prestodb/presto · error · PrestoException

HIVE_WRITER_CLOSE_ERROR

HIVE_WRITER_CLOSE_ERROR

Error message

Error rolling back write to Hive

What it means

HivePageSink.doAbort attempts to roll back every open writer when a write is aborted. If any writer's rollback itself throws, the sink wraps that exception in HIVE_WRITER_CLOSE_ERROR with this message, preserving the original as cause. It signals failure during cleanup of a failed write, not the original write failure.

Source

Thrown at presto-hive/src/main/java/com/facebook/presto/hive/HivePageSink.java:311

    }

    private void doAbort()
    {
        Optional<Exception> rollbackException = Optional.empty();
        for (HiveWriter writer : writers) {
            // writers can contain nulls if an exception is thrown when doAppend expends the writer list
            if (writer != null) {
                try {
                    writer.rollback();
                }
                catch (Exception e) {
                    log.warn("exception '%s' while rollback on %s", e, writer);
                    rollbackException = Optional.of(e);
                }
            }
        }
        if (rollbackException.isPresent()) {
            throw new PrestoException(HIVE_WRITER_CLOSE_ERROR, "Error rolling back write to Hive", rollbackException.get());
        }
    }

    @Override
    public CompletableFuture<?> appendPage(Page page)
    {
        if (page.getPositionCount() > 0) {
            // Must be wrapped in doAs entirely
            // Implicit FileSystem initializations are possible in HiveRecordWriter#addRow or #createWriter
            hdfsEnvironment.doAs(session.getUser(), () -> doAppend(page));
        }

        return NOT_BLOCKED;
    }

    private void doAppend(Page page)
    {
        while (page.getPositionCount() > MAX_PAGE_POSITIONS) {

View on GitHub (pinned to 55bb57d202)

Solutions

1. Inspect the cause exception attached to this PrestoException; it contains the real rollback failure.
2. Check storage connectivity/permissions and free space on staging directories.
3. Retry the query; abort-time rollback failures usually leave no committed data since the write was already aborted.
4. If reproducible on a specific format/compression, upgrade the Hive connector or disable that writer type.

Example fix

// The thrown exception carries the root cause; surface it
try { ... } catch (PrestoException e) {
  log.error("rollback failed: %s", e.getCause()); // cause is the writer's rollback exception
}

try {
    // sink operation / query
} catch (PrestoException e) {
    if ("HIVE_WRITER_CLOSE_ERROR".equals(e.getErrorCode().getName())) {
        Throwable root = e.getCause(); // real rollback failure
        log.error("Writer rollback failed: %s", root);
    }
    throw e;
}

When it happens

Trigger: Query failure or cancellation triggers doAbort(); a writer's rollback() call throws (e.g. underlying writer.close/abort raised an IOException, or the original write error left the writer in a broken state).

Common situations: Node losing connection to storage mid-write, ORC/Parquet writer throwing while aborting an in-progress file, disk full on local spill/staging paths during abort.

Related errors

• HIVE_TOO_MANY_OPEN_PARTITIONS
• GENERIC_INTERNAL_ERROR
• HIVE_INVALID_ENCRYPTION_METADATA
• HIVE_INVALID_ENCRYPTION_METADATA
• HIVE_INVALID_ENCRYPTION_METADATA

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/dbef35b86bc7ff70. Report an issue: GitHub.