apache/seatunnel · error · SeaTunnelException

Failed to get connection, interrupted while doing another at

Error message

Failed to get connection, interrupted while doing another attempt

What it means

JdbcConnectionFactory.connect retries acquiring a JDBC connection from the DataSource up to connectRetryTimes. If a retry attempt fails with SQLException and the intervening Thread.sleep is interrupted (InterruptedException), it wraps and rethrows as SeaTunnelException with this message, abandoning further retries.

Source

Thrown at seatunnel-connectors-v2/connector-cdc/connector-cdc-base/src/main/java/org/apache/seatunnel/connectors/cdc/base/relational/connection/JdbcConnectionFactory.java:71

                new ConnectionPoolId(
                        sourceConfig.getHostname(),
                        sourceConfig.getPort(),
                        sourceConfig.getUsername());

        HikariDataSource dataSource =
                JdbcConnectionPools.getInstance(jdbcConnectionPoolFactory)
                        .getOrCreateConnectionPool(connectionPoolId, sourceConfig);

        int i = 0;
        while (i < connectRetryTimes) {
            try {
                return dataSource.getConnection();
            } catch (SQLException e) {
                if (i < connectRetryTimes - 1) {
                    try {
                        Thread.sleep(300);
                    } catch (InterruptedException ie) {
                        throw new SeaTunnelException(
                                "Failed to get connection, interrupted while doing another attempt",
                                ie);
                    }
                    LOG.warn("Get connection failed, retry times {}", i + 1);
                } else {
                    LOG.error("Get connection failed after retry {} times", i + 1);
                    throw new SeaTunnelException(e);
                }
            }
            i++;
        }
        return dataSource.getConnection();
    }
}

View on GitHub (pinned to cf67b549a7)

Solutions

Fix the underlying database connectivity problem so retries do not linger (network, firewall, max_connections, credentials)
Re-run the job once the database is reachable; the interrupt itself was likely an intentional shutdown
Avoid cancelling the job repeatedly while connection acquisition is in flight; increase connect.max-retries so attempts succeed sooner
If interruptions happen without user cancellation, check for overly aggressive cluster shutdown/timeout settings

Example fix

// before
// job cancelled while DB unreachable -> interrupted during retry sleep
url = "jdbc:mysql://localhost:3306"  // DB down
// after
// fix DB reachability first
url = "jdbc:mysql://db-host:3306"  // reachable; connect succeeds on first attempt

Defensive patterns

Strategy: retry

Validate before calling

// verify reachability before submitting
nc -vz $DB_HOST $DB_PORT

Try / catch

try {
    connection = jdbcConnectionFactory.connect();
} catch (SeaTunnelException e) {
    if (e.getCause() instanceof InterruptedException) {
        // job was interrupted during retry; treat as shutdown, don't auto-retry here
    }
}

Prevention

Avoid cancelling jobs while connection retries are in flight
Resolve DB connectivity before job submission so retry sleeps are never reached
Monitor job shutdown behavior to prevent spurious interrupts

When it happens

Trigger: A SQLException occurs on a non-final retry, then the worker thread sleeping 300ms between attempts receives an interrupt (task cancel/shutdown, Thread.interrupt from engine stop).

Common situations: Zeta engine job cancellation or node shutdown while a CDC source is stuck retrying a dead/overloaded database connection; operator kills the job during retry backoff.

Understand the failure class

Background: Request timed out: what client-side request timeouts mean across libraries (Request timed out, TIMED_OUT, APITimeoutError) — this error's family across 39 libraries.

Related errors

AI-assisted analysis of apache/seatunnel@cf67b549a7 (2026-09-10). Data as JSON: /api/errors/f56dce0f90fad044. Report an issue: GitHub.