In which replication mode is the primary system blocked until the secondary system is available?

In full sync replication mode, the primary system is designed to ensure data consistency with the secondary system by waiting for an acknowledgment that the data has been successfully written to the secondary before it commits any changes. This means that operations on the primary system are blocked until the secondary system becomes available to accept and confirm the data, enforcing a strict consistency model.

Full sync replication is particularly useful in scenarios where data loss is unacceptable, and immediate consistency across systems is necessary. The primary system prioritizes data integrity, making it crucial for applications that require real-time synchronization and cannot tolerate any latency in updates or potential data discrepancies between the two systems.

In contrast, other modes like asynchronous replication and partial sync do not impose such requirements on the primary system, allowing it to continue operations without waiting for the secondary system's acknowledgment, thereby sacrificing immediate data consistency for improved performance or availability.