Talend Cloud Migration: On-Premise to Cloud Guide

#!/bin/bash
# talend_inventory.sh - Analyze existing Talend jobs
 
REPO_PATH="/opt/talend/repository"
OUTPUT_FILE="talend_inventory_$(date +%Y%m%d).csv"
 
echo "job_name,project,components_used,db_connections,file_connections,last_modified" > $OUTPUT_FILE
 
# Find all job files
find $REPO_PATH -name "*.item" -type f | while read job_file; do
    JOB_NAME=$(basename "$job_file" .item)
    PROJECT=$(echo "$job_file" | cut -d'/' -f5)
 
    # Extract components used
    COMPONENTS=$(grep -o 'component="[^"]*"' "$job_file" | \
                 cut -d'"' -f2 | sort -u | tr '\n' '|')
 
    # Count connection types
    DB_CONNS=$(grep -c "tDB\|tOracle\|tMySQL\|tPostgres" "$job_file" || echo "0")
    FILE_CONNS=$(grep -c "tFile\|tS3\|tAzure\|tGCS" "$job_file" || echo "0")
 
    # Get last modified date
    LAST_MOD=$(stat -c %y "$job_file" | cut -d' ' -f1)
 
    echo "$JOB_NAME,$PROJECT,$COMPONENTS,$DB_CONNS,$FILE_CONNS,$LAST_MOD" >> $OUTPUT_FILE
done
 
echo "Inventory exported to $OUTPUT_FILE"

// TalendCloudCompatibilityChecker.java
// Analyze jobs for cloud migration compatibility
 
import java.io.*;
import java.util.*;
 
public class TalendCloudCompatibilityChecker {
 
    // Components NOT supported in Talend Cloud
    private static final Set<String> UNSUPPORTED_COMPONENTS = new HashSet<>(Arrays.asList(
        "tOracleRow",          // Use tDBRow instead
        "tJDBCConnection",     // Direct JDBC needs configuration
        "tFileInputRaw",       // Raw file handling
        "tSystem",             // System calls restricted
        "tGroovy",             // Limited in cloud
        "tJMSInput",           // JMS requires Remote Engine
        "tELTOracleMap"        // ELT pushdown changes
    ));
 
    // Components requiring modification
    private static final Map<String, String> MODIFICATION_REQUIRED = new HashMap<>();
    static {
        MODIFICATION_REQUIRED.put("tFileInputDelimited",
            "Path must use cloud storage or Remote Engine file system");
        MODIFICATION_REQUIRED.put("tFileOutputDelimited",
            "Output path must be cloud-accessible");
        MODIFICATION_REQUIRED.put("tDBConnection",
            "Database must be cloud-accessible (whitelist IPs)");
        MODIFICATION_REQUIRED.put("tContextLoad",
            "Context files must be in cloud repository");
        MODIFICATION_REQUIRED.put("tRunJob",
            "Child jobs must also be migrated");
    }
 
    public static void analyzeJob(String jobXmlPath) {
        List<String> issues = new ArrayList<>();
        List<String> warnings = new ArrayList<>();
 
        try (BufferedReader reader = new BufferedReader(new FileReader(jobXmlPath))) {
            String line;
            int lineNum = 0;
 
            while ((line = reader.readLine()) != null) {
                lineNum++;
 
                for (String component : UNSUPPORTED_COMPONENTS) {
                    if (line.contains("component=\"" + component + "\"")) {
                        issues.add(String.format(
                            "Line %d: Unsupported component '%s' - requires replacement",
                            lineNum, component));
                    }
                }
 
                for (Map.Entry<String, String> entry : MODIFICATION_REQUIRED.entrySet()) {
                    if (line.contains("component=\"" + entry.getKey() + "\"")) {
                        warnings.add(String.format(
                            "Line %d: Component '%s' - %s",
                            lineNum, entry.getKey(), entry.getValue()));
                    }
                }
 
                // Check for hardcoded paths
                if (line.matches(".*[A-Z]:\\\\.*") || line.matches(".*/opt/.*") ||
                    line.matches(".*/home/.*")) {
                    warnings.add(String.format(
                        "Line %d: Hardcoded file path detected - needs parameterization",
                        lineNum));
                }
 
