Amazon Redshift 佈建叢集入門

#!/bin/bash

# Amazon Redshift Provisioned Cluster Tutorial Script
# This script creates a Redshift cluster, loads sample data, runs queries, and cleans up resources
# Version 3: Fixed IAM role usage in COPY commands

# Set up logging
LOG_FILE="redshift_tutorial.log"
exec > >(tee -a "$LOG_FILE") 2>&1

echo "Starting Amazon Redshift tutorial script at $(date)"
echo "All commands and outputs will be logged to $LOG_FILE"

# Function to handle errors
handle_error() {
    echo "ERROR: $1"
    echo "Resources created so far:"
    if [ -n "$CLUSTER_ID" ]; then echo "- Redshift Cluster: $CLUSTER_ID"; fi
    if [ -n "$ROLE_NAME" ]; then echo "- IAM Role: $ROLE_NAME"; fi
    
    echo "Attempting to clean up resources..."
    cleanup_resources
    exit 1
}

# Function to clean up resources
cleanup_resources() {
    echo "Cleaning up resources..."
    
    # Delete the cluster if it exists
    if [ -n "$CLUSTER_ID" ]; then
        echo "Deleting Redshift cluster: $CLUSTER_ID"
        aws redshift delete-cluster --cluster-identifier "$CLUSTER_ID" --skip-final-cluster-snapshot
        echo "Waiting for cluster deletion to complete..."
        aws redshift wait cluster-deleted --cluster-identifier "$CLUSTER_ID"
        echo "Cluster deleted successfully."
    fi
    
    # Delete the IAM role if it exists
    if [ -n "$ROLE_NAME" ]; then
        echo "Removing IAM role policy..."
        aws iam delete-role-policy --role-name "$ROLE_NAME" --policy-name RedshiftS3Access || echo "Failed to delete role policy"
        
        echo "Deleting IAM role: $ROLE_NAME"
        aws iam delete-role --role-name "$ROLE_NAME" || echo "Failed to delete role"
    fi
    
    echo "Cleanup completed."
}

# Function to wait for SQL statement to complete
wait_for_statement() {
    local statement_id=$1
    local max_attempts=30
    local attempt=1
    local status=""
    
    echo "Waiting for statement $statement_id to complete..."
    
    while [ $attempt -le $max_attempts ]; do
        status=$(aws redshift-data describe-statement --id "$statement_id" --query 'Status' --output text)
        
        if [ "$status" == "FINISHED" ]; then
            echo "Statement completed successfully."
            return 0
        elif [ "$status" == "FAILED" ]; then
            local error=$(aws redshift-data describe-statement --id "$statement_id" --query 'Error' --output text)
            echo "Statement failed with error: $error"
            return 1
        elif [ "$status" == "ABORTED" ]; then
            echo "Statement was aborted."
            return 1
        fi
        
        echo "Statement status: $status. Waiting... (Attempt $attempt/$max_attempts)"
        sleep 10
        ((attempt++))
    done
    
    echo "Timed out waiting for statement to complete."
    return 1
}

# Function to check if IAM role is attached to cluster
check_role_attached() {
    local role_arn=$1
    local max_attempts=10
    local attempt=1
    
    echo "Checking if IAM role is attached to the cluster..."
    
    while [ $attempt -le $max_attempts ]; do
        local status=$(aws redshift describe-clusters \
            --cluster-identifier "$CLUSTER_ID" \
            --query "Clusters[0].IamRoles[?IamRoleArn=='$role_arn'].ApplyStatus" \
            --output text)
        
        if [ "$status" == "in-sync" ]; then
            echo "IAM role is successfully attached to the cluster."
            return 0
        fi
        
        echo "IAM role status: $status. Waiting... (Attempt $attempt/$max_attempts)"
        sleep 30
        ((attempt++))
    done
    
    echo "Timed out waiting for IAM role to be attached."
    return 1
}

# Variables to track created resources
CLUSTER_ID="examplecluster"
ROLE_NAME="RedshiftS3Role-$(date +%s)"
DB_NAME="dev"
DB_USER="awsuser"
DB_PASSWORD="Changeit1"  # In production, use AWS Secrets Manager to generate and store passwords

echo "=== Step 1: Creating Amazon Redshift Cluster ==="

# Create the Redshift cluster
echo "Creating Redshift cluster: $CLUSTER_ID"
CLUSTER_RESULT=$(aws redshift create-cluster \
  --cluster-identifier "$CLUSTER_ID" \
  --node-type ra3.4xlarge \
  --number-of-nodes 2 \
  --master-username "$DB_USER" \
  --master-user-password "$DB_PASSWORD" \
  --db-name "$DB_NAME" \
  --port 5439 2>&1)

# Check for errors
if echo "$CLUSTER_RESULT" | grep -i "error"; then
    handle_error "Failed to create Redshift cluster: $CLUSTER_RESULT"
fi

echo "$CLUSTER_RESULT"
echo "Waiting for cluster to become available..."

