Spring Cloud Data Flow流部署持续卡在DEPLOYING状态求助

Let’s walk through step-by-step to diagnose why your data pipeline is stuck in the DEPLOYING state. We’ll start with core infrastructure checks, move to server logs, and then dive into application-specific issues.

1. Verify Core Infrastructure Health #

Your pipeline depends on Zookeeper and Kafka—if either is misbehaving, deployment will fail to progress.

• Check Zookeeper:
  • Confirm it’s listening on port 2181 using netstat -ano | findstr :2181 (look for a LISTENING state).
  • Review Zookeeper logs for startup success signals, like Started AdminServer on port 8080 or ZooKeeper audit is disabled. If you see connection errors or bind failures, Zookeeper didn’t start properly.
• Check Kafka Server:
  • Verify Kafka is listening on port 9092 with netstat -ano | findstr :9092.
  • Look in Kafka logs for the line started (kafka.server.KafkaServer)—this confirms it’s ready to accept connections. If you see errors about Zookeeper connection timeouts, fix Zookeeper first.

2. Validate Spring Cloud Data Flow & Skipper Server Status #

Skipper handles application deployment, and Data Flow orchestrates the stream—both need to be fully operational.

• Data Flow Server:
  • Check logs for Started DataFlowServerApplication to confirm it’s up. Look for any exceptions during startup (e.g., port conflicts, database connection issues if you’re using a persistent store).
• Skipper Server:
  • Look for Started SkipperServerApplication in logs. Then, scan for deployment-related errors:
    • Did Skipper fail to locate your application artifacts? If you’re using local JARs, ensure the file:// URI uses Windows-compatible formatting (e.g., file:///C:/projects/product-service/target/product-service-1.0.0.jar—note three slashes).
    • Are there permission errors? Skipper needs read access to your application JARs and write access to its working directories.

3. Analyze Application-Specific Logs #

Even if the stream shows DEPLOYING, the individual services may have attempted to start and failed. Check:

• Product/Discount/Courier Service Logs:
  • Look for port conflicts (e.g., Address already in use if multiple apps try to use 8080). You can override ports with deployment properties like app.product-service.server.port=8081.
  • Verify Kafka bootstrap server configuration—ensure spring.cloud.stream.kafka.binder.brokers is set to localhost:9092 (or your Kafka address) and there are no connection timeouts.
  • Check for missing dependencies or configuration errors (e.g., missing environment variables, invalid property values for product/discount logic).

4. Address Windows-Specific Quirks #

Windows has unique behavior that can break deployment:

• File Path Formatting: When using local JARs in Data Flow Shell, use forward slashes or escaped backslashes (e.g., file:///C:/my-apps/service.jar or file://C:\\my-apps\\service.jar). Avoid unescaped backslashes—they’ll be interpreted as escape characters.
• Firewall & Antivirus: Ensure Windows Firewall isn’t blocking traffic between Zookeeper, Kafka, Data Flow, Skipper, and your services. Temporarily disable it (for testing) to rule this out.
• User Permissions: Run Data Flow/Skipper Shell as Administrator if you’re seeing access denied errors for log files or application JARs.

5. Double-Check Stream Definition & Deployment #

A misconfigured stream can prevent deployment:

• Confirm your stream creation command uses valid application names and channel bindings. For example:

  stream create --name product-delivery --definition "product-service | discount-service | courier-service"
  

• Verify deployment properties align with your stream’s channel setup. Example:

  stream deploy --name product-delivery --properties "
    app.product-service.spring.cloud.stream.bindings.output.destination=products,
    app.discount-service.spring.cloud.stream.bindings.input.destination=products,
    app.discount-service.spring.cloud.stream.bindings.output.destination=discounted-products,
    app.courier-service.spring.cloud.stream.bindings.input.destination=discounted-products
  "
  

• Ensure you’re specifying correct application versions if using Maven artifacts (e.g., maven://com.example:product-service:1.0.0).

If you’ve worked through all these steps and still see the DEPLOYING state, share specific error snippets from Skipper Server or Data Flow Server logs—this will help pinpoint the exact failure point.

内容的提问来源于stack exchange，提问作者Radhakrishna