Mysql to ScyllaDB Migrator and validator

[vigneshkumarak]

Add MySQL as a migration and validation source

Summary

Adds support for migrating data from MySQL to ScyllaDB and validating the migrated data. This enables using the scylla-migrator for MySQL-to-ScyllaDB migrations with the same Spark-based workflow already supported for Cassandra and DynamoDB sources.

Migration

• New MySQL source type that reads tables via Spark JDBC with configurable partitioned parallel reads, cursor-based fetching, and optional WHERE filtering
• MySQLSchemaMapper handles MySQL-specific type conversions (e.g., TINYINT(1) / BIT(1) to Boolean, unsigned BIGINT to Decimal)
• Wired into Migrator to write to ScyllaDB via the existing Spark ScyllaDB connector

Validation

• MySQLToScyllaValidator performs row-by-row comparison between MySQL source and ScyllaDB target by joining on the configured primary key
• Supports hash-based comparison via a new hashColumns config option: large text/blob columns are replaced by an MD5 hash computed server-side in MySQL and client-side in Spark for ScyllaDB, dramatically reducing network data transfer during validation
• Comparison supports configurable tolerances for floats, timestamps, and BigDecimal values
• Validation summary now reports a breakdown by failure type (e.g., "1 missing target row(s), 1 differing field value(s)")

Configuration

• New SourceSettings.MySQL with fields: host, port, database, table, credentials, primaryKey, partitionColumn, numPartitions, lowerBound, upperBound, fetchSize, where, connectionProperties
• New optional hashColumns field in Validation config (backward-compatible — defaults to None, existing configs are unaffected)
• Added mysql-connector-j 8.3.0 dependency

Test configs

• mysql-to-scylla-basic.yaml — basic MySQL migration with validation
• mysql-to-scylla-large-blobs.yaml — large BLOB migration with hash-based validation