# Wait for the cluster to be available
aws redshift wait cluster-available --cluster-identifier "$CLUSTER_ID" || handle_error "Timeout waiting for cluster to become available"

# Get cluster status to confirm
CLUSTER_STATUS=$(aws redshift describe-clusters \
  --cluster-identifier "$CLUSTER_ID" \
  --query 'Clusters[0].ClusterStatus' \
  --output text)

echo "Cluster status: $CLUSTER_STATUS"

echo "=== Step 2: Creating IAM Role for S3 Access ==="

# Create trust policy file
echo "Creating trust policy for Redshift"
cat > redshift-trust-policy.json << EOF
{
  "Version":"2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Principal": {
        "Service": "redshift.amazonaws.com"
      },
      "Action": "sts:AssumeRole"
    }
  ]
}
EOF

# Create IAM role
echo "Creating IAM role: $ROLE_NAME"
ROLE_RESULT=$(aws iam create-role \
  --role-name "$ROLE_NAME" \
  --assume-role-policy-document file://redshift-trust-policy.json 2>&1)

# Check for errors
if echo "$ROLE_RESULT" | grep -i "error"; then
    handle_error "Failed to create IAM role: $ROLE_RESULT"
fi

echo "$ROLE_RESULT"

# Get the role ARN
ROLE_ARN=$(aws iam get-role --role-name "$ROLE_NAME" --query 'Role.Arn' --output text)
echo "Role ARN: $ROLE_ARN"

# Create policy document for S3 access
echo "Creating S3 access policy"
cat > redshift-s3-policy.json << EOF
{
  "Version":"2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "s3:GetObject",
        "s3:ListBucket"
      ],
      "Resource": [
        "arn:aws:s3:::redshift-downloads",
        "arn:aws:s3:::redshift-downloads/*"
      ]
    }
  ]
}
EOF

# Attach policy to role
echo "Attaching S3 access policy to role"
POLICY_RESULT=$(aws iam put-role-policy \
  --role-name "$ROLE_NAME" \
  --policy-name RedshiftS3Access \
  --policy-document file://redshift-s3-policy.json 2>&1)

# Check for errors
if echo "$POLICY_RESULT" | grep -i "error"; then
    handle_error "Failed to attach policy to role: $POLICY_RESULT"
fi

echo "$POLICY_RESULT"

# Attach role to cluster
echo "Attaching IAM role to Redshift cluster"
ATTACH_ROLE_RESULT=$(aws redshift modify-cluster-iam-roles \
  --cluster-identifier "$CLUSTER_ID" \
  --add-iam-roles "$ROLE_ARN" 2>&1)

# Check for errors
if echo "$ATTACH_ROLE_RESULT" | grep -i "error"; then
    handle_error "Failed to attach role to cluster: $ATTACH_ROLE_RESULT"
fi

echo "$ATTACH_ROLE_RESULT"

# Wait for the role to be attached
echo "Waiting for IAM role to be attached to the cluster..."
if ! check_role_attached "$ROLE_ARN"; then
    handle_error "Failed to attach IAM role to cluster"
fi

echo "=== Step 3: Getting Cluster Connection Information ==="

# Get cluster endpoint
CLUSTER_INFO=$(aws redshift describe-clusters \
  --cluster-identifier "$CLUSTER_ID" \
  --query 'Clusters[0].Endpoint.{Address:Address,Port:Port}' \
  --output json)

echo "Cluster endpoint information:"
echo "$CLUSTER_INFO"

echo "=== Step 4: Creating Tables and Loading Data ==="

echo "Creating sales table"
SALES_TABLE_ID=$(aws redshift-data execute-statement \
  --cluster-identifier "$CLUSTER_ID" \
  --database "$DB_NAME" \
  --db-user "$DB_USER" \
  --sql "DROP TABLE IF EXISTS sales; CREATE TABLE sales(salesid integer not null, listid integer not null distkey, sellerid integer not null, buyerid integer not null, eventid integer not null, dateid smallint not null sortkey, qtysold smallint not null, pricepaid decimal(8,2), commission decimal(8,2), saletime timestamp);" \
  --query 'Id' --output text)

echo "Sales table creation statement ID: $SALES_TABLE_ID"

