CN102156720A - Method, device and system for restoring data - Google Patents

本发明公开了一种数据恢复的方法、装置和系统，其方法为：获取所有用户操作请求，将用户操作请求以操作日志形式存储于恢复日志；提取恢复日志中的读取操作记录并进行分析，得到对目标容灾节点数据库进行更新的操作语句；接收并合并属于同一事务的操作语句，以事务为单位进行集合，获取可提交的事务包并发送；顺序接收事务包，并装载事务包至目标容灾节点数据库中完成数据恢复。通过上述本发明的方法不仅解决了数据在异构数据库容灾节点之间的数据恢复的一致性和完整性问题，而且还减轻了日志传输对网络造成的负担，以及避免执行无效的查询操作，使事务包装载效率更高，进一步提高了目标容灾节点数据库的数据恢复速度。

The invention discloses a data recovery method, device and system, the method comprising: obtaining all user operation requests, storing the user operation requests in the recovery log in the form of an operation log; extracting and analyzing the read operation records in the recovery log , to obtain the operation statement for updating the database of the target disaster recovery node; receive and merge the operation statements belonging to the same transaction, collect them in transaction units, obtain and send the transaction package that can be submitted; receive the transaction package sequentially, and load the transaction package to Complete data recovery in the target disaster recovery node database. The above-mentioned method of the present invention not only solves the problem of consistency and integrity of data recovery between heterogeneous database disaster recovery nodes, but also reduces the burden on the network caused by log transmission and avoids executing invalid query operations, The loading efficiency of the transaction package is higher, and the data recovery speed of the target disaster recovery node database is further improved.

一种数据恢复的方法、装置和系统A method, device and system for data recovery

技术领域technical field

本发明涉及数据库领域，更具体的说，是涉及一种异构数据库的容灾系统中基于操作日志重做的数据恢复的方法、装置和系统。The present invention relates to the database field, and more specifically, relates to a method, device and system for data recovery based on operation log redo in a heterogeneous database disaster recovery system.

背景技术Background technique

如图1所示，在现有技术中基于异构数据库的容灾系统的结构，由分布在网络中的若干个数据处理中心(容灾节点)，容灾管理系统中间件和应用客户端组成。容灾系统是指为计算机信息系统提供的一个能应付各种灾难的环境。在计算机系统遭受不可抗拒的自然灾难，以及计算机犯罪、病毒和人为操作错误等人为灾难时，容灾系统将保证用户数据的安全性(数据容灾)和不间断的提供应用服务(应用容灾)。异构数据库系统则是相关的多个数据库系统的集合。As shown in Figure 1, the structure of the disaster recovery system based on heterogeneous databases in the prior art consists of several data processing centers (disaster recovery nodes) distributed in the network, disaster recovery management system middleware and application clients . The disaster recovery system refers to an environment that can cope with various disasters provided for the computer information system. When the computer system suffers from irresistible natural disasters, as well as man-made disasters such as computer crimes, viruses, and human errors, the disaster recovery system will ensure the security of user data (data disaster recovery) and provide uninterrupted application services (application disaster recovery) ). A heterogeneous database system is a collection of related multiple database systems.

基于异构数据库的容灾系统以中间件的形式为应用系统提供服务，中间件系统屏蔽数据处理中心(图中未标示)的存在为用户提供一个统一的访问接口，而基于异构数据库的容灾系统将现有数据中心组织为一个整体，因此各个容灾节点之间对等，并互为灾备的关系。The disaster recovery system based on heterogeneous databases provides services for application systems in the form of middleware. The middleware system shields the existence of the data processing center (not shown in the figure) to provide users with a unified The disaster system organizes the existing data centers as a whole, so each disaster recovery node is equal and has a mutual disaster recovery relationship.

在基于异构数据库的容灾系统中，当图1中的某些容灾节点发生故障或遭受灾难时，依据容灾要求，正常的容灾节点仍然继续处理应用系统的数据更新请求。但是，当发生灾难的容灾节点硬件更换或者故障排除之后，其数据与其他正常的容灾节点会处于不一致的状态，因此，需要对发生故障或灾难的容灾节点进行数据一致性恢复。In the disaster recovery system based on heterogeneous databases, when some disaster recovery nodes in Figure 1 fail or suffer disasters, according to the disaster recovery requirements, the normal disaster recovery nodes still continue to process the data update requests of the application system. However, when the disaster recovery node hardware is replaced or the fault is eliminated, its data will be in an inconsistent state with other normal disaster recovery nodes. Therefore, it is necessary to perform data consistency recovery on the failure or disaster recovery node.

现有数据容灾通常采用数据复制的方法，在异地建立一个备份数据系统，该备份数据系统在本地的数据系统崩溃时迅速接替本地数据系统提供数据服务。目前主流的方法包括基于存储层的数据复制、基于逻辑卷的数据复制和基于重做日志(redo log)的数据复制。Existing data disaster recovery usually adopts the method of data replication, and a backup data system is established in a different place, and the backup data system quickly takes over the local data system to provide data services when the local data system crashes. The current mainstream methods include storage layer-based data replication, logical volume-based data replication, and redo log-based data replication.

其中，基于存储层的数据复制：通过基于SAN(Storage Area Network，存储局域网)进行数据复制，该复制过程针对每个I/O(输入输出设备)进行，它将主系统中对磁盘的I/O操作通过网络传输到目标端存储系统中再执行一遍。但是，基于该方法进行复制的缺点为：只支持主、备两端为同一品牌和具有同样配置的存储设备，不能整合不同品牌的存储设备；不支持不同数据库产品之间的数据复制；需要昂贵的高端磁盘阵列和光纤连接设备，实施费用也较高。Among them, data replication based on the storage layer: through data replication based on SAN (Storage Area Network, storage area network), the replication process is carried out for each I/O (input and output device), and it combines the I/O of the disk in the main system The O operation is transmitted to the target storage system through the network and executed again. However, the disadvantages of replication based on this method are: it only supports storage devices with the same brand and configuration at the primary and backup ends, and cannot integrate storage devices of different brands; it does not support data replication between different database products; it requires expensive High-end disk arrays and fiber optic connection equipment, the implementation cost is also higher.

基于逻辑卷的数据复制：通过基于TCP/IP的网络环境进行数据复制，由操作系统进程捕捉逻辑卷的变化进行数据复制。它将主系统中对磁盘逻辑卷的修改操作复制到目标端系统重新执行一遍。但是，基于该方法进行复制的缺点为：需要在存储设备与数据库之间建立一个虚拟存储层(逻辑卷管理层)，会对主机性能有一定影响；要求两端的数据库系统必须同构，这使得在整合不同数据库产品的场合受到限制；需要传输的数据量比较大，占用的网络资源也比较多。Data replication based on logical volumes: Data replication is performed through a TCP/IP-based network environment, and the operating system process captures changes in logical volumes for data replication. It copies the modification operation of the disk logical volume in the primary system to the target system and executes it again. However, the disadvantage of duplicating based on this method is: a virtual storage layer (logical volume management layer) needs to be established between the storage device and the database, which will have a certain impact on the performance of the host computer; the database systems at both ends must be isomorphic, which makes It is limited in the occasion of integrating different database products; the amount of data to be transmitted is relatively large, and the network resources occupied are relatively large.

