What causes "Invalid metadata length %s" in prestodb/presto?

Run orc-tools metadata validation on the file to confirm corruption.

prestodb/presto · error · OrcCorruptionException

Invalid metadata length %s

Error message

Invalid metadata length %s

What it means

Like the footer check, the PostScript's metadataLength must be non-negative; a negative value is impossible in a valid ORC file. The library throws OrcCorruptionException because the tail bytes cannot represent a real ORC postscript.

Source

Thrown at presto-orc/src/main/java/com/facebook/presto/orc/cache/StorageOrcFileTailSource.java:117

        checkOrcVersion(orcDataSource, postScript.getVersion());
        validateWrite(writeValidation, orcDataSource, validation -> validation.getVersion().equals(postScript.getVersion()), "Unexpected version");

        int bufferSize = toIntExact(postScript.getCompressionBlockSize());

        // check compression codec is supported
        CompressionKind compressionKind = postScript.getCompression();
        validateWrite(writeValidation, orcDataSource, validation -> validation.getCompression() == compressionKind, "Unexpected compression");

        PostScript.HiveWriterVersion hiveWriterVersion = postScript.getHiveWriterVersion();

        int footerSize = toIntExact(postScript.getFooterLength());
        int metadataSize = toIntExact(postScript.getMetadataLength());

        if (footerSize < 0) {
            throw new OrcCorruptionException(orcDataSource.getId(), "Invalid footer length %s", footerSize);
        }
        if (metadataSize < 0) {
            throw new OrcCorruptionException(orcDataSource.getId(), "Invalid metadata length %s", metadataSize);
        }

        // read DWRF stripe cache only if this feature is enabled and it has meaningful data
        boolean readDwrfStripeCache = dwrfStripeCacheEnabled
                && postScript.getDwrfStripeCacheLength().isPresent()
                && postScript.getDwrfStripeCacheMode().isPresent()
                && postScript.getDwrfStripeCacheMode().get() != DwrfStripeCacheMode.NONE;
        int dwrfStripeCacheSize = 0;
        if (readDwrfStripeCache) {
            dwrfStripeCacheSize = postScript.getDwrfStripeCacheLength().getAsInt();
            checkSizes(orcDataSource, metadataSize, dwrfStripeCacheSize);
        }

        // check if extra bytes need to be read
        Slice completeFooterSlice;
        int completeFooterSize = dwrfStripeCacheSize + metadataSize + footerSize + postScriptSize + SIZE_OF_BYTE;
        if (completeFooterSize > buffer.length) {
            // allocate a new buffer large enough for the complete footer

View on GitHub (pinned to 55bb57d202)

Solutions

Run orc-tools metadata validation on the file to confirm corruption.
Restore the file from a backup or upstream source.
Enable and check HDFS/disk checksums to find the corruption source.
Verify the writing engine version is one supported by this Presto ORC reader.

Defensive patterns

Strategy: try-catch

Try / catch

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

Prevention

Enable storage checksums
Avoid in-place edits of ORC files
Validate files after write with a metadata read-back

When it happens

Trigger: getOrcFileTail decodes a PostScript whose getMetadataLength() is negative.

Common situations: Corrupted storage blocks, truncated and re-appended files, or reads of files produced by unsupported/buggy writers.

Related errors

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/7d24badde3668d53. Report an issue: GitHub.