                // Check for environment-specific configurations
                if (line.contains("localhost") || line.contains("127.0.0.1")) {
                    issues.add(String.format(
                        "Line %d: localhost reference - must use cloud-accessible endpoint",
                        lineNum));
                }
            }
        } catch (IOException e) {
            System.err.println("Error reading job file: " + e.getMessage());
        }
 
        // Output report
        System.out.println("\n=== Cloud Compatibility Report ===");
        System.out.println("Job: " + jobXmlPath);
 
        System.out.println("\nBLOCKERS (" + issues.size() + "):");
        issues.forEach(i -> System.out.println("  ❌ " + i));
 
        System.out.println("\nWARNINGS (" + warnings.size() + "):");
        warnings.forEach(w -> System.out.println("  ⚠️ " + w));
 
        System.out.println("\nMigration Effort: " +
            (issues.isEmpty() && warnings.size() < 5 ? "LOW" :
             issues.size() < 3 ? "MEDIUM" : "HIGH"));
    }
}

migration_dependency_analysis:
  infrastructure_dependencies:
    databases:
      - name: "Oracle Production"
        current_access: "Direct TCP/IP"
        cloud_requirement: "VPN/Private Link or Cloud SQL"
        action: "Configure Talend Cloud IP whitelist"
 
      - name: "SQL Server Data Warehouse"
        current_access: "Windows Authentication"
        cloud_requirement: "SQL Authentication"
        action: "Create SQL login, update connection"
 
    file_systems:
      - name: "Network Share \\\\fileserver\\data"
        current_access: "SMB/CIFS"
        cloud_requirement: "Cloud Storage or Remote Engine"
        action: "Migrate to Azure Blob/S3 or deploy Remote Engine"
 
      - name: "Local directories /opt/talend/data"
        cloud_requirement: "Cloud Storage"
        action: "Implement S3/Azure Blob components"
 
    apis:
      - name: "Internal REST APIs"
        current_access: "Internal network"
        cloud_requirement: "Public endpoint or VPN"
        action: "Expose via API Gateway or use Remote Engine"
 
  job_dependencies:
    parent_child_jobs:
      - parent: "Master_ETL_Job"
        children:
          - "Load_Customers"
          - "Load_Orders"
          - "Load_Products"
        migration_order: "Children first, then parent"
 
    shared_resources:
      - type: "Context groups"
        items: ["DEV_Context", "PROD_Context"]
        action: "Recreate in Talend Cloud workspace"
 
      - type: "Metadata connections"
        items: ["Oracle_DW", "Salesforce_API"]
        action: "Recreate connections in Cloud"
 
      - type: "Routines"
        items: ["CustomStringUtils", "DateFormatter"]
        action: "Migrate custom code to Cloud routines"

migration_strategies:
  big_bang:
    description: "Migrate all jobs at once"
    pros:
      - "Clean cutover"
      - "No dual maintenance"
      - "Faster overall timeline"
    cons:
      - "Higher risk"
      - "Requires more testing"
      - "Longer downtime"
    recommended_for:
      - "Small job portfolios (<50 jobs)"
      - "Non-critical systems"
      - "Strong testing capability"
 
  phased_migration:
    description: "Migrate in waves by priority/domain"
    pros:
      - "Lower risk per wave"
      - "Learn from early migrations"
      - "Minimal business disruption"
    cons:
      - "Longer timeline"
      - "Dual maintenance period"
      - "Complexity managing both"
    recommended_for:
      - "Large job portfolios (100+ jobs)"
      - "Critical business processes"
      - "Limited testing resources"
    wave_example:
      wave_1: "Non-critical, simple jobs (30 jobs)"
      wave_2: "Reporting jobs (40 jobs)"
      wave_3: "Core ETL jobs (50 jobs)"
      wave_4: "Real-time integrations (20 jobs)"
 
  parallel_run:
    description: "Run both environments simultaneously"
    pros:
      - "Lowest risk"
      - "Validate results side-by-side"
      - "Easy rollback"
    cons:
      - "Double resource cost"
      - "Data sync challenges"
      - "Longest timeline"
    recommended_for:
      - "Mission-critical systems"
      - "Regulatory requirements"
      - "Zero tolerance for errors"

