Glassfish/Payara配置MySQL数据源失败及Driver与DataSource差异咨询

Hey there! I’ve worked through tons of GlassFish/Payara JDBC issues like yours, so let’s break down your questions clearly:

1. How to Get the DataSource Configuration Working? #

Your struggle with javax.sql.DataSource not working is super common—usually tied to version mismatches or incorrect setup steps. Here’s what you need to fix:

• Match MySQL Connector/J Version to Class Name
  This is the #1 mistake people make:

  • For MySQL 5.x and older Connector/J versions, use com.mysql.jdbc.jdbc2.optional.MysqlDataSource
  • For MySQL 8.x+ and Connector/J 8.x, the class name changed to com.mysql.cj.jdbc.MysqlDataSource (the old class is removed in 8.x)
    Double-check your Connector/J jar version—if you’re using a newer MySQL server, you need the 8.x connector, not the 5.x one.

• Place the Connector Jar in the Correct Directory
  Forget glassfish/lib or domain/domain1/bin/lib—the right spot for domain-specific library loading is:

  • GlassFish: glassfish/domains/domain1/lib/ext
  • Payara: payara5/glassfish/lib/ext
    This directory is scanned automatically when the domain starts, ensuring the driver class is available to the connection pool.

• Configure All Required DataSource Properties
  Don’t stop at just the resource type and class name—you need to set these mandatory properties in the connection pool:

  • serverName: Your MySQL host (e.g., localhost)
  • databaseName: The name of your target database
  • user & password: MySQL credentials
  • For MySQL 8.x, add serverTimezone=UTC (or your local timezone) and useSSL=false (if you’re not using SSL for local dev) to avoid connection errors.

• Clean Cache & Perform a Full Server Restart

  • Delete the generated cache directory: glassfish/domains/domain1/generated
  • Stop the entire GlassFish/Payara server (not just restart the domain) and start it back up. Partial restarts often leave old class references in memory.

• Check the Server Log for Exact Errors
  The ping error message is vague—dig into glassfish/domains/domain1/logs/server.log (or Payara’s equivalent log) for the full stack trace. It’ll tell you if it’s a class-not-found error, authentication failure, or timezone mismatch.

2. What’s the Difference Between Driver and DataSource? #

These are two distinct layers in the JDBC stack, with key differences:

• Core Purpose & Abstraction Level

  • java.sql.Driver: A low-level JDBC component responsible for establishing direct connections to the database. It’s the foundational class that implements the JDBC protocol to talk to MySQL.
  • javax.sql.DataSource: A higher-level abstraction that wraps connection management—it’s the standard way to implement connection pooling in Java EE containers. It doesn’t just create connections; it reuses them, manages connection limits, and handles idle connection cleanup.

• Usage & Performance

  • With Driver, you’d manually load the driver (Class.forName("com.mysql.jdbc.Driver")) and use DriverManager.getConnection() to get a new connection every time. This is inefficient for apps with multiple requests, as creating new connections is resource-heavy.
  • With DataSource, the container (GlassFish/Payara) manages a pool of pre-created connections. You fetch connections via JNDI lookup, and when you close the connection, it’s returned to the pool instead of being destroyed. This drastically improves performance in high-concurrency scenarios.

• Container Integration

  • GlassFish/Payara’s built-in connection pool features (like monitoring, connection timeouts, and pool size tuning) only work fully with DataSource configurations. When you use Driver, the container may simulate pooling, but it won’t leverage the full DataSource-based pool management capabilities.

• Advanced Features

  • DataSource supports extended functionality like XA (distributed transactions) via XADataSource, custom connection validation, and integration with container-managed security. Driver has no built-in support for these features.

内容的提问来源于stack exchange，提问作者nerdlyist