apache/iceberg · error · UncheckedInterruptedException

Interrupted in SQL command

Error message

Interrupted in SQL command

What it means

execute() runs SQL through Tasks, which can be interrupted while waiting/retrying; on InterruptedException the catalog wraps it in UncheckedInterruptedException with "Interrupted in SQL command". The calling thread's interrupt flag was set while the catalog metadata write was in flight.

Source

Thrown at core/src/main/java/org/apache/iceberg/jdbc/JdbcCatalog.java:782

  }

  private int execute(Consumer<SQLException> sqlErrorHandler, String sql, String... args) {
    try {
      return connections.run(
          conn -> {
            try (PreparedStatement preparedStatement = conn.prepareStatement(sql)) {
              for (int pos = 0; pos < args.length; pos += 1) {
                preparedStatement.setString(pos + 1, args[pos]);
              }

              return preparedStatement.executeUpdate();
            }
          });
    } catch (SQLException e) {
      sqlErrorHandler.accept(e);
      throw new UncheckedSQLException(e, "Failed to execute: %s", sql);
    } catch (InterruptedException e) {
      throw new UncheckedInterruptedException(e, "Interrupted in SQL command");
    }
  }

  @FunctionalInterface
  interface RowProducer<R> {
    R apply(ResultSet result) throws SQLException;
  }

  @SuppressWarnings("checkstyle:NestedTryDepth")
  private <R> List<R> fetch(RowProducer<R> toRow, String sql, String... args) {
    try {
      return connections.run(
          conn -> {
            List<R> result = Lists.newArrayList();

            try (PreparedStatement preparedStatement = conn.prepareStatement(sql)) {
              for (int pos = 0; pos < args.length; pos += 1) {
                preparedStatement.setString(pos + 1, args[pos]);

View on GitHub (pinned to 86d9c8fc54)

Solutions

1. Avoid interrupting threads performing catalog commits; drain tasks before shutdown
2. Increase timeouts so watchdogs do not interrupt long SQL operations
3. Check whether the write completed despite interruption (verify catalog state) before retrying
4. Catch UncheckedInterruptedException, restore the interrupt flag, and unwind cleanly

Example fix

// before
catalog.createTable(ident, schema);
// after
try {
  catalog.createTable(ident, schema);
} catch (UncheckedInterruptedException e) {
  Thread.currentThread().interrupt();
  throw e;
}

try {
  catalog.dropTable(ident);
} catch (UncheckedInterruptedException e) {
  Thread.currentThread().interrupt();
  // verify whether the write completed before retrying
  throw e;
}

When it happens

Trigger: Thread interrupt during a JdbcCatalog metadata write (e.g., task cancellation, executor shutdown, query timeout wrapper, Ctrl-C driven shutdown while committing catalog changes).

Common situations: Application shutdown cancelling in-flight commits; Spark/Flink task cancellation; custom timeouts that interrupt worker threads; blocking JDBC driver calls interrupted by watchdog threads.

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

• Interrupted in call to initialize
• Interrupted in SQL query
• Interrupted during commit
• Interrupted during tryLock
• Failed to insert: %d of %d succeeded

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/8d019b42821435ee. Report an issue: GitHub.