prestodb/presto · error · OrcCorruptionException

Invalid ORC metadata %s or DWRF stripe cache size %s

Error message

Invalid ORC metadata %s or DWRF stripe cache size %s

What it means

DWRF files may carry an optional stripe footer cache. An ORC file's metadata section and the DWRF stripe cache are mutually exclusive tails: both being > 0 simultaneously is invalid. checkSizes throws OrcCorruptionException when both metadataSize and dwrfStripeCacheSize are positive.

Source

Thrown at presto-orc/src/main/java/com/facebook/presto/orc/cache/StorageOrcFileTailSource.java:214

                        orcDataSource,
                        Joiner.on('.').join(version),
                        CURRENT_MAJOR_VERSION,
                        CURRENT_MINOR_VERSION);
            }
        }
    }

    /**
     * ORC metadata and DWRF stripe cache sizes are mutually exclusive because
     * only ORC files have metadata, and only DWRF files have stripe cache.
     * <p>
     * Let's check that either both sizes are 0, or only one of them is
     * greater than 0.
     */
    private static void checkSizes(OrcDataSource orcDataSource, int metadataSize, int dwrfStripeCacheSize)
    {
        if (metadataSize > 0 && dwrfStripeCacheSize > 0) {
            throw new OrcCorruptionException(orcDataSource.getId(),
                    "Invalid ORC metadata %s or DWRF stripe cache size %s",
                    metadataSize,
                    dwrfStripeCacheSize);
        }

        if (dwrfStripeCacheSize < 0) {
            throw new OrcCorruptionException(orcDataSource.getId(), "Invalid DWRF stripe cache length %s", dwrfStripeCacheSize);
        }
    }
}

View on GitHub (pinned to 55bb57d202)

Solutions

1. Regenerate the file with a DWRF writer that emits a consistent postscript.
2. Validate with orc-tools/vertex-tools metadata dump to inspect both length fields.
3. Check for storage corruption and restore from a known-good copy.
4. Open an issue with the writer team if the file came from a specific internal pipeline version.

try { tail = source.getOrcFileTail(ds, reader, validation, false, time); }
catch (OrcCorruptionException e) { reportWriterInconsistency(ds.getId(), e); }

When it happens

Trigger: getOrcFileTail reads a DWRF postscript where both dwrfStripeCacheLength and metadataLength are greater than zero — an inconsistent combination from a corrupt or non-conforming file.

Common situations: Files produced by DWRF writers with mismatched/buggy postscript fields, or corruption flipping a length field from 0 to a positive value.

Related errors

• Invalid DWRF stripe cache length %s
• Encrypted data size %s exceeds limit of 2^23
• Stripe encryption keys are missing, but file is encrypted
• Number of stripe encryption keys did not match number of enc
• Invalid file size %s

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/4cefd45c782da8b7. Report an issue: GitHub.