数据库系统的基于redo log的数据复制：使用数据库以外的独立进程，捕捉重做日志文件(redo log file)的信息，将其翻译成SQL语句，再通过网络传输到目标端数据库，在目标端数据库执行同样的SQL。如果其进程赶不上数据库日志切换，也可以捕捉归档日志中的内容。但是，该方法仍然存在缺点：其一，数据库的所有更改都记录在redo log中，其中记录了对数据库中的每一个变化，它的实施与具体数据库密切相关的，要将数据库的redo log file中的信息转换为SQL语句，需要对每种不同的数据库redo log进行具体分析，因此不利于异构容灾的实现；其二，在数据库的吞吐量太大时，数据会有较大延迟。Redo log-based data replication of the database system: Use an independent process other than the database to capture the information of the redo log file (redo log file), translate it into SQL statements, and then transmit it to the target database through the network, and then transfer it to the target database. Execute the same SQL. It can also capture the contents of the archived log if its process cannot catch up with the database log switch. However, this method still has disadvantages: First, all changes to the database are recorded in the redo log, which records every change in the database, and its implementation is closely related to the specific database. The redo log file of the database must be The information in the database is converted into SQL statements, which requires specific analysis of the redo logs of each different database, which is not conducive to the realization of heterogeneous disaster recovery; second, when the throughput of the database is too large, the data will have a large delay.

基于上述现有技术中的三种数据复制方法的分析，可知现有的数据复制技术都不能完全满足异构数据库容灾系统需求。因此，需要一种可以克服现有技术中的数据复制方法的缺点，即满足异构数据库容灾系统需求的数据一致性恢复方法。Based on the analysis of the above three data replication methods in the prior art, it can be seen that none of the existing data replication technologies can fully meet the requirements of the heterogeneous database disaster recovery system. Therefore, there is a need for a data consistency recovery method that can overcome the shortcomings of the data replication method in the prior art, that is, meet the requirements of a heterogeneous database disaster recovery system.

发明内容Contents of the invention

有鉴于此，本发明提供了一种数据恢复的方法、装置和系统，以克服现有技术中各种数据复制方法不能满足异构数据库容灾系统需求的问题。In view of this, the present invention provides a data recovery method, device and system to overcome the problem that various data replication methods in the prior art cannot meet the requirements of heterogeneous database disaster recovery systems.

为实现上述目的，本发明提供如下技术方案：To achieve the above object, the present invention provides the following technical solutions:

一种数据恢复方法，包括：A data recovery method comprising:

获取所有用户操作请求，将所述用户操作请求以操作日志形式存储于恢复日志；Obtain all user operation requests, and store the user operation requests in the recovery log in the form of operation logs;

提取所述恢复日志中的读取操作记录并进行分析，得到对目标容灾节点数据库进行更新的操作语句；Extracting and analyzing the read operation records in the recovery log to obtain an operation statement for updating the database of the target disaster recovery node;

接收并合并属于同一事务的所述操作语句，以所述事务为单位进行集合，获取可提交的事务包，并发送；receiving and merging the operation statements belonging to the same transaction, collecting the transaction as a unit, obtaining a submittable transaction package, and sending it;

顺序接收所述事务包，并装载所述事务包至所述目标容灾节点数据库中，完成数据恢复。The transaction packages are received sequentially, and the transaction packages are loaded into the database of the target disaster recovery node to complete data recovery.

优选的，所述提取所述恢复日志中的读取操作记录并进行分析的过程包括：Preferably, the process of extracting and analyzing the read operation records in the recovery log includes:

逐条提取存储于所述恢复日志中的操作日志记录；extracting operation log records stored in the recovery log one by one;

依次对所述操作日志记录中表示操作记录项的结构化查询语言SQL的语句内容的sqlsrc字段进行分析；The sqlsrc field of the statement content of the structured query language SQL representing the operation record item in the operation log record is analyzed in turn;

当所述sqlsrc字段的开头与关键字select匹配时，丢弃所述sqlsrc字段对应的操作日志记录的语句；When the beginning of the sqlsrc field matches the keyword select, discard the statement recorded in the operation log corresponding to the sqlsrc field;

当所述sqlsrc字段的开头与关键字select不匹配时，确定所述sqlsrc字段对应的操作日志记录的语句为对目标容灾节点数据库进行更新的操作语句。When the beginning of the sqlsrc field does not match the keyword select, it is determined that the statement recorded in the operation log corresponding to the sqlsrc field is an operation statement for updating the database of the target disaster recovery node.

优选的，所述得到对目标容灾节点数据库进行更新的操作语句中包括：提交当前事务Commit的事务、未Commit的事务和中止当前事务Rollback的事务。Preferably, the operation statement obtained to update the database of the target disaster recovery node includes: a transaction that commits the current transaction Commit, a transaction that does not Commit, and a transaction that aborts the current transaction Rollback.

优选的，所述合并属于同一事务的所述操作语句，以所述事务为单位进行集合，获取可提交的事务包的过程包括：Preferably, the merging of the operation statements belonging to the same transaction is performed in units of the transaction, and the process of obtaining a submittable transaction package includes:

根据所述恢复日志中所属的事务序列号进行事务合并，生成对应的事务包；Performing transaction merging according to the transaction sequence number in the recovery log to generate a corresponding transaction package;

依据所述恢复日志中的操作日志序列号对每个所述事务包中的操作日志记录进行升序排序；Sort the operation log records in each transaction package in ascending order according to the operation log sequence number in the recovery log;

当所述事务包为Commit的事务，获取所述事务包；When the transaction package is a Commit transaction, obtain the transaction package;

当所述事务包为Rollback的事务，丢弃所述事务包。When the transaction packet is a Rollback transaction, the transaction packet is discarded.

优选的，所述获取可提交的事务包之后，并发送之前还包括：Preferably, after obtaining the submittable transaction package and before sending it, it also includes:

存储所述事务包于发送端的缓存中。Store the transaction packet in the cache of the sender.

优选的，所述顺序接收所述事务包，之后，装载之前还包括：Preferably, the sequence receives the transaction package, and then, before loading, also includes:

将所述顺序接收的事务包存储于接收端的事务队列中。The sequentially received transaction packets are stored in a transaction queue at the receiving end.