talend_cloud_architecture:
  cloud_components:
    tmc:  # Talend Management Console
      url: "https://tmc.{region}.cloud.talend.com"
      capabilities:
        - "Job scheduling"
        - "Monitoring dashboards"
        - "Log management"
        - "Engine management"
 
    cloud_studio:
      deployment: "Web-based designer"
      capabilities:
        - "Pipeline design"
        - "Connection management"
        - "Git integration"
        - "Collaboration features"
 
    execution_engines:
      cloud_engine:
        type: "Fully managed"
        use_case: "Cloud-to-cloud integrations"
        limitations:
          - "Cannot access on-premise resources directly"
          - "Limited to cloud-accessible endpoints"
 
      remote_engine:
        type: "Customer-managed"
        use_case: "Hybrid integrations"
        deployment_options:
          - "On-premise VM"
          - "Cloud VM (AWS/Azure/GCP)"
          - "Kubernetes"
        features:
          - "Access on-premise databases"
          - "Access local file systems"
          - "Firewall-friendly (outbound only)"
 
  network_architecture:
    cloud_engine_connectivity:
      outbound_ips: "Talend-managed, published IP ranges"
      action: "Whitelist in target system firewalls"
 
    remote_engine_connectivity:
      communication: "Outbound HTTPS to Talend Cloud"
      ports: "443 only"
      no_inbound_required: true

# remote-engine-kubernetes.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: talend-remote-engine
  namespace: talend
spec:
  replicas: 2
  selector:
    matchLabels:
      app: talend-remote-engine
  template:
    metadata:
      labels:
        app: talend-remote-engine
    spec:
      containers:
      - name: remote-engine
        image: talend/remote-engine:latest
        env:
        - name: TALEND_CLOUD_ACCOUNT
          valueFrom:
            secretKeyRef:
              name: talend-secrets
              key: account-id
        - name: TALEND_CLOUD_TOKEN
          valueFrom:
            secretKeyRef:
              name: talend-secrets
              key: pairing-token
        - name: ENGINE_NAME
          value: "k8s-remote-engine-prod"
        - name: ENGINE_WORKSPACE
          value: "Production"
        - name: JAVA_OPTS
          value: "-Xms2g -Xmx8g -XX:+UseG1GC"
        resources:
          requests:
            memory: "4Gi"
            cpu: "2"
          limits:
            memory: "16Gi"
            cpu: "8"
        volumeMounts:
        - name: engine-data
          mountPath: /opt/talend/data
        - name: engine-logs
          mountPath: /opt/talend/logs
      volumes:
      - name: engine-data
        persistentVolumeClaim:
          claimName: talend-engine-data
      - name: engine-logs
        emptyDir: {}
---
apiVersion: v1
kind: Service
metadata:
  name: talend-remote-engine
  namespace: talend
spec:
  selector:
    app: talend-remote-engine
  ports:
  - port: 8989
    targetPort: 8989
    name: admin

/*
 * Common component migrations from On-Premise to Cloud
 */
 
// 1. File Components Migration
// ON-PREMISE: Local file paths
// tFileInputDelimited
//   filename: "C:\\data\\input\\customers.csv"
 
// CLOUD: Use S3/Azure/GCS components or context variables
// tS3Input (for AWS)
//   bucket: context.s3_bucket
//   key: "input/customers.csv"
// OR keep tFileInputDelimited with Remote Engine
//   filename: context.data_path + "/input/customers.csv"
 
// 2. Database Connections
// ON-PREMISE: Direct database connection
// tOracleConnection
//   host: "oracle-server.internal.com"
//   port: "1521"
//   database: "PROD"
 
// CLOUD: Use cloud-accessible endpoint
// tDBConnection (generic) or tOracleConnection
//   host: context.db_host  // Cloud SQL or VPN-accessible
//   port: context.db_port
//   database: context.db_name
// Note: Whitelist Talend Cloud IP ranges in database firewall
 
// 3. Context Loading
// ON-PREMISE: Load from files
// tContextLoad
//   filename: "/opt/talend/config/prod.properties"
 
// CLOUD: Use TMC Environment configurations
// Define variables in TMC workspace
// Access via: context.variable_name
// OR use tContextLoad with Remote Engine accessing local files
 
// 4. Job Orchestration
// ON-PREMISE: tRunJob with local jobs
// tRunJob
//   job: "child_job"
//   context: "Production"
 
