var/lib/dci/pipelines/ansible.cfg
dci_credentials: /etc/dci-openshift-agent/dci_credentials.yml
configuration: "@QUEUE"
pipeline_user: /etc/dci-openshift-agent/pipeline_user.yml
ansible_inventory: /home/dciteam/inventories/dallas/sno/sno1-cluster4.yml
ansible_extravars:
    install_type: acm
    acm_cluster_type: sno
    dci_local_log_dir: /var/lib/dci-pipeline/upload-errors
    dci_gits_to_components:
    - /var/lib/dci/dallas-config/dci-openshift-agent
    - /var/lib/dci/inventories
    - /var/lib/dci/pipelines
    dci_tags: []
    dci_cache_dir: /var/lib/dci-pipeline
    dci_base_ip: "{{ ansible_default_ipv4.address }}"
    dci_baseurl: "http://{{ dci_base_ip }}"
    dci_teardown_on_success: false
    enable_perf_addon: true
topic: OCP-4.10
components:
    - ocp
inputs:
    kubeconfig: hub_kubeconfig_path
outputs:
    kubeconfig: "kubeconfig"
success_tag: ocp-acm-sno-4.10-ok

4. Run a full pipeline

Now that we have the required components, executing these two jobs will go through the following tasks:

• Install OCP and the ACM operator on top. It generates a KUBECONFIG file.
• The second job uses the KUBECONFIG to interact with ACM and deploy the SNO instance.
• The generation of a second KUBECONFIG happens, this time for the SNO instance.
• The installation of operators is completed as the final step.

  1	$ dci-pipeline-schedule ocp-4.10-acm-hub ocp-4.10-acm-sno

5. Re-use an ACM Hub

It is not required to deploy an ACM Hub each time, it can be re-used to deploy many clusters. Setting the HUB_KUBECONFIG path environment variable before executing the acm-sno pipeline is enough once the Hub installation is complete. There should be many managed clusters listed in the ACM console.

    #!bash
    $ export HUB_KUBECONFIG=/acm-hub-kubeconfig-path
    $ dci-pipeline-schedule ocp-4.10-acm-sno

Troubleshooting

Please run the following commands to help identify issues during a cluster deployment like:

• ACM being unable to pull the images.
• The selected OCP release is wrong.
• Unable to connect to the BMC.
• Incorrect cluster settings.

  1 2 3 4 5 6

  $ oc logs -n multicluster-engine -l app=assisted-service $ oc logs -n multicluster-engine -l app=assisted-image-service $ oc project <cluster_name> $ oc describe infraenv <cluster_name> $ oc describe bmc <cluster_name> $ oc describe agentclusterinstall <cluster_name>

References

For more information about the variables used in the examples, please read the ACM's roles and DCI documentation.

• acm-setup/readme
• acm-sno/readme
• dci-openshift-agent/readme

Wrap up

Feeling overwhelmed with all the requirements to install OCP using UPI/IPI? Need something easy to manage your distributed fleet of OCP clusters? Then give ACM a try.

And if you want to start testing SNO via ACM to expand your CI, please try the DCI integration.