一种数据恢复装置，包括：A data recovery device, comprising:

容灾管理系统中间件，所述容灾管理系统中间件包括：Disaster recovery management system middleware, the disaster recovery management system middleware includes:

事务捕获模块，用于获取所有用户操作请求，将所述用户操作请求以操作日志形式存储于恢复日志；A transaction capture module, configured to obtain all user operation requests, and store the user operation requests in the recovery log in the form of an operation log;

恢复日志数据库，用于存储所述恢复日志；recovery log database, for storing the recovery log;

日志分析模块，用于提取所述恢复日志中的读取操作记录并进行分析，获取对目标容灾节点数据库进行更新的操作语句；The log analysis module is used to extract and analyze the read operation record in the recovery log, and obtain the operation statement for updating the database of the target disaster recovery node;

事务综合模块，用于接收并合并所述日志分析模块中属于同一事务的操作语句，以所述事务为单位进行集合，获取可提交的所述事务包；The transaction integration module is used to receive and merge the operation statements belonging to the same transaction in the log analysis module, collect the transaction as a unit, and obtain the transaction package that can be submitted;

日志发送模块，用于发送事务综合模块中获取的可提交的所述事务包；A log sending module, configured to send the submittable transaction package obtained in the transaction synthesis module;

日志接收模块，用于顺序接收所述事务包；A log receiving module, configured to sequentially receive the transaction packets;

日志装载模块，用于装载所述日志接收模块接收到的事务包至所述目标容灾节点数据库中，完成数据恢复。The log loading module is configured to load the transaction package received by the log receiving module into the database of the target disaster recovery node to complete data recovery.

一种数据恢复系统，包括：A data recovery system comprising:

包含一个或多个用户的用户端，用于发送用户操作请求；Contains a client of one or more users for sending user operation requests;

数据恢复装置，用于接收所述用户操作请求，将所述用户操作请求以操作日志形式进行存储，并对所述操作日志进行处理，将获取到的符合目标容灾节点数据库数据恢复的事务包装载至容灾节点数据库；A data recovery device, configured to receive the user operation request, store the user operation request in the form of an operation log, process the operation log, and package the acquired transaction that meets the target disaster recovery node database data recovery Loaded to the disaster recovery node database;

容灾节点数据库，用于依据装载的所述事务包中的操作日志进行数据恢复。The disaster recovery node database is used for data recovery according to the operation log in the loaded transaction package.

经由上述的技术方案可知，与现有技术相比，本发明公开了一种数据恢复的方法、装置和系统，通过接收用户操作请求，并以操作日志的形式存储于恢复日志数据库中，在目标容灾节点数据库恢复数据时，从恢复日志数据库中提取操作日志，经网络传输到需要恢复的容灾节点数据库中，并在这些容灾节点数据库中进行事务重做装载。通过上述过程不仅解决了数据在异构数据库容灾节点之间的恢复问题，而且通过日志分析减少了网络中的操作日志传输量，减轻日志传输对网络造成的负担，同时，通过事务综合使传输过程中以事务为单位进行传输，保证数据的一致性和完整性；以及，在目标容灾节点数据库端进行事务包装载时能够避免执行无效的查询操作，使事务包装载效率更高，进一步提高了目标容灾节点数据库的数据恢复速度。It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a method, device and system for data recovery. By receiving user operation requests and storing them in the recovery log database in the form of operation logs, the target When the disaster recovery node database restores data, the operation log is extracted from the recovery log database, and is transmitted to the disaster recovery node database to be restored through the network, and the transaction redo loading is carried out in these disaster recovery node databases. The above process not only solves the problem of data recovery between heterogeneous database disaster recovery nodes, but also reduces the amount of operation log transmission in the network through log analysis, and reduces the burden on the network caused by log transmission. In the process, the transaction is used as the unit to ensure the consistency and integrity of the data; and, when the transaction package is loaded on the database side of the target disaster recovery node, invalid query operations can be avoided, making the transaction package loading more efficient and further improving The data recovery speed of the target disaster recovery node database is improved.

附图说明Description of drawings

为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据提供的附图获得其他的附图。In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only It is an embodiment of the present invention, and those skilled in the art can also obtain other drawings according to the provided drawings without creative work.

图1为现有技术中基于异构数据库的容灾系统的结构示意图；FIG. 1 is a schematic structural diagram of a disaster recovery system based on a heterogeneous database in the prior art;

图2为本发明实施例公开的一种数据恢复的方法流程图；Fig. 2 is a flow chart of a data recovery method disclosed in an embodiment of the present invention;

图3为本发明实施例公开的捕获或获取所有用户操作请求的原理图；FIG. 3 is a schematic diagram of capturing or obtaining all user operation requests disclosed in an embodiment of the present invention;

图4为本发明实施例公开的日志分析的原理图；FIG. 4 is a schematic diagram of log analysis disclosed in an embodiment of the present invention;

图5为本发明实施例公开的日志分析的流程图；Fig. 5 is the flow chart of log analysis disclosed by the embodiment of the present invention;

图6为本发明实施例公开的实现事务综合获取事务包的原理图；FIG. 6 is a schematic diagram of realizing transaction comprehensive acquisition of a transaction package disclosed in an embodiment of the present invention;

图7为本发明实施例公开的实现事务综合获取事务包的流程图；FIG. 7 is a flow chart of realizing transaction comprehensive acquisition of a transaction package disclosed in an embodiment of the present invention;

图8为本发明实施例公开的一种数据恢复装置的结构示意图；FIG. 8 is a schematic structural diagram of a data recovery device disclosed in an embodiment of the present invention;

图9为本发明实施例公开的一种数据恢复系统的结构示意图。FIG. 9 is a schematic structural diagram of a data recovery system disclosed in an embodiment of the present invention.

具体实施方式Detailed ways

为了引用和清楚起见，下文中使用的技术名词的说明、简写或缩写总结如下：For reference and clarity, descriptions, abbreviations or abbreviations of technical terms used in the following text are summarized as follows:

事务：指访问并可能更新数据库中各种数据项的一个程序执行单元，通常由高级数据库操纵语言或编程语言(如SQL，C++或Java)书写的用户程序的执行所引起；Transaction: refers to a program execution unit that accesses and possibly updates various data items in the database, usually caused by the execution of user programs written in advanced database manipulation languages or programming languages (such as SQL, C++ or Java);

SQL：Structured Query Language，结构化查询语言；SQL: Structured Query Language, Structured Query Language;

Commit：提交当前事务；Commit: Commit the current transaction;

Rollback：中止当前事务；Rollback: abort the current transaction;

Cache：缓存；Cache: cache;

Export Process：传输进程；Export Process: transfer process;

Import Process：导入进程；Import Process: import process;

Queue：队列。Queue: Queue.

