What causes "Failed to close current writer" in apache/iceberg?

Inspect the wrapped IOException cause to identify the underlying storage failure and fix that (permissions, disk space, network).

apache/iceberg · error · UncheckedIOException

Failed to close current writer

Error message

Failed to close current writer

What it means

ClusteredWriter throws UncheckedIOException when closing the current per-partition file writer fails with an IOException. The close path is critical because writer results (file metadata) are only collected after a successful close, so a failure here means data files may be incomplete or corrupt and the write commit should not proceed.

Source

Thrown at core/src/main/java/org/apache/iceberg/io/ClusteredWriter.java:120

    }

    currentWriter.write(row);
  }

  @Override
  public void close() throws IOException {
    if (!closed) {
      closeCurrentWriter();
      this.closed = true;
    }
  }

  private void closeCurrentWriter() {
    if (currentWriter != null) {
      try {
        currentWriter.close();
      } catch (IOException e) {
        throw new UncheckedIOException("Failed to close current writer", e);
      }

      addResult(currentWriter.result());

      this.currentWriter = null;
    }
  }

  @Override
  public final R result() {
    Preconditions.checkState(closed, "Cannot get result from unclosed writer");
    return aggregatedResult();
  }
}

View on GitHub (pinned to 86d9c8fc54)

Solutions

Inspect the wrapped IOException cause to identify the underlying storage failure and fix that (permissions, disk space, network).
Retry the write job; Iceberg commits are transactional so failed partial files are discarded.
Check storage endpoint health/quotas and increase timeouts for flaky object stores.
If reproducible, reduce file sizes/roll settings to shrink the amount of data buffered per writer.

Example fix

// before
try {
  writer.write(row);
} catch (UncheckedIOException e) {
  // silently swallowed
}
// after
try {
  writer.write(row);
} catch (UncheckedIOException e) {
  LOG.error("Writer close failed, aborting task", e);
  throw e; // surface so the commit is not attempted
}

Defensive patterns

Strategy: try-catch

Validate before calling

// ensure writer inputs are valid and storage reachable before writing
try (OutputStream probe = io.newOutputFile(targetPath).create()) { probe.write(0); }

Try / catch

try {
  writer.close();
} catch (UncheckedIOException e) {
  LOG.error("Failed to close clustered writer: {}", e.getCause().getMessage(), e);
  throw e; // abort commit; retry task
}

Prevention

Monitor disk space and object-store error rates on write nodes
Keep table write retry policies enabled (Tasks.retry) for transient IO
Avoid extremely large files per writer to reduce flush duration
Inspect the cause chain — the root IOException names the real storage problem

When it happens

Trigger: Calling write() when the clustered writer rolls over to a new partition/roll interval and must close the previous writer, or calling close() at end of write, and currentWriter.close() throws IOException (e.g. underlying filesystem/IO failure, full disk, closed connection).

Common situations: Underlying object store throttling or connection resets during flush; disk full on local scratch; HDFS lease recovery issues; task preempted while writer is mid-flush.

Understand the failure class

Background: "failed to write file", "Could not save figure", "Error saving remote file" — file write failed: causes and fixes across languages and libraries — this error's family across 38 libraries.

Related errors

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/5a694741dfdc5933. Report an issue: GitHub.