摘要：【中文】本发明涉及COW快照技术领域，且公开了一种用于数据存储及数据灾备的COW快照技术，包括以下步骤，使用连续快照系统对逻辑存储单元进行连续快照，实现对存储服务器逻辑存储单元的持续数据保护，将快照信息存储在数据库中，加快数据恢复时，根据时间点查找相关快照的速度，构建WEB管理模块，提供易用的管理接口，通过WEB管理模块创建个持续数据保护任务。该用于数据存储及数据灾备的COW快照技术，通过快照获取磁盘或者逻辑存储单元的一个副本，通过连续快照可以获取磁盘连续多的副本，从而实现数据的连续保护，通过回滚快照将逻辑存储单元恢复到之前的一个状态，或者通过克隆快照产生一个新的逻辑磁盘，供用户挂载使用。 【EN】The invention relates to the technical field of COW snapshot, and discloses a COW snapshot technology for data storage and data disaster recovery. According to the COW snapshot technology for data storage and data disaster recovery, one copy of a disk or a logic storage unit is obtained through snapshot, and multiple continuous copies of the disk can be obtained through continuous snapshot, so that continuous protection of data is achieved, the logic storage unit is restored to a previous state through rollback snapshot, or a new logic disk is generated through clone snapshot and is used for being mounted by a user.

摘要：【中文】本发明涉及块索引技术领域，且公开了一种用于灾备客户端的btree块索引技术，其特征在于，包括以下部分：将有序的索引列的值和rowid存进Oracle的各个数据块中，形成索引块，存在内存中，这些数据块以树结构的形式组织起来，父节点只记录子节点的键值位置信息，不存具体数据。该用于灾备客户端的btree块索引技术，通过将有序的索引列的值和rowid存进Oracle的各个数据块中，形成索引块，存在内存中，这些数据块以树结构的形式组织起来，父节点只记录子节点的键值位置信息，不存具体数据，查询的时候，根据树的结构去查询，本技术定位高效、利用率高且自我平衡，特别适用于高基数字段，定位单条或小范围数据非常高效，适用于灾备客户端。 【EN】The invention relates to the technical field of block indexing, and discloses a btree block indexing technology for a disaster recovery client, which is characterized by comprising the following parts: and storing the values of the ordered index columns and the rowid into each data block of the Oracle to form index blocks, storing the index blocks in a memory, organizing the data blocks in a tree structure, and only recording the key value position information of the child node by the parent node without storing specific data. According to the btree block indexing technology for the disaster recovery client, values of ordered index columns and rowids are stored in data blocks of Oracle to form index blocks, the index blocks are stored in a memory, the data blocks are organized in a tree structure mode, a father node only records key value position information of child nodes, specific data are not stored, and query is carried out according to the tree structure during query.