Skip to content

Latest commit

 

History

History
460 lines (409 loc) · 14.7 KB

README.md

File metadata and controls

460 lines (409 loc) · 14.7 KB

Cloud Migration Workflow Management

This is a subsystem of the Cloud-Barista platform that provides workflow management for cloud migration.

Overview

  • Create and management workflow through Airflow.
  • Create workflow based on gusty.

Development environment

  • Tested operating systems (OSs):
    • Ubuntu 24.04, Ubuntu 22.04, Ubuntu 18.04
  • Language:
    • Go: 1.23.0

How to run

1. Build and run

Build and run the binary with Airflow server

make run

Or, you can run it within Docker by this command.

make run_docker

If you want to stop the binary with Airflow server, run this command.

make stop

2. Write configuration file

  • Configuration file name is 'cm-cicada.yaml'
  • The configuration file must be placed in one of the following directories.
    • .cm-cicada/conf directory under user's home directory
    • 'conf' directory where running the binary
    • 'conf' directory where placed in the path of 'CMCICADA_ROOT' environment variable
  • Configuration options
    • task_component
      • load_examples : Load task component examples if true.
      • examples_directory : Specify directory where task component examples are located. Must be set if 'load_examples' is true.
    • workflow_template
      • templates_directory : Specify directory where workflow templates are located.
    • airflow-server
      • address : Specify Airflow server's address ({IP or Domain}:{Port})
      • use_tls : Must be true if Airflow server uses HTTPS.
      • skip_tls_verify : Skip TLS/SSL certificate verification. Must be set if 'use_tls' is true.
      • init_retry : Retry count of initializing Airflow server connection used by cm-cicada.
      • timeout : HTTP timeout value as seconds.
      • username : Airflow login username.
      • password : Airflow login password.
      • connections : Pre-define Airflow connections (Set multiple connections)
        • id : ID of connection
        • type : Type of connection
        • description : Description of connection
        • host : Host address or URL of connection
        • port : Port number for use connection
        • schema : Connection schema
        • login : Username for use connection
        • password : Password for use connection
    • dag_directory_host : Specify DAG directory of the host. (Mounted DAG directory used by Airflow container.)
    • dag_directory_container : Specify DAG directory of Airflow container. (DAG directory inside the container.)
  • listen
    • port : Listen port of the API.
  • Configuration file example
    cm-cicada:
      task_component:
          load_examples: true
          examples_directory: "./lib/airflow/example/task_component/"
      workflow_template:
          templates_directory: "./lib/airflow/example/workflow_template/"
      airflow-server:
          address: 127.0.0.1:8080
          use_tls: false
          # skip_tls_verify: true
          init_retry: 5
          timeout: 10
          username: "airflow"
          password: "airflow_pass"
          connections:
            - id: honeybee_api
              type: http
              description: HoneyBee API
              host: 127.0.0.1
              port: 8081
              schema: http
            - id: beetle_api
              type: http
              description: Beetle API
              host: 127.0.0.1
              port: 8056
              schema: http
              login: default
              password: default
            - id: tumblebug_api
              type: http
              description: TumbleBug API
              host: 127.0.0.1
              port: 1323
              schema: http
              login: default
              password: default
      dag_directory_host: "./_airflow/airflow-home/dags"
      dag_directory_container: "/usr/local/airflow/dags" # Use dag_directory_host for dag_directory_container, if this value is empty
      listen:
          port: 8083

3. Check workflow template

Check workflow template list.

curl -X 'GET' \
  'http://127.0.0.1:8083/cicada/workflow_template' \
  -H 'accept: application/json'

Get workflow template and copy the content.

curl -X 'GET' \
  'http://127.0.0.1:8083/cicada/workflow_template/81bbeb23-2c48-4536-9f01-55796e0fa394' \
  -H 'accept: application/json'

4. Create the workflow

Create the workflow by pasting copied workflow template content. Modify all of 'sgId' params. (Create the source group from honeybee.)