下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

由于，现有技术中的各类数据复制的方法都存在不能满足异构数据库的容灾系统的需求，即不能满足确保异构的数据库容灾节点之间的数据恢复的一致性，也不能满足提高容灾节点数据恢复效率的需求，因此，本发明提供了一种数据恢复的方法、装置和系统，基本原理是：在容灾管理系统中设置恢复日志(Recovery Log)数据库，通过该恢复日志数据库接收应用系统发送至容灾管理系统的所有事务请求，即用户操作请求，并以操作日志的形式存储到文件系统或数据库系统中。当对一个容灾节点数据库进行恢复时，从恢复日志数据库中提取操作日志，经网络传输到需要恢复的容灾节点数据库中，并在这些容灾节点数据库中进行事务重做装载，从而实现各个容灾节点数据库的数据一致性恢复。具体过程通过以下实施例进行说明。Because all kinds of data replication methods in the prior art cannot meet the requirements of disaster recovery systems for heterogeneous databases, that is, they cannot meet the requirements of ensuring the consistency of data recovery between disaster recovery nodes of heterogeneous databases, nor can they meet the requirements of disaster recovery systems for heterogeneous databases. The demand for improving the data recovery efficiency of disaster recovery nodes, therefore, the present invention provides a method, device and system for data recovery, the basic principle is: a recovery log (Recovery Log) database is set in the disaster recovery management system, through the recovery log The database receives all transaction requests sent by the application system to the disaster recovery management system, that is, user operation requests, and stores them in the file system or database system in the form of operation logs. When recovering a disaster recovery node database, the operation log is extracted from the recovery log database, and transmitted to the disaster recovery node database to be recovered through the network, and transaction redo loading is carried out in these disaster recovery node databases, so as to realize each Data consistency recovery of disaster recovery node database. The specific process is illustrated by the following examples.

请参阅附图2，为本发明实施例公开的一种数据恢复的方法流程图，主要包括以下步骤：Please refer to accompanying drawing 2, it is a flow chart of a method for data recovery disclosed by the embodiment of the present invention, mainly including the following steps:

步骤S101，获取所有用户操作请求，将所述用户操作请求以操作日志形式存储于恢复日志中。In step S101, all user operation requests are acquired, and the user operation requests are stored in a recovery log in the form of an operation log.

在本发明上述公开的实施例中，在用户和容灾节点数据库之间设置有中间层，该中间层也可称为容灾管理系统中间件。如图3所示为捕获或获取所有用户操作请求的原理图(图中包括：用户1、用户2、用户n、容灾管理系统中间件A、恢复日志数据库B和容灾节点数据库C)，所有用户操作请求由该容灾管理系统中间件A接收，并由该容灾管理系统中间件A将用户操作请求转发给所有容灾节点数据库C，使用户操作请求不再直接提交给容灾节点数据库C，同时，在该容灾管理系统中间件A将用户操作请求进行复制，并以操作日志形式存储于恢复日志数据库B中。In the above-disclosed embodiments of the present invention, an intermediate layer is set between the user and the database of the disaster recovery node, and the intermediate layer may also be called the middleware of the disaster recovery management system. As shown in Figure 3, it is a schematic diagram of capturing or obtaining all user operation requests (the figure includes: user 1, user 2, user n, disaster recovery management system middleware A, recovery log database B and disaster recovery node database C), All user operation requests are received by the disaster recovery management system middleware A, and the disaster recovery management system middleware A forwards the user operation requests to all disaster recovery node databases C, so that user operation requests are no longer directly submitted to the disaster recovery nodes Database C, at the same time, in the disaster recovery management system middleware A copies the user operation request, and stores it in the recovery log database B in the form of an operation log.

需要说明的是，以操作日志形式将用户操作请求存储于恢复日志数据库B中，具体为将该用户操作请求以操作日志记录形式保存于文件系统或者数据库系统中，但是，由于业务应用系统会产生大量日志记录，利用文件系统管理日志记录可能导致效率不高，因此可选择使用专门的数据库系统保存上述的操作日志。其中，恢复日志数据库B中的恢复日志的数据结构的格式如表1所示。It should be noted that the user operation request is stored in the recovery log database B in the form of an operation log, specifically saving the user operation request in the file system or database system in the form of an operation log record. However, due to the business application system will generate A large number of log records, using the file system to manage log records may lead to inefficiency, so you can choose to use a dedicated database system to save the above operation logs. Wherein, the format of the data structure of the recovery log in the recovery log database B is shown in Table 1.

表1：Table 1:

  字段名field name   类型 type   字段意义说明Description of field meaning   log_idlog_id   长整形long shaping   操作日志ID，根据请求提交的时间生成Operation log ID, generated according to the time when the request is submitted   vloginvlogin   字符串String   用户登录信息User login information   sqlsrcsqlsrc   文本text   该操作记录项的SQL语句内容The SQL statement content of the operation record item   sqlsrc_paramsqlsrc_param   文本text   SQL语句的参数Parameters of the SQL statement   transaction_sntransaction_sn   长整形long shaping   所属的事务序列号，根据事务提交时间生成The sequence number of the transaction to which it belongs, generated according to the transaction commit time   exec_statusexec_status   字符characters   SQL语句执行状态SQL statement execution status   update_countupdate_count   整数integer   该SQL语句执行更新的记录数The number of records updated by the SQL statement

因为用户数据在经过数据库管理系统处理之前是与具体数据库无关的，因此，捕获或截获的数据无须进行额外的数据库之间的数据格式转换。Because the user data is irrelevant to the specific database before being processed by the database management system, the captured or intercepted data does not require additional data format conversion between databases.

此外，在该容灾管理系统中间件A实现用户操作请求的获取或捕获时，用户利用通用驱动E(图3中用符号E表示)，即利用一个能够实现java.sql.Driver接口的类作为容灾管理系统中间件的通用驱动，使用户可以使用该通用驱动E代替专门的JDBC驱动D访问容灾管理系统中间件，而容灾管理系统中间件A则根据具体的数据库JDBC驱动D(在图3中用符号D表示)转发接收到的用户操作请求。In addition, when the middleware A of the disaster recovery management system implements the acquisition or capture of user operation requests, the user uses a general driver E (represented by symbol E in Figure 3), that is, uses a class that can implement the java.sql.Driver interface as the The general driver of the disaster recovery management system middleware enables users to use the general driver E instead of the special JDBC driver D to access the disaster recovery management system middleware, and the disaster recovery management system middleware A is based on the specific database JDBC driver D (in In Fig. 3, indicated by symbol D), the received user operation request is forwarded.

步骤S102，提取所述恢复日志中的读取操作记录并进行分析，获取对目标容灾节点数据库进行更新的操作语句。Step S102, extracting and analyzing the read operation records in the recovery log, and obtaining an operation statement for updating the database of the target disaster recovery node.