// CLOUD: Same pattern, ensure child job is also in cloud
// tRunJob
//   job: "child_job"  // Must exist in same cloud workspace
//   context: "Production"
// Alternative: Use TMC task orchestration for complex workflows
 
// 5. System Commands
// ON-PREMISE: tSystem for shell commands
// tSystem
//   command: "mv /data/input/* /data/archive/"
 
// CLOUD: Avoid tSystem; use native components
// For file operations: Use tFileCopy, tFileDelete
// For complex logic: Use tJava with Java APIs
// For unavoidable system calls: Remote Engine required

// Before Migration: On-Premise Job
// Processing local files with Oracle database
 
/* Original tFileInputDelimited configuration:
 * filename = "C:\\ETL\\data\\daily_sales.csv"
 * encoding = "UTF-8"
 * header = 1
 */
 
/* Original tOracleOutput configuration:
 * host = "oracle.internal.company.com"
 * port = "1521"
 * schema = "SALES"
 * table = "DAILY_TRANSACTIONS"
 */
 
// After Migration: Cloud-Ready Job
 
/* New tS3Connection configuration:
 * accessKey = context.aws_access_key
 * secretKey = context.aws_secret_key
 * region = context.aws_region
 */
 
/* New tS3Input configuration:
 * bucket = context.s3_data_bucket
 * key = "sales/daily/" + context.processing_date + "/daily_sales.csv"
 * encoding = "UTF-8"
 */
 
/* New tDBConnection (Cloud-accessible Oracle):
 * host = context.db_host  // RDS endpoint or VPN-accessible
 * port = context.db_port
 * database = context.db_name
 * username = context.db_user
 * password = context.db_password  // Stored in TMC secrets
 */
 
// Context Variables defined in TMC:
// Environment: Production
// Variables:
//   aws_access_key: (encrypted)
//   aws_secret_key: (encrypted)
//   aws_region: us-east-1
//   s3_data_bucket: company-data-prod
//   db_host: oracle-prod.abc123.us-east-1.rds.amazonaws.com
//   db_port: 1521
//   db_name: SALES
//   db_user: etl_user
//   db_password: (encrypted)

// Migrating custom routines to Talend Cloud
 
// Original on-premise routine: CustomStringUtils.java
package routines;
 
public class CustomStringUtils {
 
    // This method works the same in Cloud
    public static String maskSSN(String ssn) {
        if (ssn == null || ssn.length() < 9) return ssn;
        return "XXX-XX-" + ssn.substring(ssn.length() - 4);
    }
 
    // This needs modification - file system access
    // ON-PREMISE version:
    public static String readConfig(String filename) {
        try {
            return new String(java.nio.file.Files.readAllBytes(
                java.nio.file.Paths.get(filename)));
        } catch (Exception e) {
            return null;
        }
    }
 
    // CLOUD version - use context variables instead
    // Remove file system access, use TMC configuration
    public static String getConfigValue(String key,
            java.util.Map<String, Object> context) {
        return context.get(key) != null ?
            context.get(key).toString() : null;
    }
 
    // This needs modification - network call
    // ON-PREMISE version with direct HTTP:
    public static String callInternalAPI(String endpoint) {
        // Direct call to internal API
        return httpGet("http://internal-api.company.com" + endpoint);
    }
 
    // CLOUD version - parameterized endpoint
    public static String callAPI(String baseUrl, String endpoint,
            String apiKey) {
        // Use context-provided base URL and API key
        java.net.HttpURLConnection conn = null;
        try {
            java.net.URL url = new java.net.URL(baseUrl + endpoint);
            conn = (java.net.HttpURLConnection) url.openConnection();
            conn.setRequestProperty("Authorization", "Bearer " + apiKey);
            // ... rest of HTTP call
            return readResponse(conn);
        } catch (Exception e) {
            throw new RuntimeException("API call failed", e);
        }
    }
}

migration_testing_strategy:
  unit_testing:
    scope: "Individual job functionality"
    approach:
      - "Compare input/output datasets"
      - "Verify transformation logic"
      - "Test error handling paths"
    tools:
      - "Talend tAssert components"
      - "Sample data comparison"
 
