docs(ansible): refresh dual-oop workflow

Deploy Operator Platform environments with Ansible.

This repository is mainly used to:

Repository scope:

- prepare target hosts

- create Kind-based Kubernetes clusters

- prepare remote hosts

- create Kind clusters

- deploy Operator Platform components

- run single-host or dual-host federation scenarios

- deploy two-host federation setups

- fetch kubeconfigs back to local machine

## Start here

- `docs/getting-started.md`: environment, inventory, secrets, first deployment

- `docs/deployment.md`: scenario selection, commands, kubeconfigs, undeploy

- `docs/deployment.md`: working deployment, undeploy, verification, smoke test flow

## Quick start

pip install -r requirements.txt

ansible-galaxy collection install -r requirements.yml

ansible-playbook playbooks/scenarios/all_in_one/deploy.yml -e @secrets.yml

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml --limit openop_2,openop_3

```

## Main scenarios

## Working scenarios

- `playbooks/scenarios/all_in_one/deploy.yml`: one host, easiest test setup

- `playbooks/scenarios/full_oop/deploy.yml`: one complete Operator Platform

- `playbooks/scenarios/dual_oop/deploy.yml`: two hosts, real federation

- `playbooks/scenarios/dual_oop/deploy.yml`: two complete Operator Platforms on two hosts

- `playbooks/scenarios/full_oop/deploy.yml`: one complete Operator Platform on one host

## Kubeconfigs

- `~/kind-cluster-configs/openop_3/op1-kubeconfig.yaml`

- `~/kind-cluster-configs/openop_2/op2-kubeconfig.yaml`

## Notes

- Run playbooks from `ansible/`.

- `ansible/secrets.yml` is local-only and gitignored.

- `dual_oop` and `full_oop` load `secrets.yml` automatically.

## Docs preview

```bash

## Scenarios

### `all_in_one`

One host, easiest way to test the stack.

```bash

ansible-playbook playbooks/scenarios/all_in_one/deploy.yml -e @secrets.yml

```

### `full_oop`

One complete Operator Platform on one host.

```bash

ansible-playbook playbooks/scenarios/full_oop/deploy.yml -e @secrets.yml

ansible-playbook playbooks/scenarios/full_oop/deploy.yml --limit openop_3

```

### `dual_oop`

Two complete Operator Platforms on two hosts.

```bash

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml -e @secrets.yml

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml --limit openop_2,openop_3

```

Use tags to deploy only one side:

Deploy one side only:

```bash

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml -e @secrets.yml --tags op1

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml -e @secrets.yml --tags op2

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml --limit openop_3

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml --limit openop_2

```

## Undeploy

Examples:

```bash

ansible-playbook playbooks/scenarios/all_in_one/quick_undeploy.yml

ansible-playbook playbooks/scenarios/dual_oop/graceful_undeploy.yml

ansible-playbook playbooks/scenarios/full_oop/quick_undeploy.yml --limit openop_3

ansible-playbook playbooks/scenarios/dual_oop/quick_undeploy.yml --limit openop_2,openop_3

```

## Inventory expectations

## Notes

- Run from the `ansible/` directory.

- Pass `-e @secrets.yml` for private images.

- `secrets.yml` is loaded by `dual_oop` and `full_oop`.

- For remote deployments, kubeconfigs are fetched to `~/kind-cluster-configs/`.

        cluster_name: "op1"

        kind_cluster_delete_name: "op1"

    - name: Delete OP1 Kind Cluster

    - name: Get existing OP1 Kind clusters

      ansible.builtin.command:

        cmd: kind delete cluster --name {{ kind_cluster_delete_name }}

        cmd: kind get clusters

      register: op1_kind_clusters

      changed_when: false

      failed_when: false

    - name: Select OP1 clusters to delete

      ansible.builtin.set_fact:

        op1_clusters_to_delete: >-

          {{

            [kind_cluster_delete_name, 'operator-platform']

            | select('in', op1_kind_clusters.stdout_lines | default([]))

            | unique

            | list

          }}

    - name: Delete OP1 Kind clusters

      ansible.builtin.command:

        cmd: kind delete cluster --name {{ item }}

      loop: "{{ op1_clusters_to_delete }}"

      register: delete_op1_result

      changed_when: delete_op1_result.rc == 0

      failed_when: false

    - name: Display OP1 deletion result

      ansible.builtin.debug:

        msg: "OP1 cluster deletion ({{ kind_cluster_delete_name }}) {{ 'successful' if delete_op1_result.rc == 0 else 'failed or cluster not found' }}"

        msg: >-

          OP1 deleted clusters: {{ op1_clusters_to_delete | join(', ') if op1_clusters_to_delete | length > 0 else 'none found' }}

    - name: Clean up OP1 kubeconfig directory

      ansible.builtin.file:

        cluster_name: "op2"

        kind_cluster_delete_name: "op2"

    - name: Delete OP2 Kind Cluster

    - name: Get existing OP2 Kind clusters

      ansible.builtin.command:

        cmd: kind get clusters

      register: op2_kind_clusters

      changed_when: false

      failed_when: false

    - name: Select OP2 clusters to delete

      ansible.builtin.set_fact:

        op2_clusters_to_delete: >-

          {{

            [kind_cluster_delete_name, 'operator-platform']

            | select('in', op2_kind_clusters.stdout_lines | default([]))

            | unique

            | list

          }}

    - name: Delete OP2 Kind clusters

      ansible.builtin.command:

        cmd: kind delete cluster --name {{ kind_cluster_delete_name }}

        cmd: kind delete cluster --name {{ item }}

      loop: "{{ op2_clusters_to_delete }}"

      register: delete_op2_result

      changed_when: delete_op2_result.rc == 0

      failed_when: false

    - name: Display OP2 deletion result

      ansible.builtin.debug:

        msg: "OP2 cluster deletion ({{ kind_cluster_delete_name }}) {{ 'successful' if delete_op2_result.rc == 0 else 'failed or cluster not found' }}"

        msg: >-

          OP2 deleted clusters: {{ op2_clusters_to_delete | join(', ') if op2_clusters_to_delete | length > 0 else 'none found' }}

