apache/cassandra · error · IOException

Corrupted file: integrity check (%s) failed for %s: %d != %d

Error message

Corrupted file: integrity check (%s) failed for %s: %d != %d

What it means

FileSegmentInputStream/ChecksummedDataInput's FileMetadataValidator (bytes variant) compares a stored checksum read from the checksum file against a CRC/Adler32 computed over the supplied byte range. On mismatch it throws this IOException, indicating the data bytes are corrupt relative to the sidecar checksum file.

Source

Thrown at src/java/org/apache/cassandra/io/util/DataIntegrityMetadata.java:67

        public void seek(long offset)
        {
            long start = chunkStart(offset);
            reader.seek(((start / chunkSize) * 4L) + 4); // 8 byte checksum per chunk + 4 byte header/chunkLength
        }

        public long chunkStart(long offset)
        {
            long startChunk = offset / chunkSize;
            return startChunk * chunkSize;
        }

        public void validate(byte[] bytes, int start, int end) throws IOException
        {
            int calculatedValue = (int) checksumType.of(bytes, start, end);
            int storedValue = reader.readInt();
            if (calculatedValue != storedValue)
                throw new IOException(String.format("Corrupted file: integrity check (%s) failed for %s: %d != %d", checksumType.name(), reader.getPath(), storedValue, calculatedValue));
        }

        /**
         * validates the checksum with the bytes from the specified buffer.
         *
         * Upon return, the buffer's position will
         * be updated to its limit; its limit will not have been changed.
         */
        public void validate(ByteBuffer buffer) throws IOException
        {
            int calculatedValue = (int) checksumType.of(buffer);
            int storedValue = reader.readInt();
            if (calculatedValue != storedValue)
                throw new IOException(String.format("Corrupted file: integrity check (%s) failed for %s: %d != %d", checksumType.name(), reader.getPath(), storedValue, calculatedValue));
        }

        public void close()
        {

View on GitHub (pinned to 88fd0f6a0e)

Solutions

  1. Run scrub/rebuild on the affected SSTable (nodetool scrub) or restore the component from a replica/backup.
  2. Ensure the data file and its checksum sidecar belong to the same SSTable generation — re-copy the whole SSTable set together.
  3. Check disk health (SMART/fsck); replace failing hardware.
  4. If corruption is reproducible on read, stop using the disk and rebuild the node.

Example fix

// before
validator.validate(bytes, 0, bytes.length);
// after
try {
    validator.validate(bytes, 0, bytes.length);
} catch (IOException e) {
    logger.error("Integrity check failed for {}", reader.getPath(), e);
    markSSTableCorrupted(sstable);
    throw e;
}
Defensive patterns

Strategy: validation

Validate before calling

// ensure data and checksum sidecar exist and belong to the same generation before validating
if (!checksumFileExists(dataFile)) throw new IOException("No checksum sidecar for " + dataFile);

Try / catch

try { validator.validate(bytes, start, end); }
catch (IOException e) { quarantine(sstable); throw e; }

Prevention

When it happens

Trigger: Calling validate(bytes, start, end) during SSTable component reads when the data component bytes differ from the checksum recorded in the -Checksum.db / digest sidecar file — bit rot, partial write, or mismatched data/checksum file pairs.

Common situations: Bootstrapping/repair reading SSTables whose data file was corrupted on disk; copying SSTable components out of order so data and checksum files come from different generations; failed disk or silent filesystem corruption.

Understand the failure class

Background: Checksum mismatch errors: "checksum verification failed", "digest mismatch", "expected vs actual checksum" — what they mean and how to fix them — this error's family across 41 libraries.

Related errors

AI-assisted analysis of apache/cassandra@88fd0f6a0e (2026-09-10). Data as JSON: /api/errors/2e04526aefa54702. Report an issue: GitHub.