  integration_testing:
    scope: "End-to-end data flows"
    approach:
      - "Test connectivity to all systems"
      - "Verify job chains work"
      - "Test scheduling triggers"
    checklist:
      - "Database connections active"
      - "Cloud storage accessible"
      - "APIs responding correctly"
      - "Authentication working"
 
  performance_testing:
    scope: "Processing speed and resource usage"
    metrics:
      - "Job execution time"
      - "Memory consumption"
      - "Network throughput"
      - "CPU utilization"
    comparison:
      - "On-premise baseline vs Cloud"
      - "Acceptable variance: ±20%"
 
  data_validation:
    scope: "Data accuracy verification"
    methods:
      - "Row count comparison"
      - "Checksum validation"
      - "Sample record verification"
      - "Aggregation comparison"

// TalendCloudMigrationTest.java
// Automated comparison testing
 
import java.sql.*;
import java.util.*;
 
public class TalendCloudMigrationTest {
 
    public static void main(String[] args) throws Exception {
        String jobName = args[0];
 
        System.out.println("Testing migration for job: " + jobName);
 
        // Run on-premise job
        System.out.println("Running on-premise job...");
        JobResult onPremResult = runOnPremiseJob(jobName);
 
        // Run cloud job
        System.out.println("Running cloud job...");
        JobResult cloudResult = runCloudJob(jobName);
 
        // Compare results
        System.out.println("\n=== Migration Test Results ===");
 
        // Row count comparison
        boolean rowCountMatch = onPremResult.rowCount == cloudResult.rowCount;
        System.out.println("Row Count: " +
            (rowCountMatch ? "✅ PASS" : "❌ FAIL") +
            " (On-Prem: " + onPremResult.rowCount +
            ", Cloud: " + cloudResult.rowCount + ")");
 
        // Execution time comparison
        double timeVariance = Math.abs(
            (cloudResult.executionTimeMs - onPremResult.executionTimeMs) /
            (double) onPremResult.executionTimeMs * 100);
        boolean timeAcceptable = timeVariance <= 20; // 20% variance allowed
        System.out.println("Execution Time: " +
            (timeAcceptable ? "✅ PASS" : "⚠️ WARNING") +
            " (Variance: " + String.format("%.1f", timeVariance) + "%)");
 
        // Data checksum comparison
        boolean checksumMatch = onPremResult.dataChecksum.equals(
            cloudResult.dataChecksum);
        System.out.println("Data Checksum: " +
            (checksumMatch ? "✅ PASS" : "❌ FAIL"));
 
        // Sample data comparison
        boolean samplesMatch = compareSampleRecords(
            onPremResult.sampleRecords,
            cloudResult.sampleRecords);
        System.out.println("Sample Records: " +
            (samplesMatch ? "✅ PASS" : "❌ FAIL"));
 
        // Overall result
        boolean overallPass = rowCountMatch && checksumMatch && samplesMatch;
        System.out.println("\n=== Overall: " +
            (overallPass ? "✅ MIGRATION VALIDATED" : "❌ ISSUES FOUND") + " ===");
    }
 
    static class JobResult {
        long rowCount;
        long executionTimeMs;
        String dataChecksum;
        List<Map<String, Object>> sampleRecords;
    }
 
    // Implementation methods...
    private static JobResult runOnPremiseJob(String jobName) {
        // Call TAC API to trigger job
        // Wait for completion
        // Query result metrics
        return new JobResult();
    }
 
    private static JobResult runCloudJob(String jobName) {
        // Call TMC API to trigger job
        // Wait for completion
        // Query result metrics
        return new JobResult();
    }
 
    private static boolean compareSampleRecords(
            List<Map<String, Object>> list1,
            List<Map<String, Object>> list2) {
        // Deep comparison of sample records
        return true;
    }
}

cutover_checklist:
  pre_cutover:
    week_before:
      - "Final validation tests passed"
      - "Stakeholder sign-off obtained"
      - "Communication sent to users"
      - "Rollback plan documented"
      - "Support team briefed"
 
    day_before:
      - "Disable on-premise job schedules"
      - "Complete final data sync"
      - "Verify cloud connections active"
      - "Confirm monitoring in place"
      - "Test rollback procedure"
 
