发明名称 Method and system of stateless data replication in a distributed database system
摘要 In one exemplary embodiment, a method includes the step of executing, with at least one processor, a main database node server in a source-database node of a source-database cluster. The main database node server manages database server operations. The main database node server communicates a source-database cluster state and a specified server operation metadata to a data-shipping daemon. The specified server operation metadata, includes an information about server operation occurring on the source-database node. Another step includes executing a data-shipping daemon in the source-database node. The data-shipping daemon receives the specified server operation metadata and stores the specified server operation metadata in a digest log. The data log includes a file of the specified server operation metadata. The data-shipping daemon reads the digest log and communicate data referred to in the specified server operation metadata in the digest log to a destination database node in a destination database cluster.
申请公布号 US9514208(B2) 申请公布日期 2016.12.06
申请号 US201314024181 申请日期 2013.09.11
申请人 Srinivasan Vekatachary;Sayyaparaju Sunil;Iyer Rajkumar 发明人 Srinivasan Vekatachary;Sayyaparaju Sunil;Iyer Rajkumar
分类号 G06F17/30 主分类号 G06F17/30
代理机构 代理人
主权项 1. A computerized distributed database system comprising: a main database node server executing, with at least one processor, in a source-database node of a source-database duster, wherein the main database node server manages database server operations, and wherein the main database node server communicates a source-database cluster state and a specified server operation metadata to a data-shipping daemon, wherein the specified server operation metadata comprises an information about server operation occurring, on the source-database node; a data-shipping daemon executing in the source-database node, wherein the data-shipping daemon comprises, and wherein the data-shipping daemon further comprises: a failure-handling module that resumes an operation of the data-shipping daemon at a substantially current point of the digest log in the event of a recovery of the source-database node from a failed state, and that takes responsibility of shipping of data for other source-database cluster nodes which are not currently present in the source-database cluster, wherein the digest log comprises a ring-buffer,wherein the ring-buffer comprises a most recent specified server operation metadata, and wherein the digest logger overwrites the oldest specified server operation metadata when the ring buffer is full, andwherein the data-shipping daemon further comprises: an event logger for receiving a database event information from the destination database node over a named pipe and for storing the database event information in an event log;an event processor for reading the event logs and for messaging an event-queue manager to execute the asynchronous database trigger functions, andan event-queue manager or managing an event queue, wherein the event queue, comprises a list of asynchronous database trigger functions, and wherein the event-queue manager calls an asynchronous database trigger function based on an order of the event queue; a digest logger receiving the specified server operation metadata and storing the specified server operation metadata in a digest log, and wherein the data log comprises a file of the specified server operation metadata; and a data shipper module reading the digest log and communicate data referred to in the specified server operation metadata in the digest log to a destination database node in a destination database cluster.
地址 Sunnyvale CA US