在步骤S102中，依据恢复日志中的读取操作记录分析出SQL语句的关联信息，该关联信息包括：操作对象、操作类型和操作数据等。并进一步区分、筛选出对目标容灾节点数据库进行更新和非更新的语句。In step S102, analyze the associated information of the SQL statement according to the read operation records in the recovery log, and the associated information includes: operation object, operation type, and operation data. And further distinguish and filter out the statements for updating and non-updating the database of the target disaster recovery node.

由于，在实际应用的过程中，读写事务都会保持一定的比例。例如，按10∶1的读写事务比例来计算，步骤S101中的恢复日志中只有1/10的记录可以对目标容灾节点数据库进行数据更新的。Because, in the process of actual application, read and write transactions will maintain a certain proportion. For example, based on the ratio of read-write transactions of 10:1, only 1/10 of the records in the recovery log in step S101 can update the data of the target disaster recovery node database.

因此，在执行步骤S102时，通过对恢复日志进行分析，将恢复日志中对目标容灾节点数据库进行更新的操作记录筛选出来。同时，在对恢复日志进行分析的过程中，丢弃查询语句，只保留关键字为Insert into、Update、Commit等对目标容灾节点数据库进行更新的SQL语句的记录。Therefore, when step S102 is executed, by analyzing the recovery log, the operation records for updating the database of the target disaster recovery node are screened out from the recovery log. At the same time, in the process of analyzing the recovery log, the query statement is discarded, and only the records of the SQL statement whose keywords are Insert into, Update, Commit, etc. to update the database of the target disaster recovery node are kept.

通过执行步骤S102对恢复日志进行分析，不仅能够减轻之后进行事务综合处理任务的负担，而且还可以减少需要传输的日志记录数量，进一步减轻日志传输时网络所承担的负担；同时，在分析的过程中丢弃了查询语句，因此，在需要恢复的容灾节点数据库端也无须执行无效的查询操作，使其加载效率更高，进而恢复数据的速度也更快。By executing step S102 to analyze the recovery log, it can not only reduce the burden of subsequent comprehensive transaction processing tasks, but also reduce the number of log records that need to be transmitted, and further reduce the burden on the network during log transmission; at the same time, in the process of analysis The query statement is discarded, so there is no need to perform invalid query operations on the database side of the disaster recovery node that needs to be restored, so that the loading efficiency is higher, and the data recovery speed is also faster.

需要说明的是，根据步骤S101中提到的恢复日志中操作记录的存储格式定义(格式如表1所示)，在执行步骤S102时判断一条语句是否对数据库中的数据产生了影响，只需要对恢复日志的sqlsrc字段内容进行分析即可。It should be noted that, according to the storage format definition of the operation record in the recovery log mentioned in step S101 (the format is shown in Table 1), when step S102 is executed to determine whether a statement has affected the data in the database, only Just analyze the contents of the sqlsrc field of the recovery log.

具体的分析过程如图4和图5所示，其中，图4为日志分析的原理图，图5为具体分析的流程图，主要包括以下步骤：The specific analysis process is shown in Figure 4 and Figure 5, where Figure 4 is a schematic diagram of log analysis, and Figure 5 is a flow chart of specific analysis, which mainly includes the following steps:

步骤S201，逐条提取存储于所述恢复日志中的操作日志记录。Step S201, extracting operation log records stored in the recovery log one by one.

步骤S202，依次对所述操作日志记录中的sqlsrc字段进行分析，判断所述sqlsrc字段的开头是否与“select”关键字匹配，如果是，则执行步骤S203；如果否，则执行步骤S204。Step S202, analyze the sqlsrc field in the operation log record sequentially, and judge whether the beginning of the sqlsrc field matches the "select" keyword, if yes, execute step S203; if not, execute step S204.

步骤S203，丢弃所述sqlsrc字段对应的操作日志记录的语句。Step S203, discarding the statement recorded in the operation log corresponding to the sqlsrc field.

步骤S204，确定所述sqlsrc字段对应的操作日志记录的语句，所述语句为对目标容灾节点数据库进行更新的操作语句。Step S204, determining the statement recorded in the operation log corresponding to the sqlsrc field, and the statement is an operation statement for updating the database of the target disaster recovery node.

通过执行上述步骤S201至步骤S204，将操作日志记录逐条从恢复日志中取出，并对日志的sqlsrc字段内容进行分析，如果该sqlsrc字段开头与“select”关键字匹配，说明该语句不会对数据库更改发生影响，将其丢弃，反之，则确定该操作日志记录的语句，该语句即为对容灾节点数据库进行更新的操作语句。By executing the above steps S201 to S204, the operation log records are taken out one by one from the recovery log, and the content of the sqlsrc field of the log is analyzed. If the beginning of the sqlsrc field matches the "select" keyword, it means that the statement will not affect the database. If the change is affected, it is discarded; otherwise, the statement recorded in the operation log is determined, and the statement is an operation statement for updating the database of the disaster recovery node.

步骤S103，接收所述操作语句，并合并属于同一事务的操作语句，以所述事务为单位进行集合，获取可提交的事务包并发送。Step S103, receiving the operation statement, merging the operation statements belonging to the same transaction, collecting the transaction as a unit, obtaining a submittable transaction package and sending it.

经过执行步骤S102分析恢复日志之后，获取到对目标容灾节点数据库进行更新的操作语句，该语句也可以统称为事务SQL语句，通过执行步骤S102获取到的事务SQL语句是交叉出现的，并且属于同一个事务的多条SQL指令是非连续存储的，而多个事务的SQL语句之间是相互穿插的；同时，获取到的事务SQL语句包含所有commit的事务以及没有commit的事务。After performing step S102 to analyze the recovery log, the operation statement for updating the database of the target disaster recovery node is obtained. This statement can also be collectively referred to as a transaction SQL statement. The transaction SQL statement obtained by performing step S102 is interleaved and belongs to Multiple SQL commands of the same transaction are stored discontinuously, and the SQL statements of multiple transactions are interspersed with each other; at the same time, the obtained transaction SQL statements include all committed transactions and transactions without commit.

为了提高系统的可控制性、保证逻辑完整性、避免数据丢失，因此，执行步骤S103将恢复日志处理的最小单位设定为一个事务，而不是单个SQL语句，通过事务综合使需要恢复数据的目标容灾节点数据库端的事务装载更加容易控制。In order to improve the controllability of the system, ensure logical integrity, and avoid data loss, the minimum unit of recovery log processing is set as a transaction instead of a single SQL statement in step S103. The transaction loading on the database side of the disaster recovery node is easier to control.

需要说明的是，只有经过Commit的数据对于目标容灾节点数据库是有意义的，而对于Rollback的数据则不需要复制到目标数据库上。It should be noted that only committed data is meaningful to the target disaster recovery node database, while Rollback data does not need to be copied to the target database.

