Skip to content

Latest commit

 

History

History
116 lines (114 loc) · 5.17 KB

connect_to_k8s_env.md

File metadata and controls

116 lines (114 loc) · 5.17 KB

Deploying Host Inventory and Xjoin-search to Kubernetes Namespaces.

Debugging Host Inventory code requires a database, kakfa broker, Kafka Zookeeper, Debezium Connector, and xjoin-search. These required services are deployed to Kubernetes using Kubernetes operators and bonfire.

Kubernetes Setup

Option 1: Minikube

  1. Start 'Minikube' with the following suggested configuration: 
        - disk-size: 200GB
    - driver: kvm2
    - memory: 16384
    - cpus: 6
  2. Connect to image repository for generating authentication token, which is used by bonfire for pulling images from image repository: 
        docker/podmin login -u <usrename> quay.io
  3. Clone the xjoin-operator code repository and cd into the xjoin-operator directory.
  4. Deploy host-inventory: 
        ./dev/setup-clowder.sh
  5. Install CRDs: 
        make install
  6. Run the xjoin-operator: 
        make run ENABLE_WEBHOOKS=false
  7. Create a new xjoin pipeline: 
        kubectl apply -f ./config/samples/xjoin_v1alpha1_xjoinpipeline.yaml -n test

Option 2: Ephemeral Cluster

  1. Login to Ephemeral cluster.
  2. Reserve a namespace: 
        bonfire namespace reserve -d xxh    # the default reservation time is one hour
  3. Deploy host-inventory: 
        bonfire deploy host-inventory -n <namespace>
  4. Exposed service ports by downloading the forward-ports-clowder.sh script and running: 
        forward-ports-clowder.sh <namespace>

Connecting Local Host-Inventory to Kubernetes Setup

  1. In the host-inventory project directory, run: 
        pipenv install --dev
    pipenv shell
  2. Get Kafka broker address: 
        make run_inv_mq_service_test_producer
    The make command complains that the kafka broker is not available and should provide the address, which should look like env-ephemeral-spxayh-509fc239-kafka-0.env-ephemeral-spxayh-509fc239-kafka-brokers.ephemeral-spxayh.svc
  3. Add the Kafka broker address to your local /etc/hosts file 
        127.0.0.1 env-ephemeral-spxayh-509fc239-kafka-0.env-ephemeral-spxayh-509fc239-kafka-brokers.ephemeral-spxayh.svc
  4. To connect to database and elasticsearch pods, get credentials by downloading the get_credentials.sh script and running: 
        get_credentials.sh <namespace>
    get_credentials provides the credentials to access database and elasticsearch.
  5. Create a new host by running: 
        make run_inv_mq_service_test_producer
  6. Verify the newly created host is available: 
        curl --location 'http://localhost:8000/api/inventory/v1/hosts' \
         --header 'x-rh-identity: <base64-encoded idenity>' \
         --header 'Content-Type: application/json' \
         --header 'x-rh-cloud-bulk-query-source: xjoin'
    Though API GET is used, the host has been provided by xjoin-search which gets it from elasticsearch index.
  7. To get hosts directly from the Elasticsearch index, run: 
        curl --location 'http://localhost:9200/xjoin.inventory.hosts/_search' \
         --header 'Authorization: Basic <elastic-user value from get_credentials>='
    Note: index name, "xjoin.inventory.hosts" may be different in your case.

Setup Development environment for Debugging API Code

This section describes how to launch debugger using VS Code to execute the local API server which in turn uses resources deployed in the ephemeral namespace.

  1. Load the host-inventory project to VS Code
  2. Add a launch configuration to run the api-server run.py with the following environment variables: 
    {
    "version": "0.2.0",
    "configurations": [
        {
            "name": "Python: Current File",
            "type": "python",
            "request": "launch",
            "program": "run.py",
            "console": "integratedTerminal",
            "justMyCode": false,
            "env": {
                "BYPASS_RBAC": "true",
                "INVENTORY_DB_USER": "<user provided get_credentials>",
                "INVENTORY_DB_PASS": "<password provided get_credentials>",
                "INVENTORY_DB_NAME": "<DB name provided get_credentials>",
                "PROMETHEUS_MULTIPROC_DIR": "./temp/prometheus_multiproc_dir/",
                "prometheus_multiproc_dir": "./temp/prometheus_multiproc_dir/",
                "FLASK_ENV": "development"
            },
        },
    ]
}
  3. Launch the api-server on a port different from the one used by the host-inventory-service
  4. Use the curl statements provided above to debug and step through the local code.