Skip to content

Connector support

This matrix is generated from the maintained capability source and checked against the Java manager registry. Roles describe whether a manager can be a source or sink; source and sink modes are listed separately so a readable source mode is never presented as a sink guarantee. Connector pages add connection, driver, query, job-count, type, staging, and security details.

ReplicaDB connector capability matrix
ConnectorSource modesSink modesStagingCaveat
MySQLcompletecomplete-atomicincrementalcompletecomplete-atomicincrementalSupported where the mode requires itUses the MariaDB JDBC driver and supports LOAD DATA style optimizations where available.
MariaDBcompletecomplete-atomicincrementalcompletecomplete-atomicincrementalSupported where the mode requires itUse MariaDB-compatible connection properties and validate server-specific SQL behavior before production runs.
PostgreSQLcompletecomplete-atomicincrementalcompletecomplete-atomicincrementalSupported where the mode requires itBinary COPY is optimized for compatible shapes; complex values may use the text path.
HSQLDBcompletecompleteNot providedThe maintained generic path is complete-mode and single-job only.
Oraclecompletecomplete-atomicincrementalcompletecomplete-atomicincrementalSupported where the mode requires itFlashback and partition behavior depend on Oracle permissions and the selected query shape.
SQL Server and Azure SQLcompletecomplete-atomicincrementalcompletecomplete-atomicincrementalSupported where the mode requires itMicrosoft Entra interactive authentication requires jobs=1 and is not a headless container flow.
CUBRIDcompletecompleteNot providedThe maintained generic path is complete-mode and single-job only.
jTDS SQL ServercompletecompleteNot providedThis legacy driver path is complete-mode and single-job only; prefer the Microsoft driver for new SQL Server work.
IBM Db2completecomplete-atomicincrementalcompletecomplete-atomicincrementalSupported where the mode requires itPartitioning and type behavior differ between Db2 LUW and other Db2 environments.
IBM i Db2completecomplete-atomicincrementalcompletecomplete-atomicincrementalSupported where the mode requires itUse IBM i naming and permissions appropriate to the AS/400 JDBC driver.
NetezzacompletecompleteNot providedThe maintained generic path is complete-mode and single-job only.
Denodocompletecomplete-atomicincrementalNot supportedNot providedDenodo is source-only; it does not implement sink writes or staging.
KafkaNot supportedcompleteNot providedKafka is sink-only and the maintained capability is complete-mode row publishing.
Amazon S3Not supportedcompleteNot providedS3 is sink-only and uses the runtime credential chain; it does not provide table merge semantics.
CSV and local filescompletecomplete-atomicincrementalcompleteincrementalSupported where the mode requires itFile replication is single-job only and has no complete-atomic sink mode.
SQLitecompletecomplete-atomicincrementalcompleteincrementalSupported where the mode requires itSQLite does not support the complete-atomic sink capability and is constrained by file locking.
MongoDBcompletecomplete-atomicincrementalcompleteincrementalSupported where the mode requires itMongoDB sink mode has no complete-atomic capability; document-to-row mapping is connector-specific.
MongoDB SRVcompletecomplete-atomicincrementalcompleteincrementalSupported where the mode requires itSRV discovery and TLS settings are supplied by the MongoDB connection string and deployment environment.