如图6所示，为执行步骤S103实现事务综合获取对应同一事务的操作语句对应的事务包的原理图。由于接收到的对应事务记录的操作语句，即SQL语句中包括Commit的事务、没有Commit的事务和Rollback的事务三种情况，因此，首先，按照事务序号对SQL语句进行划分，每个事务包含多条SQL语句；然后，以事务为单位进行处理，获取已经Commit的事务并发送；将未提交的Commit事务保存在本地，直至通过日志得知保存的未提交的Commit事务已提交，则立即获取该Commit事务并发送；而对Rollback的事务则作丢弃处理。As shown in FIG. 6 , for executing step S103 to implement transaction synthesis, obtain a schematic diagram of a transaction package corresponding to an operation statement corresponding to the same transaction. Since the received operation statement corresponding to the transaction record, that is, the SQL statement includes three cases of Commit transaction, non-Commit transaction and Rollback transaction, therefore, firstly, the SQL statement is divided according to the transaction sequence number, and each transaction contains multiple Then, process in units of transactions, get the committed transactions and send them; save the uncommitted Commit transactions locally, until it is known from the log that the saved uncommitted Commit transactions have been committed, then immediately obtain the Commit transactions Commit the transaction and send it; and discard the Rollback transaction.

具体实现事务综合获取事务包的过程请参阅附图7，主要包括以下步骤：Please refer to Figure 7 for the specific implementation of the process of transaction integration and acquisition of transaction packages, which mainly includes the following steps:

步骤S301，根据所述恢复日志中的transaction_sn字段进行事务合并，生成对应的事务包。Step S301, perform transaction merging according to the transaction_sn field in the recovery log, and generate a corresponding transaction package.

在执行步骤S301中，确定该transaction_sn字段相同的日志记录属于同一个事务，并将它们合并到同一个事务包之中构成。In step S301, it is determined that the log records with the same transaction_sn field belong to the same transaction, and merge them into the same transaction package to form.

步骤S302，依据所述恢复日志中的log_id字段对每个所述事务包中的操作日志记录进行升序排序，确保每个所述事务包中的操作日志记录顺序与提交时的顺序相同。Step S302, sort the operation log records in each transaction package in ascending order according to the log_id field in the recovery log, to ensure that the operation log records in each transaction package are in the same order as they were submitted.

步骤S303，判断所述事务包是否为Commit事务，如果是，将执行步骤S304；如果否，则执行步骤S305。Step S303, judging whether the transaction package is a Commit transaction, if yes, execute step S304; if not, execute step S305.

步骤S304，获取所述事务包。Step S304, acquiring the transaction package.

步骤S305，丢弃所述事务包。Step S305, discarding the transaction packet.

经由上述描述可知，在执行步骤S303中对当前的事务包进行判断会获取两种结果，或为Commit事务，或为Rollback事务，通过判断后将对后续操作有用的Commit事务包进行保留，而将无用的Rollback事务包丢弃。It can be seen from the above description that in step S303, judging the current transaction package will obtain two results, either a Commit transaction or a Rollback transaction. After the judgment is passed, the Commit transaction package useful for subsequent operations will be retained, and the Useless Rollback transaction packets are discarded.

需要说明的是，由于网络环境的复杂性，网络故障、系统断电等不可预知的情况都有可能发生，为确保事务包，即日志在传输中断后能继续正确传输，保证日志的在源端和目的端的一致性，即为保证数据传输的安全、可靠，在本发明中设置了一个具有针对性的日志发送和接收机制。主要包括以下特点：It should be noted that due to the complexity of the network environment, unpredictable situations such as network failures and system power outages may occur. In order to ensure that the transaction package, that is, the log can continue to be transmitted correctly after the transmission is interrupted, ensure that the log is at the source. Consistency with the destination, that is, to ensure the safety and reliability of data transmission, a targeted log sending and receiving mechanism is set in the present invention. It mainly includes the following features:

其一，数据(事务)在传输之前首先存入发送端的Cache，Export Process从Cache中读取事务数据封装为TCP/IP数据包(事务包)传送给接收端的Import Process。First, the data (transaction) is first stored in the Cache of the sending end before transmission, and the Export Process reads the transaction data from the Cache and encapsulates it into a TCP/IP data packet (transaction packet) and sends it to the Import Process of the receiving end.

其二，在接收端，Import Process在收到传输的事务包后，首先存入事务队列中，然后由装载进程从事务队列中严格按照事务的顺序装载事务信息。Second, at the receiving end, after the Import Process receives the transmitted transaction package, it first stores it in the transaction queue, and then the loading process loads the transaction information from the transaction queue in strict accordance with the order of transactions.

依据上述传输中的特点，负责传输的Export Process从本地队列中按照先进先出的原则抓取需要传输的事务包，并将事务包封装成一个数据包后通过TCP/IP协议传递给接收端系统。According to the characteristics of the above transmission, the Export Process responsible for the transmission grabs the transaction packets to be transmitted from the local queue according to the first-in-first-out principle, encapsulates the transaction packets into a data packet, and then transmits it to the receiving end system through the TCP/IP protocol .

由于，在发送端封装的数据包的包头部分描述包的大小，因此，接收端系统在接收到传来的数据包后，首先根据包头描述的包的大小进行传输的合法性检查，判断传输过程是否完整。Since the header part of the data packet encapsulated at the sending end describes the size of the packet, after receiving the incoming data packet, the receiving end system first checks the validity of the transmission according to the size of the packet described in the header to judge the transmission process. is it complete.

此外，在日志源和目标容灾节点数据库两端都设置有磁盘缓存，事务文件按照事务顺序传输，并确定目标容灾节点数据库端将接收的事务文件写入磁盘后删除源端数据，能够确保即使出现网络中断、系统断电等故障时，数据恢复仍然保证其连续和一致性。In addition, disk caches are set at both ends of the log source and target disaster recovery node databases, transaction files are transmitted in the order of transactions, and the target disaster recovery node database is determined to write the received transaction files to the disk and then delete the source data, which can ensure Even in the event of network interruption, system power failure and other failures, data recovery still guarantees its continuity and consistency.

即通过上述步骤S103的操作，以事务为单位进行传输，而不是以SQL语句为单位进行，保证了数据恢复的一致性和完整性。That is, through the operation of the above step S103, the transmission is performed in units of transactions instead of in units of SQL statements, which ensures the consistency and integrity of data recovery.

步骤S104，顺序接收所述事务包，使所述事务包装载至所述目标容灾节点数据库中，完成数据恢复。In step S104, the transaction packets are received in sequence, and the transaction packets are loaded into the database of the target disaster recovery node to complete data recovery.