  cutover_day:
    execution_order:
      1: "Stop all on-premise scheduled jobs"
      2: "Verify no jobs running"
      3: "Enable cloud job schedules in TMC"
      4: "Trigger initial cloud job runs"
      5: "Monitor execution closely"
      6: "Validate output data"
      7: "Confirm downstream systems receiving data"
 
    validation_gates:
      gate_1:
        name: "First job execution"
        criteria: "Job completes without errors"
        action_if_fail: "Pause, investigate, fix"
 
      gate_2:
        name: "Data validation"
        criteria: "Output matches expected"
        action_if_fail: "Consider rollback"
 
      gate_3:
        name: "Performance check"
        criteria: "Within 20% of baseline"
        action_if_fail: "Tune or accept"
 
  post_cutover:
    immediate:
      - "Send success notification"
      - "Update documentation"
      - "Archive on-premise configuration"
 
    first_week:
      - "Monitor job executions daily"
      - "Address any issues promptly"
      - "Collect feedback from users"
      - "Document lessons learned"
 
    first_month:
      - "Performance baseline in cloud"
      - "Optimize job configurations"
      - "Decommission on-premise (if applicable)"
      - "Final project closure"

#!/bin/bash
# rollback_to_onprem.sh
# Emergency rollback procedure
 
echo "=== TALEND CLOUD MIGRATION ROLLBACK ==="
echo "WARNING: This will revert to on-premise execution"
read -p "Are you sure you want to proceed? (yes/no): " confirm
 
if [ "$confirm" != "yes" ]; then
    echo "Rollback cancelled"
    exit 0
fi
 
# Step 1: Disable cloud schedules
echo "Step 1: Disabling cloud job schedules..."
curl -X POST "https://tmc.us.cloud.talend.com/api/v1/jobs/disable-all" \
    -H "Authorization: Bearer $TMC_TOKEN" \
    -H "Content-Type: application/json"
 
# Step 2: Verify no cloud jobs running
echo "Step 2: Checking for running cloud jobs..."
RUNNING_JOBS=$(curl -s "https://tmc.us.cloud.talend.com/api/v1/jobs/running" \
    -H "Authorization: Bearer $TMC_TOKEN" | jq '.count')
 
if [ "$RUNNING_JOBS" -gt 0 ]; then
    echo "WARNING: $RUNNING_JOBS jobs still running. Wait or cancel them."
    exit 1
fi
 
# Step 3: Re-enable on-premise schedules
echo "Step 3: Re-enabling on-premise job schedules..."
curl -X POST "http://tac.internal.company.com:8080/api/schedules/enable-all" \
    -H "Authorization: Basic $TAC_AUTH"
 
# Step 4: Trigger critical jobs manually
echo "Step 4: Triggering critical jobs..."
for job in "Master_ETL" "Daily_Load" "Report_Generator"; do
    curl -X POST "http://tac.internal.company.com:8080/api/jobs/$job/run" \
        -H "Authorization: Basic $TAC_AUTH"
done
 
# Step 5: Notify stakeholders
echo "Step 5: Sending rollback notification..."
curl -X POST "https://slack.webhook.url" \
    -H "Content-Type: application/json" \
    -d '{"text":"⚠️ ALERT: Talend Cloud migration rolled back to on-premise"}'
 
echo "=== ROLLBACK COMPLETE ==="
echo "Please monitor on-premise jobs and investigate cloud issues"

cloud_migration_best_practices:
  planning:
    - "Assess ALL jobs before starting migration"
    - "Identify blockers early (unsupported components)"
    - "Plan for hybrid state during transition"
    - "Document all dependencies thoroughly"
 
  execution:
    - "Start with non-critical jobs"
    - "Migrate in small waves"
    - "Validate each wave before proceeding"
    - "Maintain parallel capability during migration"
 
  connectivity:
    - "Deploy Remote Engines for on-premise access"
    - "Whitelist Talend Cloud IPs proactively"
    - "Use secure connections (SSL/TLS everywhere)"
    - "Parameterize all connection details"
 
  testing:
    - "Automate comparison testing"
    - "Test with production-scale data"
    - "Validate performance benchmarks"
    - "Test rollback procedures"
 
  operations:
    - "Set up monitoring from day one"
    - "Configure alerting for failures"
    - "Document runbooks for cloud operations"
    - "Train operations team on TMC"