CN113111122A - Real estate registration data read-write separation method facing RAC cluster - Google Patents

Abstract

The invention provides a real estate registration data read-write separation method facing RAC cluster, which comprises the steps of establishing a database A, B, automatically distributing data read-write task distribution of double nodes through the CPU and IO performance of a server on the double nodes checked by a program on a C node, accessing an A node by a high-frequency read-only module in a design system, accessing the C node by a read-write mixed module, and accessing a B node by a data batch writing module. The RAC and Exadata Smart Scan technologies and the like adopted by the invention realize flexible scheduling of multi-node distributed data read-write resources of the database and read-write separation application according to the demand scene design, comprehensively improve the processing capacity and throughput of the system, effectively improve the data read-write efficiency, maximally, reasonably and effectively utilize the performance of the server and improve the working efficiency.

Description

Real estate registration data read-write separation method facing RAC cluster

Technical Field

The invention belongs to the technical field of real estate data application and utilization, and particularly relates to a real estate registration data read-write separation method facing an RAC cluster.

Background

The Guizhou province real estate registration cloud platform is a large centralized platform of the whole province, covers registration and certification work of 88 counties, carries out service declaration by thousands of financial institutions and real estate development enterprises of the whole province, simultaneously supports data sharing and utilization of 12 provincial horizontal departments, has the maximum online number of more than 5000 people and the maximum single-day write record of more than 100 thousands, reasonably improves efficiency in the face of high-frequency data reading and writing, improves system performance by adopting a multi-node data reading and writing mode, and prevents the problem of time waiting caused by data communication blockage on a single node.

Disclosure of Invention

Aiming at the defects in the prior art, the ORALCE RAC-oriented data read-write separation method provided by the invention solves the problems.

In order to achieve the above purpose, the invention adopts the technical scheme that:

the scheme provides a real estate registration data read-write separation method facing RAC clusters, which comprises the following steps:

s1, establishing three ABC nodes on an ORALCE RAC, and constructing a C node as a virtual node;

s2, designing a multi-node database and server performance scanning and task allocation mechanism on the C node;

s3, providing ABC three-node database connection of ODP.NET by the application system;

s4, designing system functions according to read-only application, write-only application and read-write mixed application;

and S5, when the C node task distribution mechanism receives a client-initiated SQL instruction, preprocessing the data and distributing the tasks through the Exadata Smart Scan technology, and finally returning the result to the client by the database server.

The invention has the beneficial effects that:

the invention provides a method for designing an ORALCE RAC-oriented data read-write separation and application system, which comprises the following steps: two nodes of a database A, B are established, the data read-write task allocation of the double nodes is automatically allocated through checking the CPU and IO performance of a server on the double nodes through a program on the C node, a read-only module with high frequency in a design system accesses the A node, a read-write mixed module accesses the C node, and data are written into the module in batch to access the B node. The technologies such as RAC and Exadata Smart Scan adopted in the design realize flexible scheduling of multi-node distributed data read-write resources of the database and design read-write separation application according to a demand scene, comprehensively improve the processing capacity and throughput of the system, effectively improve the data read-write efficiency, maximize and rationalize the performance of the server and improve the working efficiency.

Drawings

FIG. 1 is a flow chart of the method of the present invention.

Detailed Description

The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.

Examples

As shown in fig. 1, the present invention provides a real estate registration data read-write separation method facing an RAC cluster, which includes the following steps:

s1, establishing three ABC nodes on an ORALCE RAC, and constructing a C node as a virtual node;

s2, designing a multi-node database and server performance scanning and task allocation mechanism on the node C, which is specifically as follows:

(1) the client initiates an SQL instruction, the task distribution mechanism constructs an exadata specific iDB instruction for the query through the exadata, and then the instruction is sent to all the exadata storage server cells through a certain algorithm;

(2) the Exadata storage software scans all related data blocks according to the instruction and screens out the rows and columns meeting the requirements;

(3) the Exadata storage software returns data meeting the conditions to the database instance by using an iDB protocol, and meanwhile, the data cannot be cached in a buffer cache because the Exadata storage software is not used for caching;

(4) and integrating the results returned by all the storage cell nodes by the data block kernel, and then assembling a complete result set required by the client, wherein the process is similar to the whole parallel query result set.

