alibaba/druid · error · UnsupportedOperationException

Not supported by HighAvailableDataSource.

Error message

Not supported by HighAvailableDataSource.

What it means

HighAvailableDataSource.getConnection(String username, String password) is explicitly unimplemented — it always throws UnsupportedOperationException. The HA datasource routes through a selector that picks one of several underlying DruidDataSources and only supports the no-arg getConnection(), because per-call credential routing is not part of its design. Use the no-arg form with credentials configured on the underlying data sources.

Source

Thrown at core/src/main/java/com/alibaba/druid/pool/ha/HighAvailableDataSource.java:234

    public boolean isInBlackList(String name) {
        return blacklist.contains(name);
    }

    public void setSelector(String name) {
        DataSourceSelector selector = DataSourceSelectorFactory.getSelector(name, this);
        if (selector != null) {
            selector.init();
            setDataSourceSelector(selector);
        }
    }

    public String getSelector() {
        return selector == null ? null : selector.getName();
    }

    @Override
    public Connection getConnection(String username, String password) throws SQLException {
        throw new UnsupportedOperationException("Not supported by HighAvailableDataSource.");
    }

    @Override
    public void setLoginTimeout(int seconds) {
        DriverManager.setLoginTimeout(seconds);
    }

    @Override
    public int getLoginTimeout() {
        return DriverManager.getLoginTimeout();
    }

    public Logger getParentLogger() throws SQLFeatureNotSupportedException {
        throw new SQLFeatureNotSupportedException();
    }

    public void setConnectionProperties(String connectionProperties) {
        this.connectionProperties = connectionProperties;

View on GitHub (pinned to fa8dc99126)

Solutions

1. Use the no-arg getConnection() and set username/password on each underlying DruidDataSource in the HA group.
2. If per-user connections are required, do not use HighAvailableDataSource — use a routing/abstract datasource that supports credential parameters, or obtain a specific underlying DataSource explicitly.
3. Configure the framework (e.g. Spring) to call getConnection() rather than getConnection(username, password).

Example fix

// before
Connection c = haDataSource.getConnection("alice", "secret"); // throws

// after
// configure credentials on each underlying source once:
//   ds1.setUsername("alice"); ds1.setPassword("secret");
Connection c = haDataSource.getConnection();

// Do not call the two-arg form on HA data sources:
if (ds instanceof HighAvailableDataSource) {
    throw new IllegalArgumentException("use getConnection() and set credentials on the underlying sources");
}

public static boolean supportsCredentialedGetConnection(DataSource ds) {
    return !(ds instanceof HighAvailableDataSource);
}

try {
    conn = ds.getConnection(user, pass);
} catch (UnsupportedOperationException e) {
    if (e.getMessage().contains("HighAvailableDataSource")) {
        conn = ds.getConnection(); // credentials already on underlying sources
    } else throw e;
}

When it happens

Trigger: Calling highAvailableDataSource.getConnection(user, pass) directly, or wiring HighAvailableDataSource into a framework that invokes the two-argument getConnection (e.g. some connection providers, JdbcTemplate with explicit credentials, or user/password supplied in code).

Common situations: Migrating from a plain DruidDataSource (which accepts user/password per call) to HighAvailableDataSource without removing the per-call credentials; container-managed auth that resolves a subject and calls getConnection(user,pwd).

Related errors

• Not supported by DruidDataSource
• not support
• statement is closed
• No Observer(such as PoolUpdater) specified.
• ZK Client is NULL, Please set the zkConnectString.

AI-assisted analysis of alibaba/druid@fa8dc99126 (2026-08-14). Data as JSON: /api/errors/38a16eb9fd5ea80b. Report an issue: GitHub.