在执行步骤S104中，利用事务包对目标容灾节点数据库进行数据恢复的装载方式采用通用的SQL接口，通过执行接收到的事务包中Insert、Update、Delete等SQL语句实现数据的装载，完成对目标容灾节点数据库的数据恢复。In the execution step S104, the loading method of data recovery of the target disaster recovery node database using the transaction package adopts a general SQL interface, and the loading of data is realized by executing SQL statements such as Insert, Update, and Delete in the received transaction package, and the database is completed. Data recovery of the target disaster recovery node database.

此外，事务包在目标容灾节点数据库中的重做，即数据恢复可以和接收并行执行。在接收缓存队列接收事务包的同时，目标容灾节点数据库从缓存队列的头部取出事务包执行装载，在确定事务包成功执行装载之后再将缓存中的事务包删除，确保每个事务包被成功执行，并且只执行一次。In addition, the redo of the transaction package in the database of the target disaster recovery node, that is, the data recovery can be executed in parallel with the reception. While receiving the transaction package in the receiving cache queue, the database of the target disaster recovery node takes out the transaction package from the head of the cache queue to perform loading, and deletes the transaction package in the cache after confirming that the transaction package is successfully loaded to ensure that each transaction package is Executed successfully, and only once.

由于，在恢复目标容灾节点数据库时所执行的事务包是经过日志分析、事务综合处理后获得的事务包序列，不需要执行无效的查询操作，因此，其在目标容灾节点数据库中具有很高的加载执行效率，从而在保证数据恢复一致性的基础上，更进一步的提高了数据恢复的速度。Since the transaction package executed when restoring the database of the target disaster recovery node is a sequence of transaction packages obtained after log analysis and comprehensive transaction processing, there is no need to perform invalid query operations, so it has a lot of advantages in the database of the target disaster recovery node. High loading execution efficiency further improves the speed of data recovery on the basis of ensuring the consistency of data recovery.

因此，通过上述本发明实施例公开的数据恢复的方法，基于对操作日志的处理，当对一个容灾节点数据库进行恢复时，提取经过分析处理的操作日志，经网络传输到需要恢复的容灾节点数据库中，并在这些容灾节点数据库中进行事务重做装载，从而实现各个容灾节点的数据一致性恢复。Therefore, through the data recovery method disclosed in the above-mentioned embodiments of the present invention, based on the processing of the operation log, when restoring a disaster recovery node database, the analyzed and processed operation log is extracted and transmitted to the recovery recovery node via the network. Node databases, and transaction redo loading is performed in these disaster recovery node databases, so as to achieve data consistency recovery of each disaster recovery node.

上述本发明公开的实施例中详细描述了一种数据恢复的方法，对于本发明的方法可采用多种形式的装置实现，因此本发明还公开了一种数据恢复的装置，下面给出具体的实施例进行详细说明。A data recovery method is described in detail in the above-mentioned disclosed embodiments of the present invention. The method of the present invention can be realized by various devices. Therefore, the present invention also discloses a data recovery device. The specific details are given below Examples are described in detail.

请参阅附图8为本发明实施例公开的一种数据恢复的装置结构示意图，主要包括：容灾管理系统中间件11、日志接收模块12和日志装载模块13。Please refer to FIG. 8 , which is a schematic structural diagram of a data recovery device disclosed in an embodiment of the present invention, which mainly includes: a disaster recovery management system middleware 11 , a log receiving module 12 and a log loading module 13 .

该容灾管理系统中间件11，用于接收用户操作请求，将所述用户操作请求以操作日志形式进行存储，并对所述操作日志进行处理，获取符合目标容灾节点数据库数据恢复的事务包a后发送。The middleware 11 of the disaster recovery management system is used to receive user operation requests, store the user operation requests in the form of operation logs, process the operation logs, and obtain transaction packets that meet the target disaster recovery node database data recovery sent after a.

该容灾管理系统中间件11中具体包括：事务捕获模块110、恢复日志数据库111、日志分析模块112、事务综合模块113和日志发送模块114。The disaster recovery management system middleware 11 specifically includes: a transaction capture module 110 , a recovery log database 111 , a log analysis module 112 , a transaction synthesis module 113 and a log sending module 114 .

其中，事务捕获模块110，用于获取所有用户操作请求，将所述用户操作请求以操作日志形式存储于恢复日志中。Wherein, the transaction capture module 110 is configured to obtain all user operation requests, and store the user operation requests in the recovery log in the form of an operation log.

恢复日志数据库111，用于存储所述恢复日志。The recovery log database 111 is configured to store the recovery log.

日志分析模块112，用于提取所述恢复日志中的读取操作记录并进行分析，获取对目标容灾节点数据库进行更新的操作语句，并发送。The log analysis module 112 is configured to extract and analyze the read operation records in the recovery log, obtain and send an operation statement for updating the database of the target disaster recovery node.

事务综合模块113，用于接收所述日志分析模块112发送的所述操作语句，并合并属于同一事务的操作语句，以所述事务为单位进行集合，获取可提交的事务包a。The transaction integration module 113 is configured to receive the operation statements sent by the log analysis module 112, and combine the operation statements belonging to the same transaction, and gather the transaction as a unit to obtain a submittable transaction package a.

日志发送模块114，用于发送事务综合模块113中获取的可提交的所述事务包a。The log sending module 114 is configured to send the committable transaction package a obtained in the transaction synthesis module 113 .

日志接收模块12，用于顺序接收所述事务包a。The log receiving module 12 is configured to sequentially receive the transaction packets a.

日志装载模块13，用于装载所述日志接收模块12接收到的事务包a至所述目标容灾节点数据库中，完成数据恢复。The log loading module 13 is configured to load the transaction package a received by the log receiving module 12 into the database of the target disaster recovery node to complete data recovery.

