Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Update the DMS Sample DAG and Docs #23681

Merged
merged 1 commit into from
May 19, 2022
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
347 changes: 347 additions & 0 deletions airflow/providers/amazon/aws/example_dags/example_dms.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,347 @@
#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
"""
Note: DMS requires you to configure specific IAM roles/permissions. For more information, see
https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Security.html#CHAP_Security.APIRole
"""

import json
import os
from datetime import datetime

import boto3
from sqlalchemy import Column, MetaData, String, Table, create_engine

from airflow import DAG
from airflow.decorators import task
from airflow.models.baseoperator import chain
from airflow.operators.python import get_current_context
from airflow.providers.amazon.aws.operators.dms import (
DmsCreateTaskOperator,
DmsDeleteTaskOperator,
DmsDescribeTasksOperator,
DmsStartTaskOperator,
DmsStopTaskOperator,
)
from airflow.providers.amazon.aws.sensors.dms import DmsTaskBaseSensor, DmsTaskCompletedSensor

S3_BUCKET = os.getenv('S3_BUCKET', 's3_bucket_name')
ROLE_ARN = os.getenv('ROLE_ARN', 'arn:aws:iam::1234567890:role/s3_target_endpoint_role')

# The project name will be used as a prefix for various entity names.
# Use either PascalCase or camelCase. While some names require kebab-case
# and others require snake_case, they all accept mixedCase strings.
PROJECT_NAME = 'DmsDemo'

# Config values for setting up the "Source" database.
RDS_ENGINE = 'postgres'
RDS_PROTOCOL = 'postgresql'
RDS_USERNAME = 'username'
# NEVER store your production password in plaintext in a DAG like this.
# Use Airflow Secrets or a secret manager for this in production.
RDS_PASSWORD = 'rds_password'

# Config values for RDS.
RDS_INSTANCE_NAME = f'{PROJECT_NAME}-instance'
RDS_DB_NAME = f'{PROJECT_NAME}_source_database'

# Config values for DMS.
DMS_REPLICATION_INSTANCE_NAME = f'{PROJECT_NAME}-replication-instance'
DMS_REPLICATION_TASK_ID = f'{PROJECT_NAME}-replication-task'
SOURCE_ENDPOINT_IDENTIFIER = f'{PROJECT_NAME}-source-endpoint'
TARGET_ENDPOINT_IDENTIFIER = f'{PROJECT_NAME}-target-endpoint'

# Sample data.
TABLE_NAME = f'{PROJECT_NAME}-table'
TABLE_HEADERS = ['apache_project', 'release_year']
SAMPLE_DATA = [
('Airflow', '2015'),
('OpenOffice', '2012'),
('Subversion', '2000'),
('NiFi', '2006'),
]
TABLE_DEFINITION = {
'TableCount': '1',
'Tables': [
{
'TableName': TABLE_NAME,
'TableColumns': [
{
'ColumnName': TABLE_HEADERS[0],
'ColumnType': 'STRING',
'ColumnNullable': 'false',
'ColumnIsPk': 'true',
},
{"ColumnName": TABLE_HEADERS[1], "ColumnType": 'STRING', "ColumnLength": "4"},
],
'TableColumnsTotal': '2',
}
],
}
TABLE_MAPPINGS = {
'rules': [
{
'rule-type': 'selection',
'rule-id': '1',
'rule-name': '1',
'object-locator': {
'schema-name': 'public',
'table-name': TABLE_NAME,
},
'rule-action': 'include',
}
]
}


def _create_rds_instance():
print('Creating RDS Instance.')

rds_client = boto3.client('rds')
rds_client.create_db_instance(
DBName=RDS_DB_NAME,
DBInstanceIdentifier=RDS_INSTANCE_NAME,
AllocatedStorage=20,
DBInstanceClass='db.t3.micro',
Engine=RDS_ENGINE,
MasterUsername=RDS_USERNAME,
MasterUserPassword=RDS_PASSWORD,
Comment on lines +112 to +123
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Don't we have operator in Rds for this?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nope, not that I can tell. There is Create/Copy/Delete Snapshot, Start/Cancel Export, and Create/Delete Event Subscription.

