Research And Implementation Of A High Availabile Mid-framework That Orients To Cluster Servers

High-availability systems, also known as the disaster recovery system. It providesan environment to cope with various disasters. When the computer system suffersirresistible natural disasters such as fire, flood, earthquake, war, and computer crime,computer viruses, power-down, network/communication failures, hardware/softwareerrors and human error and other man-made disasters, disaster recovery system protectsdata, recovers errors from disaster, and also provides uninterrupted service. The goal ofhigh availability technology is to quickly and accurately process failure and reduce theloss. A high availability cluster-oriented middle framework has been researched anddesigned in the thesis. The framework provides the underlying universal secondarydevelopment interface, is easy to use and able to be a low-level framework for a highavailability or disaster recovery system. During the research and design period, thesearch engine system, cloud platform, as well as big data areas are referenced andstudied. Both existing technology and trends are researched.Technology that requires by high-availability system has been researched andsummarized, including skills for cross-platform programming, cross-architectureprograming, object-oriented programming, component-oriented programming, designpattern and so on. Several sub-modules have been designed and implemented, includinglow-level transportation protocol layer-used for the network communications, groupand membership automatic broadcasting module-used to maintain global state, fencingmodule–used to handle exceptions of framework and avoid brain-split, andhigh-availability components daemon-the only running instance on a node for globalmanagement and command communications. In addition to above, a sample ofhigh-availability file sharing server is created with the framework, which exports CIFSservice to windows OS users on Linux.