通过上述容灾管理系统中间件11中的各个模块之间的相互协作，完成对操作日志的处理，以保证基于该处理后的操作日志完成对目标容灾节点数据库的数据恢复。本发明实施例公开的装置内各个模块之间的协作过程具体为：事务捕获模块110捕获所有客户请求，以操作日志形式存储至恢复日志数据库111中；日志分析模块112从恢复日志数据库111中的读取操作记录并进行分析，筛选出对数据库进行更新的操作语句，将分析结果交给事务综合模块113；事务综合模块113则将属于同一事务的语句进行合并，将日志记录综合为以事务为最小单位的集合，将集合后的事务包交给日志发送模块114进行发送或传输；日志发送模块114则负责将集合的事务包无丢失的，按照原顺序传输到日志接收模块12；最后，由日志装载模块13将接收的事务包装载至目标容灾节点数据库中，完成对目标容灾节点数据库的数据恢复。由于，通过日志分析减少了网络中的操作日志传输量，减轻日志传输对网络造成的负担，而利用事务综合使传输过程中以事务为单位进行传输，而不是以SOL语句为单位进行，保证数据的一致性和完整性；并且，在执行目标容灾节点数据库的事务装载过程中不需要执行无效的查询操作，使得装载效率更高，进一步加快了目标容灾节点数据库的数据恢复速度。The processing of the operation log is completed through the mutual cooperation among various modules in the disaster recovery management system middleware 11, so as to ensure that the data recovery of the target disaster recovery node database is completed based on the processed operation log. The cooperation process among the various modules in the device disclosed in the embodiment of the present invention is specifically as follows: the transaction capture module 110 captures all client requests, and stores them in the recovery log database 111 in the form of operation logs; Read and analyze the operation record, screen out the operation statement for updating the database, and deliver the analysis result to the transaction synthesis module 113; the transaction synthesis module 113 merges the statements belonging to the same transaction, and synthesizes the log records into transaction-based The collection of the smallest unit, the transaction package after the collection is handed over to the log sending module 114 for sending or transmission; the log sending module 114 is then responsible for transmitting the collected transaction package to the log receiving module 12 according to the original order without loss; finally, by The log loading module 13 loads the received transaction package into the database of the target disaster recovery node, and completes the data recovery of the database of the target disaster recovery node. Because, through log analysis, the amount of operation log transmission in the network is reduced, and the burden on the network caused by log transmission is reduced, and the use of transaction synthesis enables the transmission to be carried out in units of transactions rather than in units of SOL statements, ensuring data Consistency and integrity; and, in the process of executing the transaction loading process of the target disaster recovery node database, it is not necessary to perform invalid query operations, which makes the loading efficiency higher and further speeds up the data recovery speed of the target disaster recovery node database.

上述本发明公开的实施例中详细描述了方法和装置，对于本发明的方法和装置可应用于对应的系统中，因此本发明还公开了一种数据恢复的系统，下面给出具体的实施例进行详细说明。The method and device are described in detail in the above disclosed embodiments of the present invention. The method and device of the present invention can be applied to corresponding systems. Therefore, the present invention also discloses a data recovery system. Specific embodiments are given below Describe in detail.

请参阅附图9为本发明实施例公开的一种数据恢复的系统结构示意图，主要包括：用户端21、数据恢复装置22和容灾节点数据库23。Please refer to FIG. 9 , which is a schematic structural diagram of a data recovery system disclosed by an embodiment of the present invention, which mainly includes: a client 21 , a data recovery device 22 and a disaster recovery node database 23 .

用户端21包含有一个或多个用户，用于发送用户操作请求。The user terminal 21 includes one or more users, and is used to send user operation requests.

数据恢复装置22，用于接收用户操作请求，将所述用户操作请求以操作日志形式进行存储，并对所述操作日志进行处理，将获取到的符合目标容灾节点数据库数据恢复的事务包装载至容灾节点数据库23中。The data recovery device 22 is configured to receive a user operation request, store the user operation request in the form of an operation log, process the operation log, and load the obtained transaction package that meets the target disaster recovery node database data recovery to the disaster recovery node database 23.

容灾节点数据库23，用于依据装载的所述事务包中的操作日志进行数据恢复。The disaster recovery node database 23 is configured to perform data recovery according to the operation log in the loaded transaction package.

本发明该实施例公开的数据恢复的系统基于上述实施例中公开的数据恢复方法和装置存在，通过数据恢复装置22接收用户端21发送的用户操作请求，将所述用户操作请求以操作日志形式进行存储，并基于该操作日志进行相关的分析和处理，经网络传输到需要恢复的容灾节点数据库中，并在这些容灾节点数据库中进行事务重做装载，从而实现各个容灾节点的数据一致性恢复。The data recovery system disclosed in this embodiment of the present invention exists based on the data recovery method and device disclosed in the above-mentioned embodiments. The data recovery device 22 receives the user operation request sent by the client 21, and stores the user operation request in the form of an operation log. Store, and perform relevant analysis and processing based on the operation log, and transmit it to the database of the disaster recovery node to be restored via the network, and perform transaction redo loading in these databases of the disaster recovery node, so as to realize the data of each disaster recovery node Consistency recovery.

综上所述：In summary:

通过上述本发明实施例公开的数据恢复的方法、装置和系统，不仅解决了数据在异构数据库容灾节点之间的恢复问题，而且通过日志分析减少了网络中的操作日志传输量，减轻日志传输对网络造成的负担，同时，通过事务综合使传输过程中以事务为单位进行传输，而不是以SOL语句为单位进行，保证数据的一致性和完整性；以及，在目标容灾节点数据库端进行事务包装载时避免执行无效的查询操作，使事务包装载效率更高，进一步提高了目标容灾节点数据库的数据恢复速度。The data recovery method, device and system disclosed in the above-mentioned embodiments of the present invention not only solve the problem of data recovery between heterogeneous database disaster recovery nodes, but also reduce the operation log transmission volume in the network through log analysis, and reduce the log The burden caused by transmission on the network, at the same time, through transaction synthesis, the transmission process is carried out in units of transactions instead of in units of SOL statements to ensure data consistency and integrity; and, on the database side of the target disaster recovery node Avoid executing invalid query operations during transaction package loading, making transaction package loading more efficient, and further improving the data recovery speed of the target disaster recovery node database.

本说明书中各个实施例采用递进的方式描述，每个实施例重点说明的都是与其他实施例的不同之处，各个实施例之间相同相似部分互相参见即可。对于实施例公开的装置而言，由于其与实施例公开的方法相对应，所以描述的比较简单，相关之处参见方法部分说明即可。Each embodiment in this specification is described in a progressive manner, each embodiment focuses on the difference from other embodiments, and the same and similar parts of each embodiment can be referred to each other. As for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and for the related information, please refer to the description of the method part.

结合本文中所公开的实施例描述的方法或算法的步骤可以直接用硬件、处理器执行的软件模块，或者二者的结合来实施。软件模块可以置于随机存储器(RAM)、内存、只读存储器(ROM)、电可编程ROM、电可擦除可编程ROM、寄存器、硬盘、可移动磁盘、CD-ROM、或技术领域内所公知的任意其它形式的存储介质中。The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be directly implemented by hardware, software modules executed by a processor, or a combination of both. Software modules can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other Any other known storage medium.

对所公开的实施例的上述说明，使本领域专业技术人员能够实现或使用本发明。对这些实施例的多种修改对本领域的专业技术人员来说将是显而易见的，本文中所定义的一般原理可以在不脱离本发明的精神或范围的情况下，在其它实施例中实现。因此，本发明将不会被限制于本文所示的这些实施例，而是要符合与本文所公开的原理和新颖特点相一致的最宽的范围。The above description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention will not be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.