How to fix ERR checksum error in Redis

Check the Redis version on both the source and destination instances:

redis-cli INFO server | grep redis_version

RDB files from newer Redis versions may not be compatible with older versions. Ensure both instances are running compatible Redis versions. If your destination is on an older version, upgrade it or obtain an RDB file created by that version.

Use the redis-check-rdb utility to diagnose RDB file corruption:

redis-check-rdb /path/to/dump.rdb

This tool verifies the RDB file structure and checksum. If it reports an error, the file is corrupted. The utility may suggest a recovery point or identify the exact byte offset where corruption occurred.

If your RDB file is corrupted, restore from a previously verified backup:

cp /backup/dump.rdb.backup /var/lib/redis/dump.rdb
chown redis:redis /var/lib/redis/dump.rdb
systemctl restart redis-server

Always keep multiple dated backups of your RDB files. Test backups periodically by loading them on a spare Redis instance to ensure they're valid before you need them.

When transferring RDB files between systems, preserve binary data integrity:

# Correct: Use cat with redirection or scp for binary files
LC_ALL=C cat dump.rdb | redis-cli -h destination-host -x restore mykey 0 REPLACE

# Or use scp directly (preserves binary data)
scp /var/lib/redis/dump.rdb user@destination:/backup/

# Incorrect: Avoid head, tail, sed, awk on binary files
head -c 1000000 dump.rdb | redis-cli -x restore ...  # Unsafe

Set LC_ALL=C when piping binary data to avoid locale-based character transformations.

When working with the DUMP and RESTORE commands in client libraries, ensure binary safety:

Node.js (ioredis):

const data = await client.dump('sourceKey');
await client.restore('targetKey', 0, data, 'REPLACE');

Python (redis-py):

data = redis_client.dump('source_key')
redis_client.restore('target_key', 0, data, replace=True)

Go:
Use the raw bytes returned by the library, don't convert to string. Most Redis clients handle this automatically, but always verify binary payloads are kept as bytes, not strings.

If you absolutely must load a file with a checksum mismatch (not recommended), you can disable verification in redis.conf:

rdbchecksum no

Add this to /etc/redis/redis.conf, then restart Redis. This will skip checksum validation on startup but is dangerous because it allows silent data corruption to pass undetected. Only use this if you have no backups and must recover partial data. Re-enable it after recovery.