Apache Software Foundation / SeaTunnel
Merged featureFeatureSelected contributionMerged Jul 21, 2026

Support GTID specific startup offsets for MySQL CDC

Added MySQL CDC support for GTID-aware specific startup offsets with skip controls, documentation, unit tests, and E2E coverage.

apache/seatunnel · #11123

Problem

SeaTunnel MySQL CDC supported specific startup from file and position, but production replay and cutover scenarios also need GTID-set startup plus bounded skip-events and skip-rows controls that survive startup, checkpoint, and restore paths.

Approach

Extended the connector option surface, startup config, Debezium-compatible offset materialization, split-state preservation, docs, and MySQL CDC E2E coverage so GTID metadata and skip controls flow through the real runtime startup path.

Impact and scope

  • Gives MySQL CDC users a GTID-aware resume path for precise cutover and replay workflows.
  • Keeps file and position as the loader-compatible anchor while carrying GTID and skip metadata through the same offset map.
  • Improves confidence in CDC restart behavior with focused parsing, state-preservation, comparison, and runtime coverage.

Validation

  • Added unit coverage for GTID parsing, invalid configuration rejection, skip ordering, and split-state preservation.
  • Added MySQL CDC E2E coverage proving the GTID-specific startup path reaches the runtime connector contract.
  • Maintainer approval praised the contribution quality and attention to detail after the blockers were addressed.