Research And Implementation Of Tuning Strategy For RAC

The database system is the core and foundation of the information application system. Inorder to adapt to the issues such as growing and contiuous operation of data volume andbusiness volume, yield high availability cluster technology. However how use this technologyto provide high availability and high performance of database systems has become animportant subject of the database application.Exploration and development database provide information security for Daqing oilfields'production and operation, with the rapid development of oilfield's information technology,staff need to acess the information of database at any time, so in this system, it chooses to useOracle RAC to ensure that the7*24hours of uninterrupted operation.However, due to thecomplexity of RAC architecture, how to manage and optimize the RAC database becomesnow faced difficulties.This article discussed the architecture of Oracle RAC cluster database system at first. Itcompared and analyzed the difference between the RAC architecture and traditionalsingle-instance database architecture on the basis of existing theory. At the same time, itpresented the methods to collect and diagnosis the information of database at first,and thenajusted the database. Besides it studited diagnostic methods of session-based andinstance-based in-depth respectively, and through the analyzed principle of Cache Fusionwhich is the core technology of the RAC thoroughly, this paper proposed scheme ofcomposite partition table and Service-TAF for dispersing hot datablock, in order to achievetrue availability in the conditions of load-balancing in RAC system.Information Provide System of Exploration and Development provides function ofinformation query and download for the professional departments of the Research Institute,whether the system operate is stable or not has a significant impact for each professionaldepartment, In this paper, we applied scheme of composite partition table and Service-TAF toInformation Provide System of Exploration and Development at the same time, and used theAWR data collection tools to diagnose the performance of database before and afteroptimization. Through the comparing data, it shows that after optimization of the Oracle RACcluster system is running more stable, the overall performance is better than that of the formeroptimization system obviously, so it has important realistic meaning.