Hugo Brito 05b7684f3e Create app-distributed-cloud prototype

This commit introduces the prototype for the app-distributed-cloud

All dcmanager services are being containerized, and the app will
utilize Keystone, certmanager, and rabbitmq from the platform.

The prototype is in its initial phase, and DC functionalities/services
have not yet been thoroughly tested. In this first phase, the subcloud
add command is functional.

This commit includes the dcmanager and dc-vault helm charts, and a
a structure of dcorch, that will be added in a following commit.

Before testing the app, the platform must be configured. Refer to the
README file for details.

Test Plan:
PASS: Execute a subcloud add successfully.

Story: 2011312
Task: 51841

Change-Id: Ifa6dbbb39d5bdd48eedd06c732d24d26a48f6ae8
Co-Authored-By: Matt Peters <matt.peters@windriver.com>
Co-Authored-By: Enzo Candotti <enzo.candotti@windriver.com>
Co-Authored-By: Yuxing Jiang <yuxing.jiang@windriver.com>
Signed-off-by: Hugo Brito <hugo.brito@windriver.com>

2025-03-25 15:08:11 -03:00

10 KiB

Raw Blame History

app-distributed-cloud (Prototype)

This tutorial provides a step-by-step guide on containerizing DC Services using the app-distributed-cloud prototype.

Note: All dcmanager operations are not fully tested or operational.

Disable Service Management

Disable the dcmanager services on the platform

source /etc/platform/openrc

sudo sm-unmanage service dcmanager-manager
sudo sm-unmanage service dcmanager-api
sudo sm-unmanage service dcmanager-audit
sudo sm-unmanage service dcmanager-audit-worker
sudo sm-unmanage service dcmanager-orchestrator
sudo sm-unmanage service dcmanager-state

sudo sm-unmanage service dcorch-engine
sudo sm-unmanage service dcorch-engine-worker
sudo sm-unmanage service dcorch-sysinv-api-proxy
sudo sm-unmanage service dcorch-patch-api-proxy
sudo sm-unmanage service dcorch-identity-api-proxy

sudo sm-unmanage service dcdbsync-api


sudo pkill -f ^".*/bin/dcmanager.*"
sudo pkill -f ^".*/bin/dcorch.*"
sudo pkill -f ^".*/bin/dcdbsync.*"

Platform Setup

system host-label-assign controller-0 starlingx.io/distributed-cloud=enabled
system host-label-assign controller-1 starlingx.io/distributed-cloud=enabled

Note: If you have issues with downloading the nginx image for dc-vault-nginx, assign the distributed-cloud label just for the controller-0

Create the namespace and root-ca secret

# Create distributed-cloud namespace

kubectl create namespace distributed-cloud

# Create system-local-ca secret

cp /etc/ssl/certs/dc-adminep-root-ca.pem /home/sysadmin/root-ca.pem

kubectl -n distributed-cloud create secret generic root-ca --from-file=ca.crt=/home/sysadmin/root-ca.pem

Distributed Cloud Application Deployment (development)

# Configure Docker Image
# Create or download the docker image used for the dcmanager pods
DOCKER_IMAGE=registry.local:9001/docker.io/starlingx/stx-distributed-cloud:master-debian-stable-latest

sudo docker login registry.local:9001

sudo docker image pull <POD_IMAGE>
sudo docker image tag <POD_IMAGE> ${DOCKER_IMAGE}
sudo docker image push ${DOCKER_IMAGE}

# Upload the prototype
system application-upload /usr/local/share/applications/helm/distributed-cloud-25.09-0.tgz

# Set Password Variables
ADMIN_KS_PASSWORD=$(keyring get CGCS admin)
RABBITMQ_PASSWORD=$(keyring get amqp rabbit)
DCMANAGER_DB_PASSWORD=$(keyring get dcmanager database)
DCMANAGER_KS_PASSWORD=$(keyring get dcmanager services)
DCORCH_DB_PASSWORD=$(keyring get dcorch database)
DCORCH_KS_PASSWORD=$(keyring get dcorch services)