# Wait for statement to complete
if ! wait_for_statement "$SALES_TABLE_ID"; then
    handle_error "Failed to create sales table"
fi

echo "Creating date table"
DATE_TABLE_ID=$(aws redshift-data execute-statement \
  --cluster-identifier "$CLUSTER_ID" \
  --database "$DB_NAME" \
  --db-user "$DB_USER" \
  --sql "DROP TABLE IF EXISTS date; CREATE TABLE date(dateid smallint not null distkey sortkey, caldate date not null, day character(3) not null, week smallint not null, month character(5) not null, qtr character(5) not null, year smallint not null, holiday boolean default('N'));" \
  --query 'Id' --output text)

echo "Date table creation statement ID: $DATE_TABLE_ID"

# Wait for statement to complete
if ! wait_for_statement "$DATE_TABLE_ID"; then
    handle_error "Failed to create date table"
fi

echo "Loading data into sales table"
SALES_LOAD_ID=$(aws redshift-data execute-statement \
  --cluster-identifier "$CLUSTER_ID" \
  --database "$DB_NAME" \
  --db-user "$DB_USER" \
  --sql "COPY sales FROM 's3://redshift-downloads/tickit/sales_tab.txt' DELIMITER '\t' TIMEFORMAT 'MM/DD/YYYY HH:MI:SS' REGION 'us-east-1' IAM_ROLE '$ROLE_ARN';" \
  --query 'Id' --output text)

echo "Sales data load statement ID: $SALES_LOAD_ID"

# Wait for statement to complete
if ! wait_for_statement "$SALES_LOAD_ID"; then
    handle_error "Failed to load data into sales table"
fi

echo "Loading data into date table"
DATE_LOAD_ID=$(aws redshift-data execute-statement \
  --cluster-identifier "$CLUSTER_ID" \
  --database "$DB_NAME" \
  --db-user "$DB_USER" \
  --sql "COPY date FROM 's3://redshift-downloads/tickit/date2008_pipe.txt' DELIMITER '|' REGION 'us-east-1' IAM_ROLE '$ROLE_ARN';" \
  --query 'Id' --output text)

echo "Date data load statement ID: $DATE_LOAD_ID"

# Wait for statement to complete
if ! wait_for_statement "$DATE_LOAD_ID"; then
    handle_error "Failed to load data into date table"
fi

echo "=== Step 5: Running Example Queries ==="

echo "Running query: Get definition for the sales table"
QUERY1_ID=$(aws redshift-data execute-statement \
  --cluster-identifier "$CLUSTER_ID" \
  --database "$DB_NAME" \
  --db-user "$DB_USER" \
  --sql "SELECT * FROM pg_table_def WHERE tablename = 'sales';" \
  --query 'Id' --output text)

echo "Query 1 statement ID: $QUERY1_ID"

# Wait for statement to complete
if ! wait_for_statement "$QUERY1_ID"; then
    handle_error "Query 1 failed"
fi

# Get and display results
echo "Query 1 results (first 10 rows):"
aws redshift-data get-statement-result --id "$QUERY1_ID" --max-items 10

echo "Running query: Find total sales on a given calendar date"
QUERY2_ID=$(aws redshift-data execute-statement \
  --cluster-identifier "$CLUSTER_ID" \
  --database "$DB_NAME" \
  --db-user "$DB_USER" \
  --sql "SELECT sum(qtysold) FROM sales, date WHERE sales.dateid = date.dateid AND caldate = '2008-01-05';" \
  --query 'Id' --output text)

echo "Query 2 statement ID: $QUERY2_ID"

# Wait for statement to complete
if ! wait_for_statement "$QUERY2_ID"; then
    handle_error "Query 2 failed"
fi

# Get and display results
echo "Query 2 results:"
aws redshift-data get-statement-result --id "$QUERY2_ID"

echo "=== Tutorial Complete ==="
echo "The following resources were created:"
echo "- Redshift Cluster: $CLUSTER_ID"
echo "- IAM Role: $ROLE_NAME"

echo ""
echo "==========================================="
echo "CLEANUP CONFIRMATION"
echo "==========================================="
echo "Do you want to clean up all created resources? (y/n): "
read -r CLEANUP_CHOICE

if [[ "$CLEANUP_CHOICE" =~ ^[Yy] ]]; then
    cleanup_resources
    echo "All resources have been cleaned up."
else
    echo "Resources were not cleaned up. You can manually delete them later."
    echo "To avoid incurring charges, remember to delete the following resources:"
    echo "- Redshift Cluster: $CLUSTER_ID"
    echo "- IAM Role: $ROLE_NAME"
fi

echo "Script completed at $(date)"