Copy link
Contributor

@eladkal eladkal May 18, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Example dags should be simple to comprehend this one feels a little overwhelming it has many code lines that are not really in the essence but more of the setups needed to make the point.
It feels like we should have proper operators for all the "setup" part. Consider adding the needed operators in a separate PR and then adjust this PR with updated code.
This is just an observation - if you feel comfortable with it as is... I'm happy to proceed.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We can add those new operators, but that shouldn't block documenting the existing stuff. The support code is there because we are transitioning it all over to System Tests which need to be self-sufficient. These can't just be the most basic versions anymore if they are going to be used as end to end tests, they'll need to create and destroy the required resources to actually work.

I can add the new operators to my backlog list, and make updating this part of that task, but I would rather get this merged in and update it later than shelf this until that gets done.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I suggest to create a followup issue and point also to handle the example dag here after auch operators are added

)

rds_client.get_waiter('db_instance_available').wait(DBInstanceIdentifier=RDS_INSTANCE_NAME)

response = rds_client.describe_db_instances(DBInstanceIdentifier=RDS_INSTANCE_NAME)
return response['DBInstances'][0]['Endpoint']


def _create_rds_table(rds_endpoint):
print('Creating table.')

hostname = rds_endpoint['Address']
port = rds_endpoint['Port']
rds_url = f'{RDS_PROTOCOL}://{RDS_USERNAME}:{RDS_PASSWORD}@{hostname}:{port}/{RDS_DB_NAME}'
engine = create_engine(rds_url)

table = Table(
TABLE_NAME,
MetaData(engine),
Column(TABLE_HEADERS[0], String, primary_key=True),
Column(TABLE_HEADERS[1], String),
)

with engine.connect() as connection:
# Create the Table.
table.create()
load_data = table.insert().values(SAMPLE_DATA)
connection.execute(load_data)

# Read the data back to verify everything is working.
connection.execute(table.select())


def _create_dms_replication_instance(ti, dms_client):
print('Creating replication instance.')
instance_arn = dms_client.create_replication_instance(
ReplicationInstanceIdentifier=DMS_REPLICATION_INSTANCE_NAME,
ReplicationInstanceClass='dms.t3.micro',
)['ReplicationInstance']['ReplicationInstanceArn']

ti.xcom_push(key='replication_instance_arn', value=instance_arn)
return instance_arn


def _create_dms_endpoints(ti, dms_client, rds_instance_endpoint):
print('Creating DMS source endpoint.')
source_endpoint_arn = dms_client.create_endpoint(
EndpointIdentifier=SOURCE_ENDPOINT_IDENTIFIER,
EndpointType='source',
EngineName=RDS_ENGINE,
Username=RDS_USERNAME,
Password=RDS_PASSWORD,
ServerName=rds_instance_endpoint['Address'],
Port=rds_instance_endpoint['Port'],
DatabaseName=RDS_DB_NAME,
)['Endpoint']['EndpointArn']

print('Creating DMS target endpoint.')
target_endpoint_arn = dms_client.create_endpoint(
EndpointIdentifier=TARGET_ENDPOINT_IDENTIFIER,
EndpointType='target',
EngineName='s3',
S3Settings={
'BucketName': S3_BUCKET,
'BucketFolder': PROJECT_NAME,
'ServiceAccessRoleArn': ROLE_ARN,
'ExternalTableDefinition': json.dumps(TABLE_DEFINITION),
},
)['Endpoint']['EndpointArn']

ti.xcom_push(key='source_endpoint_arn', value=source_endpoint_arn)
ti.xcom_push(key='target_endpoint_arn', value=target_endpoint_arn)


def _await_setup_assets(dms_client, instance_arn):
print("Awaiting asset provisioning.")
dms_client.get_waiter('replication_instance_available').wait(
Filters=[{'Name': 'replication-instance-arn', 'Values': [instance_arn]}]
)


def _delete_rds_instance():
print('Deleting RDS Instance.')