# Create dcmanager and dcorch overrides
cat<<EOF>dcmanager.yaml
images:
  tags:
    dcmanager: ${DOCKER_IMAGE}
    ks_user: ${DOCKER_IMAGE}
    ks_service: ${DOCKER_IMAGE}
    ks_endpoints: ${DOCKER_IMAGE}
    dcmanager_db_sync: ${DOCKER_IMAGE}
    db_init: ${DOCKER_IMAGE}
    db_drop: ${DOCKER_IMAGE}
  pullPolicy: Always
pod:
  image_pull_secrets:
    default:
      - name: default-registry-key
  tolerations:
    dcmanager:
      enabled: true
conf:
  dcmanager:
    DEFAULT:
      log_config_append: /etc/dcmanager/logging.conf
      transport_url: rabbit://guest:${RABBITMQ_PASSWORD}@controller.internal:5672
      auth_strategy: keystone
      playbook_timeout: 3600
      use_usm: False
      workers: 1
      orch_workers: 1
      state_workers: 1
      audit_workers: 1
      audit_worker_workers: 1
    cache:
      auth_uri: http://controller.internal:5000/v3
      admin_tenant: admin
      admin_username: admin
      admin_password: ${ADMIN_KS_PASSWORD}
    endpoint_cache:
      auth_uri: http://controller.internal:5000/v3
      auth_plugin: password
      username: dcmanager
      password: ${DCMANAGER_KS_PASSWORD}
      project_name: services
      user_domain_name: Default
      project_domain_name: Default
      http_connect_timeout: 15
    database:
      connection_recycle_time: 3600
      max_pool_size: 105
      max_overflow: 100
    keystone_authtoken:
      auth_version: v3
      auth_type: password
dependencies:
  static:
    api:
      jobs:
        - dcmanager-ks-user
        - dcmanager-ks-service
        - dcmanager-ks-endpoints
    ks_endpoints:
      jobs:
        - dcmanager-ks-user
        - dcmanager-ks-service
endpoints:
  cluster_domain_suffix: cluster.local
  oslo_db:
    auth:
      admin:
        username: admin-dcmanager
        password: ${DCMANAGER_DB_PASSWORD}
      dcmanager:
        username: admin-dcmanager
        password: ${DCMANAGER_DB_PASSWORD}
    hosts:
      default: postgresql
    host_fqdn_override:
      default: controller.internal
    port:
      postgresql:
        default: 5432
    path: /dcmanager
    scheme: postgresql+psycopg2
  oslo_messaging:
    auth:
      admin:
        username: guest
        password: ${RABBITMQ_PASSWORD}
      dcmanager:
        username: guest
        password: ${RABBITMQ_PASSWORD}
    hosts:
      default: rabbitmq
    host_fqdn_override:
      default: controller.internal
    path: /
    scheme: rabbit
    port:
      amqp:
        default: 5672
      http:
        default: 15672
  identity:
    name: keystone
    auth:
      admin:
        username: admin
        password: ${ADMIN_KS_PASSWORD}
        region_name: RegionOne
        project_name: admin
        user_domain_name: Default
        project_domain_name: Default
      dcmanager:
        role: admin
        username: dcmanager
        password: ${DCMANAGER_KS_PASSWORD}
        region_name: RegionOne
        project_name: services
        user_domain_name: Default
        project_domain_name: Default
    hosts:
      default: keystone-api
      public: keystone
    host_fqdn_override:
      default: controller.internal
    path:
      default: /v3
    scheme:
      default: http
    port:
      api:
        default: 80
        internal: 5000
  dcmanager:
    name: dcmanager
    hosts:
      default: dcmanager-api
      public: dcmanager
    host_fqdn_override:
      default: null
    path:
      default: /v1.0
    scheme:
      default: 'http'
    port:
      api:
        default: 8119
        public: 80
EOF