In this embodiment, a user creates a task form (table 1) in Oracle according to the task content of the service system, and the form information includes: unique identification of the database, task name, task group, task program set, task processing program set, state and remark.

TABLE 1

Serial number	Name of field	Name of Chinese character	Type of field	Length of field
					1	id	Unique identifier of database	int	11
2	JobName	Task name	varchar	50
					3	JobGroup	Task group	varchar	50
4	JobTypeFullName	Task program set	varchar	200
					5	JobClassAssembly	Task processing class assembly	varchar	50
6	Status	Status of state	int	11
					7	Remark	Remarks for note	varchar	1000

S3, providing ABC three-node database connection of ODP.NET by the application system;

s4, designing system functions according to read-only application, write-only application and read-write mixed application, which specifically comprises the following steps:

(1) establishing A, B, C three database connection character strings at the server;

(2) the data extraction main content of designing and simultaneously using the A node to read data and the B node to write data for the big data analysis of real estate comprises the following steps: real estate classification information, transaction process information, rights achievement information, house information, land information, financial information, and rights-of-interest information; the step of simultaneously inputting the extracted data into a real estate analysis subject database comprises the following steps: total analysis subject bank, transaction rate analysis subject bank, transaction information subject bank, mortgage condition subject bank, internet + subject bank, external inquiry subject bank, efficiency analysis subject bank. The design uses the C node function, is used for reading and writing the module that is used mixedly, distribute the main content of the person's task according to A, B performance situation automatically by the task distribution mechanism and include: the system comprises a result report module for external inquiry of rightful persons, a business acceptance and handling process module, a right book result library entry module, a data modification module, a data exchange module and the like.

S5, when the C node task distribution mechanism receives a client-side initiated SQL instruction, preprocessing and task distribution are carried out on data through an Exadata Smart Scan technology (Exadata data scanning and distributing technology, and Smart Scan data technology for reducing I/O interaction), and finally the database server side returns the result to the client-side.

In this embodiment, the use modes of the three nodes are set A, B, C according to the application mode of reading and writing.

In this embodiment, the real estate registration platform and the statistical supervision and analysis system are split into data extraction and writing into the point C according to the functions of the application modules, and the data extraction and writing into the point C are completed quickly by using the automatic task allocation mechanism. The data model analysis of the statistical supervision analysis system mainly uses the application of reading data to be placed on the node A as a command for priority processing, and the real estate registration platform and the statistical supervision analysis system reasonably utilize the data reading and writing of multiple nodes to improve the efficiency.

Through the design, the invention provides a multi-node high-frequency data mass reading and writing method, the performance of a database and a server is utilized to the maximum extent, the reading and writing channel can be set in a user-defined mode according to the application scene of real estate registration service, along with the improvement of the service volume and the coverage area, the system provides a processing method which is better in flexibility and more flexible, the resource utilization rate is improved, and the hardware cost is reduced.

Claims (1)

1. A real estate registration data read-write separation method facing RAC cluster is characterized by comprising the following steps:

s1, establishing three ABC nodes on an ORALCE RAC, and constructing a C node as a virtual node;

s2, designing a multi-node database and server performance scanning and task allocation mechanism on the C node;

s3, providing ABC three-node database connection of ODP.NET by the application system;

s4, designing system functions according to read-only application, write-only application and read-write mixed application;

and S5, when the C node task distribution mechanism receives a client-initiated SQL instruction, preprocessing the data and distributing the tasks through the Exadata Smart Scan technology, and finally returning the result to the client by the database server.

Images