SAP Job Scheduler Architecture — Internal Technical Reference

Version 5.2 — Last updated: March 2026 — Internal Technical Reference

1. Architecture Overview

FocusPoint uses Quartz.NET with a persistent ADO Job Store (SQL Server) for SAP integration jobs. Jobs can run in two mutually exclusive locations:

Both schedulers share the same Quartz tables (prefix: SCHEDULEJOB_) in the application database but use different SCHED_NAME values to isolate their job data.

2. Operating Modes

The master toggle is SapIntegration:EnableSapJobs in App_Data/appsettings.json:

Important: Only ONE mode should be active at a time. Running both will cause duplicate message processing, race conditions, and data corruption.

3. Scheduler Instances

4. SAP Job Registry

All SAP jobs are defined in SapServiceRegistration.ScheduleSapJobs() and mapped to interval groups:

Intervals are configurable via SAPIntegrationSettings in the admin panel and applied at startup via RescheduleJobIntervalsAsync().

5. Configuration Reference

App_Data/appsettings.json

Database Settings (Admin Panel)

6. Startup Behavior

Web App Startup (NopStartup.cs, Order = 5002)

When EnableSapJobs = true:
ConfigureServices → Register SAP Job Types (DI) → Configure() → Schedule SAP Jobs on Core Scheduler → Cleanup SapWorker Rows (SQL) → Reschedule Intervals from DB

When EnableSapJobs = false:
ConfigureServices → Skip SAP Job Registration → Configure() → Remove SAP Jobs from Core Scheduler (SQL) → Return (no further setup)

Worker Startup (Program.cs)

Build Host & DI → Create SAP Scheduler → Schedule SAP Jobs → Start Scheduler → Cleanup Core Scheduler SAP Jobs (SQL) → Reschedule Intervals from DB

7. Database Cleanup Logic

Cleanup is performed via CleanupSapSchedulerEntriesAsync() using direct SQL (not Quartz API) because the target scheduler instance may not be started or even created in the current process.

SQL Deletion Order (FK Dependencies)

Quartz tables have foreign key constraints. Rows must be deleted in this order:

Note: Steps 8–11 are only executed for full scheduler cleanup (cleanupSapWorkerScheduler). Targeted SAP job removal (removeSapJobsFromCoreScheduler) only deletes steps 1–7 filtered by job name.

8. Cache Invalidation (Worker Mode)

When the Integration Worker processes SAP inbound messages that modify products, the product price cache in the web app process must be invalidated. The worker cannot clear this cache directly because it lives in a separate process.

Flow:
Worker processes items → Collects product IDs in _processedItems queue → Publishes entity events → HTTP POST to web app

The worker sends a POST request to {FirstStoreUrl}/api/ClearProductPriceCache with the list of affected product IDs as form-encoded data. The web app's ApiController.ClearProductPriceCache enqueues them into ProductCacheClearQueue.

Condition: This HTTP call only happens when EnableSapJobs = false (worker mode) and there are processed items. In web app mode, cache is cleared in-process.

9. Setup & Deployment

Scenario A: SAP Jobs in Web App (Default)

No additional setup required. SAP jobs run on the core scheduler inside the web app process.

// App_Data/appsettings.json (web app)
{
  "SapIntegration": {
    "EnableSapJobs": true    // <-- default, can be omitted
  }
}

Scenario B: SAP Jobs in Integration Worker

Step 1: Configure Web App

// App_Data/appsettings.json (web app)
{
  "SapIntegration": {
    "EnableSapJobs": false,
    "WorkerEndpointUrl": "http://localhost:5100"
  }
}

Step 2: Deploy Worker via PowerShell

# Basic deployment (reads settings from appsettings.json)
.\deploy-worker.ps1

# Deploy for a specific client
.\deploy-worker.ps1 -NodeName "IntegrationWorker-Acme" `
                    -Database "AcmeDB" `
                    -DataSource "SQL01" `
                    -HealthCheckPort 5101

# Deploy multiple workers (multi-tenant)
.\deploy-worker.ps1 -NodeName "Worker-ClientA" -Database "ClientA_DB" -HealthCheckPort 5100
.\deploy-worker.ps1 -NodeName "Worker-ClientB" -Database "ClientB_DB" -HealthCheckPort 5101

Step 3: Verify

# Check service status
sc query "FocusPoint.IntegrationWorker.IntegrationWorker-Acme"

# Health check
curl http://localhost:5101/health

# Readiness check
curl http://localhost:5101/ready

Deploy Script Parameters

10. Operational Commands

Windows Service Management

Manual CLI (without deploy script)

Manual Database Cleanup (SQL)

-- View all SAP jobs in the core scheduler
SELECT JOB_NAME, JOB_GROUP, SCHED_NAME
FROM SCHEDULEJOB_JOB_DETAILS
WHERE SCHED_NAME = 'FocusPoint.Core.Scheduler'
  AND JOB_NAME IN ('FileReaderJob','SalesOrderConfirmationJob',
      'SalesQuoteConfirmationJob','OrderUpdateJob','QuoteUpdateJob',
      'NewUserResponseJob','ShipmentAdviceJob','PaymentCaptureJob',
      'InvoiceReceiptJob','SalesDraftReceiptJob','SapCustomerJob',
      'BusinessPartnerContactJob','SalesEmployeeJob','TaxInfoJob',
      'ItemMappingJob','JsonMessageJob','SapOuboundBPJob','ItemJob',
      'InventoryJob','BPAddressesJob','SapPromotionJob',
      'ServiceContractJob','BusinessPartnerCategoryJob');

-- View all SapWorker scheduler entries
SELECT JOB_NAME, JOB_GROUP, SCHED_NAME
FROM SCHEDULEJOB_JOB_DETAILS
WHERE SCHED_NAME = 'FocusPoint.SapWorker.Scheduler';

-- View scheduler state (which nodes are connected)
SELECT * FROM SCHEDULEJOB_SCHEDULER_STATE;

Warning: Do not manually delete Quartz rows while the owning scheduler is running. Stop the service first, then delete, then restart.

11. Troubleshooting

FocusPoint Platform — SAP Job Scheduler Architecture — Internal Document — March 2026