rds_client = boto3.client('rds')
rds_client.delete_db_instance(
DBInstanceIdentifier=RDS_INSTANCE_NAME,
SkipFinalSnapshot=True,
)

rds_client.get_waiter('db_instance_deleted').wait(DBInstanceIdentifier=RDS_INSTANCE_NAME)


def _delete_dms_assets(dms_client):
ti = get_current_context()['ti']
replication_instance_arn = ti.xcom_pull(key='replication_instance_arn')
source_arn = ti.xcom_pull(key='source_endpoint_arn')
target_arn = ti.xcom_pull(key='target_endpoint_arn')

print('Deleting DMS assets.')
dms_client.delete_replication_instance(ReplicationInstanceArn=replication_instance_arn)
dms_client.delete_endpoint(EndpointArn=source_arn)
dms_client.delete_endpoint(EndpointArn=target_arn)


def _await_all_teardowns(dms_client):
print('Awaiting tear-down.')
dms_client.get_waiter('replication_instance_deleted').wait(
Filters=[{'Name': 'replication-instance-id', 'Values': [DMS_REPLICATION_INSTANCE_NAME]}]
)

dms_client.get_waiter('endpoint_deleted').wait(
Filters=[
{
'Name': 'endpoint-id',
'Values': [SOURCE_ENDPOINT_IDENTIFIER, TARGET_ENDPOINT_IDENTIFIER],
}
]
)


@task
def set_up():
ti = get_current_context()['ti']
dms_client = boto3.client('dms')

rds_instance_endpoint = _create_rds_instance()
_create_rds_table(rds_instance_endpoint)
instance_arn = _create_dms_replication_instance(ti, dms_client)
_create_dms_endpoints(ti, dms_client, rds_instance_endpoint)
_await_setup_assets(dms_client, instance_arn)


@task(trigger_rule='all_done')
def clean_up():
dms_client = boto3.client('dms')

_delete_rds_instance()
_delete_dms_assets(dms_client)
_await_all_teardowns(dms_client)


with DAG(
dag_id='example_dms',
schedule_interval=None,
start_date=datetime(2021, 1, 1),
tags=['example'],
catchup=False,
) as dag:

# [START howto_operator_dms_create_task]
create_task = DmsCreateTaskOperator(
task_id='create_task',
replication_task_id=DMS_REPLICATION_TASK_ID,
source_endpoint_arn='{{ ti.xcom_pull(key="source_endpoint_arn") }}',
target_endpoint_arn='{{ ti.xcom_pull(key="target_endpoint_arn") }}',
replication_instance_arn='{{ ti.xcom_pull(key="replication_instance_arn") }}',
table_mappings=TABLE_MAPPINGS,
)
# [END howto_operator_dms_create_task]

# [START howto_operator_dms_start_task]
start_task = DmsStartTaskOperator(
task_id='start_task',
replication_task_arn=create_task.output,
)
# [END howto_operator_dms_start_task]

# [START howto_operator_dms_describe_tasks]
describe_tasks = DmsDescribeTasksOperator(
task_id='describe_tasks',
describe_tasks_kwargs={
'Filters': [
{
'Name': 'replication-instance-arn',
'Values': ['{{ ti.xcom_pull(key="replication_instance_arn") }}'],
}
]
},
do_xcom_push=False,
)
# [END howto_operator_dms_describe_tasks]

await_task_start = DmsTaskBaseSensor(
task_id='await_task_start',
replication_task_arn=create_task.output,
target_statuses=['running'],
termination_statuses=['stopped', 'deleting', 'failed'],
)

# [START howto_operator_dms_stop_task]
stop_task = DmsStopTaskOperator(
task_id='stop_task',
replication_task_arn=create_task.output,
)
# [END howto_operator_dms_stop_task]

# TaskCompletedSensor actually waits until task reaches the "Stopped" state, so it will work here.
# [START howto_operator_dms_task_completed_sensor]
await_task_stop = DmsTaskCompletedSensor(
task_id='await_task_stop',
replication_task_arn=create_task.output,
)
# [END howto_operator_dms_task_completed_sensor]

# [START howto_operator_dms_delete_task]
delete_task = DmsDeleteTaskOperator(
task_id='delete_task',
replication_task_arn=create_task.output,
trigger_rule='all_done',
)
# [END howto_operator_dms_delete_task]

chain(
set_up()
>> create_task
>> start_task
>> describe_tasks
>> await_task_start
>> stop_task
>> await_task_stop
>> delete_task
>> clean_up()
)
Loading