Check the ID of the created workflow.

curl -X 'POST' \
  'http://127.0.0.1:8083/cicada/workflow' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "name": "migrate_infra_workflow",
  "data": {
    "description": "Migrate Server",
    "task_groups": [
      {
        "name": "migrate_infra",
        "description": "Migrate Server",
        "tasks": [
          {
            "name": "infra_import",
            "task_component": "honeybee_task_import_infra",
            "request_body": "",
            "path_params": {
              "sgId": "ba695bbb-d673-4092-9821-c9cb05676228"
            },
            "dependencies": []
          },
          {
            "name": "infra_get",
            "task_component": "honeybee_task_get_infra_refined",
            "request_body": "",
            "path_params": {
              "sgId": "ba695bbb-d673-4092-9821-c9cb05676228"
            },
            "dependencies": [
              "infra_import"
            ]
          },
          {
            "name": "infra_recommend",
            "task_component": "beetle_task_recommend_infra",
            "request_body": "infra_get",
            "path_params": null,
            "dependencies": [
              "infra_get"
            ]
          },
          {
            "name": "infra_migration",
            "task_component": "beetle_task_infra_migration",
            "request_body": "infra_recommend",
            "path_params": null,
            "dependencies": [
              "infra_recommend"
            ]
          },
          {
            "name": "register_target_to_source_group",
            "task_component": "honeybee_register_target_info_to_source_group",
            "request_body": "infra_migration",
            "path_params": {
              "sgId": "ba695bbb-d673-4092-9821-c9cb05676228"
            },
            "dependencies": [
              "infra_migration"
            ]
          }
        ]
      }
    ]
  }
}'

5. Run the workflow

curl -X 'POST' \
  'http://127.0.0.1:8083/cicada/workflow/4420da6c-c50f-4d8b-bc2e-f02d8b557fad/run' \
  -H 'accept: application/json' \
  -d ''

About Task Component

Each task in the workflow references a Task Component.

The Task Component part frequently leads to mistakes, and when the Request Body structure is complex or changes, it becomes difficult to keep up. Therefore, we have implemented a feature that automatically generates Task Components by reading JSON files.

As shown below, using JSONs containing name, description, api_connection_id, swagger_yaml_endpoint, and endpoint, the cicada reads the Swagger YAML and finds the endpoint to automatically construct task components.

The api_connection_id corresponds to one of the connection ids defined in https://github.com/cloud-barista/cm-cicada/blob/main/conf/cm-cicada.yaml

{
  "name": "tumblebug_mci_dynamic",
  "description": "Create MCI Dynamically from common spec and image.",
  "api_connection_id": "tumblebug_api",
  "swagger_yaml_endpoint": "/tumblebug/api/doc.yaml",
  "endpoint": "/ns/{nsId}/mciDynamic"
}

Examples of these JSONs can be found at: https://github.com/cloud-barista/cm-cicada/tree/main/lib/airflow/example/task_component

The Task Component automatically generated from the above JSON is as follows:

Expand
{
  "id": "796645fc-c594-4263-a9ff-243051d1f3a5",
  "name": "tumblebug_mci_dynamic",
  "description": "Create MCI Dynamically from common spec and image.",
  "data": {
    "options": {
      "api_connection_id": "tumblebug_api",
      "endpoint": "/tumblebug/ns/{nsId}/mciDynamic",
      "method": "POST",
      "request_body": "{\n    \"description\": \"Made in CB-TB\",\n    \"installMonAgent\": \"no\",\n    \"label\": {},\n    \"name\": \"mci01\",\n    \"systemLabel\": \"\",\n    \"vm\": [\n        {\n            \"commonImage\": \"ubuntu18.04\",\n            \"commonSpec\": \"aws+ap-northeast-2+t2.small\",\n            \"connectionName\": \"string\",\n            \"description\": \"Description\",\n            \"label\": {},\n            \"name\": \"g1-1\",\n            \"rootDiskSize\": \"default, 30, 42, ...\",\n            \"rootDiskType\": \"default, TYPE1, ...\",\n            \"subGroupSize\": \"3\",\n            \"vmUserPassword\": \"string\"\n        }\n    ]\n}"
    },
    "body_params": {
      "required": [
        "name",
        "vm"
      ],
      "properties": {
        "description": {
          "type": "string",
          "example": "Made in CB-TB"
        },
        "installMonAgent": {
          "type": "string",
          "description": "InstallMonAgent Option for CB-Dragonfly agent installation ([yes/no] default:no)",
          "default": "no",
          "enum": [
            "yes",
            "no"
          ],
          "example": "no"
        },
        "label": {
          "type": "object",
          "description": "Label is for describing the object by keywords"
        },
        "name": {
          "type": "string",
          "example": "mci01"
        },
        "systemLabel": {
          "type": "string",
          "description": "SystemLabel is for describing the mci in a keyword (any string can be used) for special System purpose",
          "example": ""
        },
        "vm": {
          "type": "array",
          "items": {
            "type": "object",
            "properties": {
              "commonImage": {
                "type": "string",
                "description": "CommonImage is field for id of a image in common namespace",
                "example": "ubuntu18.04"
              },
              "commonSpec": {
                "type": "string",
                "description": "CommonSpec is field for id of a spec in common namespace",
                "example": "aws+ap-northeast-2+t2.small"
              },
              "connectionName": {
                "type": "string",
                "description": "if ConnectionName is given, the VM tries to use associtated credential.\nif not, it will use predefined ConnectionName in Spec objects"
              },
              "description": {
                "type": "string",
                "example": "Description"
              },
              "label": {
                "type": "object",
                "description": "Label is for describing the object by keywords"
              },
              "name": {
                "type": "string",
                "description": "VM name or subGroup name if is (not empty) && (> 0). If it is a group, actual VM name will be generated with -N postfix.",
                "example": "g1-1"
              },
              "rootDiskSize": {
                "type": "string",
                "description": "\"default\", Integer (GB): [\"50\", ..., \"1000\"]",
                "default": "default",
                "example": "default, 30, 42, ..."
              },
              "rootDiskType": {
                "type": "string",
                "description": "\"\", \"default\", \"TYPE1\", AWS: [\"standard\", \"gp2\", \"gp3\"], Azure: [\"PremiumSSD\", \"StandardSSD\", \"StandardHDD\"], GCP: [\"pd-standard\", \"pd-balanced\", \"pd-ssd\", \"pd-extreme\"], ALIBABA: [\"cloud_efficiency\", \"cloud\", \"cloud_essd\"], TENCENT: [\"CLOUD_PREMIUM\", \"CLOUD_SSD\"]",
                "default": "default",
                "example": "default, TYPE1, ..."
              },
              "subGroupSize": {
                "type": "string",
                "description": "if subGroupSize is (not empty) && (> 0), subGroup will be generated. VMs will be created accordingly.",
                "default": "1",
                "example": "3"
              },
              "vmUserPassword": {
                "type": "string"
              }
            }
          }
        }
      }
    },
    "path_params": {
      "required": [
        "nsId"
      ],
      "properties": {
        "nsId": {
          "type": "string",
          "description": "Namespace ID",
          "default": "default"
        }
      }
    },
    "query_params": {
      "properties": {
        "option": {
          "type": "string",
          "description": "Option for MCI creation",
          "enum": [
            "hold"
          ]
        }
      }
    }
  },
  "created_at": "2024-11-01T14:16:21.541043563+09:00",
  "updated_at": "2024-11-01T17:03:37.326385637+09:00",
  "is_example": true
}

SMTP

1. Add SMTP info

file path : /_airflow/docker-compose.yml

modify docker-compose.yml file and enter your smtp info.

gmail example : https://support.google.com/a/answer/176600?hl=en

...
    airflow-server:
        environment:
            AIRFLOW__SMTP__SMTP_HOST: 'smtp.gmail.com'
            AIRFLOW__SMTP__SMTP_USER: '[email protected]'
            AIRFLOW__SMTP__SMTP_PASSWORD: 'wtknvaprkkwyaurd'
            AIRFLOW__SMTP__SMTP_PORT: 587
            AIRFLOW__SMTP__SMTP_MAIL_FROM: '[email protected]'
...

2. Modify mail.py

file path : /_airflow/airflow-home/dags/mail.py

Modify the recipient's email address in the email_task.

...
    email_task = EmailOperator(
        task_id='send_email',
        to='Your [email protected]',
        subject='DAG 상태 보고서',
        ...
    )
...

3. Add taskComponent

Add trigger_email task component at the bottom of the workflow to receive email alarms.

...
         {
           "name": "trigger_email",
           "task_component": "trigger_email",
           "request_body": "",
           "path_params": {},
           "dependencies": [
             "{$Pre_taskName}"
           ]
         }

...

GET task log

1. GET workflow RunId

[GET] /workflow/{wfId}/runs image

2. GET taskId and task_Try_Num

[GET] /workflow/{wfId}/workflowRun/{wfRunId}/taskInstances image

3. GET execution task log

[GET] /workflow/{wfId}/workflowRun/{wfRunId}/task/{taskId}/taskTryNum/{taskTyNum}/logs image

Health-check

Check if CM-Cicada is running

curl http://127.0.0.1:8083/cicada/readyz

# Output if it's running successfully
# {"message":"CM-Cicada API server is ready"}

Check out all APIs