Purpose <p>The high availability requirements for supernova detection on data acquisition systems are exceptionally stringent, primarily stemming from the extremely low probability, unpredictability of supernova burst events, and the irreproducible nature of their scientific data. This paper proposes a high availability scheme for distributed data acquisition systems.</p> Method <p>By collaborating with the online software, dynamic service registration and discovery have been implemented, and an enhanced state machine ensures that instances correctly resume the operational state after restart. Corresponding solutions have also been formulated to address the single points of failure identified in each module. Moreover, a novel task scheduling architecture has been proposed and implemented, which achieves distributed task scheduling through state externalization and self-scheduling mechanisms.</p> Result <p>This scheme significantly ensures service continuity even in the event of a single-point failure, thereby meeting the stringent requirements for uninterrupted monitoring of rare supernova bursts.</p> Conclusion <p>This study explores a promising scheme that could potentially provide crucial technical support for the long-term stable operation of the supernova detection experiment.</p>