cat<<EOF>dcorch.yaml
images:
  tags:
    dcorch: ${DOCKER_IMAGE}
    ks_user: ${DOCKER_IMAGE}
    ks_service: ${DOCKER_IMAGE}
    ks_endpoints: ${DOCKER_IMAGE}
    dcorch_db_sync: ${DOCKER_IMAGE}
    db_init: ${DOCKER_IMAGE}
    db_drop: ${DOCKER_IMAGE}
  pullPolicy: Always
pod:
  image_pull_secrets:
    default:
      - name: default-registry-key
  tolerations:
    dcorch:
      enabled: true
  replicas:
    dcorch_engine_worker: 1
    dcorch_sysinv_api_proxy: 1
    keystone_api_proxy: 1
    dcorch_patch_api_proxy: 1
    dcorch_usm_api_proxy: 1
conf:
  dcorch:
    DEFAULT:
      log_config_append: /etc/dcorch/logging.conf
      transport_url: rabbit://guest:${RABBITMQ_PASSWORD}@controller.internal:5672
      auth_strategy: keystone
      playbook_timeout: 3600
      use_usm: False
    endpoint_cache:
      password: ${DCMANAGER_KS_PASSWORD}
    database:
      connection_recycle_time: 3600
      max_pool_size: 105
      max_overflow: 100
    keystone_authtoken:
      auth_version: v3
      auth_type: password
dependencies:
  static:
    api:
      jobs:
        - dcorch-ks-user
        - dcorch-ks-service
        - dcorch-ks-endpoints
    ks_endpoints:
      jobs:
        - dcorch-ks-user
        - dcorch-ks-service

endpoints:
  cluster_domain_suffix: cluster.local
  oslo_db:
    auth:
      admin:
        username: admin-dcorch
        password: ${DCORCH_DB_PASSWORD}
      dcorch:
        username: admin-dcorch
        password: ${DCORCH_DB_PASSWORD}
      dcmanager:
        username: admin-dcmanager
        password: ${DCMANAGER_DB_PASSWORD}
    hosts:
      default: postgresql
    host_fqdn_override:
      default: controller.internal
    port:
      postgresql:
        default: 5432
    path: /dcorch
    scheme: postgresql+psycopg2
  oslo_messaging:
    auth:
      admin:
        username: guest
        password: ${RABBITMQ_PASSWORD}
      dcmanager:
        username: guest
        password: ${RABBITMQ_PASSWORD}
    hosts:
      default: rabbitmq
    host_fqdn_override:
      default: controller.internal
    path: /
    scheme: rabbit
    port:
      amqp:
        default: 5672
      http:
        default: 15672
  identity:
    name: keystone
    auth:
      admin:
        username: admin
        password: ${ADMIN_KS_PASSWORD}
        region_name: RegionOne
        project_name: admin
        user_domain_name: Default
        project_domain_name: Default
      dcorch:
        role: admin
        username: dcorch
        password: ${DCORCH_KS_PASSWORD}
        region_name: RegionOne
        project_name: services
        user_domain_name: Default
        project_domain_name: Default
    hosts:
      default: keystone-api
      public: keystone
    host_fqdn_override:
      default: controller.internal
    path:
      default: /v3
    scheme:
      default: http
    port:
      api:
        default: 80
        internal: 5000
  dcorch:
    name: dcorch
    hosts:
      default: dcorch-api
      public: dcorch
    host_fqdn_override:
      default: null
    path:
      default: /v1.0
    scheme:
      default: 'http'
    port:
      api:
        default: 8118
        public: 80
EOF

system helm-override-update distributed-cloud dcmanager distributed-cloud --values dcmanager.yaml
system helm-override-update distributed-cloud dcorch distributed-cloud --values dcorch.yaml

system helm-override-show distributed-cloud dcmanager distributed-cloud
system helm-override-show distributed-cloud dcorch distributed-cloud

Apply app-distributed-cloud

system application-apply distributed-cloud
system application-show distributed-cloud

To remove

system application-remove distributed-cloud
system application-delete distributed-cloud

Check dcmanager endpoints

openstack endpoint list | grep dcmanager

Check if dcmanager-api endpoint works

kubectl get svc dcmanager-api -n distributed-cloud
kubectl get endpoints dcmanager-api -n distributed-cloud

# Get Token
openstack token issue

curl -i http://<endpoint>/v1.0/subclouds -X GET -H "Content-Type: application/json" -H "Accept: application/json" -H "X-Auth-Token:${TOKEN}"

Configure dcmanager-client

Edit file: /usr/lib/python3/dist-packages/dcmanagerclient/api/v1/client.py

_DEFAULT_DCMANAGER_URL = (
    "http://dcmanager-api.distributed-cloud.svc.cluster.local:8119/v1.0"
)

# delete if not dcmanager_url: to always set default
dcmanager_url = _DEFAULT_DCMANAGER_URL

Check dcmanager-manager is working

dcmanager subcloud-group add --name test
dcmanager subcloud update --group 2 subcloud2-stx-latest

10 KiB Raw Blame History