    - name: Clean up OP2 kubeconfig directory

      ansible.builtin.file:

- name: Deploy Full OOP (Single Federation Manager)

  hosts: k8s_clusters

  gather_facts: true

  vars_files:

    - ../../../secrets.yml

  vars:

    # Scenario-specific configuration

# Deployment Guide

## Choose a scenario

## Working scenarios

### `all_in_one`

### `dual_oop`

Use this when you want the simplest setup on one host.

Use this for real federation across two hosts.

```bash

ansible-playbook playbooks/scenarios/all_in_one/deploy.yml -e @secrets.yml

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml --limit openop_2,openop_3

```

Good for:

- first deployment

- local validation

- quick federation-related testing on one VM

### `full_oop`

Use this when you want one complete Operator Platform on one host.

```bash

ansible-playbook playbooks/scenarios/full_oop/deploy.yml -e @secrets.yml

```

This deploys:

### `dual_oop`

- OP1 on `op1_nodes`

- OP2 on `op2_nodes`

Use this when you want two separate Operator Platforms on two hosts.

You can deploy one side only if needed:

```bash

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml -e @secrets.yml

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml --limit openop_3

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml --limit openop_2

```

This deploys:

- OP1 on `op1_nodes`

- OP2 on `op2_nodes`

### `full_oop`

You can also deploy one side only:

Use this for one complete Operator Platform on one host.

```bash

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml -e @secrets.yml --tags op1

ansible-playbook playbooks/scenarios/dual_oop/deploy.yml -e @secrets.yml --tags op2

ansible-playbook playbooks/scenarios/full_oop/deploy.yml --limit openop_3

```

## Where kubeconfigs go

For remote deployments, the playbooks fetch kubeconfigs to your local machine:

For remote deployments, playbooks fetch kubeconfigs locally to:

```text

~/kind-cluster-configs/<inventory-host>/

## Access points

Default service URLs:

Useful service URLs:

- OEG northbound: `http://<host-ip>:32263/oeg/1.0.0`

- Federation Manager: `http://<host-ip>:30989`

- Homer: `http://<host-ip>:30088`

- Zot: `http://<host-ip>:30050`

- Federation Manager: `http://<host-ip>:30989`

- Prometheus: `http://<host-ip>:30090`

- Grafana: `http://<host-ip>:30091`

## End-to-end smoke test

After `dual_oop` is green, run smoke from `local-docker-deployment/` in repository root.

Command:

```bash

bash smoke-test.sh \

  --image "docker.io/nginx:latest" \

  --app-name "test-app" \

  --local-url "http://192.168.123.155:32263/oeg/1.0.0" \

  --remote-url "http://192.168.123.178:32263/oeg/1.0.0" \

  --local-provider "Local Operator" \

  --federated-provider "Remote Operator"

```

Notes:

- use OEG northbound URL on port `32263`, not Federation Manager on `30989`

- current smoke script auto-resolves local and remote zone IDs

- current smoke script auto-normalizes invalid app names like `test-app` to `test_app`

- script deploys and then cleans up app instances in one run

Expected success signals:

- local app instance reaches `ready`

- federated app instance reaches `ready`

- cleanup leaves `Remaining matching app instances: []`

## Undeploy

Quick removal deletes the cluster directly:

Quick removal deletes clusters directly:

```bash

ansible-playbook playbooks/scenarios/all_in_one/quick_undeploy.yml

ansible-playbook playbooks/scenarios/full_oop/quick_undeploy.yml

ansible-playbook playbooks/scenarios/dual_oop/quick_undeploy.yml

ansible-playbook playbooks/scenarios/dual_oop/quick_undeploy.yml --limit openop_2,openop_3

ansible-playbook playbooks/scenarios/full_oop/quick_undeploy.yml --limit openop_3

```

Graceful removal undeploys components first:

```bash

ansible-playbook playbooks/scenarios/all_in_one/graceful_undeploy.yml

ansible-playbook playbooks/scenarios/full_oop/graceful_undeploy.yml

ansible-playbook playbooks/scenarios/dual_oop/graceful_undeploy.yml

ansible-playbook playbooks/scenarios/dual_oop/graceful_undeploy.yml --limit openop_2,openop_3

ansible-playbook playbooks/scenarios/full_oop/graceful_undeploy.yml --limit openop_3

```

## Troubleshooting

### Port already in use

The Kind role now checks port conflicts before cluster creation. If a deployment fails early, free the reported ports or remove old containers/clusters on the target host.

If deployment fails before Kind cluster creation, old clusters or containers still own NodePort mappings. Remove stale Kind clusters on target hosts, then retry.

### SRM or another component stays in `FAILED - RETRYING`

### Deployment succeeds but a component is slow

Read pod status with fetched kubeconfig:

`FAILED - RETRYING` during readiness checks is normal if the task later ends with `ok`.

```bash

export KUBECONFIG=~/kind-cluster-configs/openop_3/op1-kubeconfig.yaml

kubectl get pods -A

kubectl describe pod -n oop <pod-name>

```

### Verify cluster health