| With the increasing in various applications of Internet, the server has to do much more work,particularly in web server.In some business services,the outages of service or the loss of key data will result in great loss,these.The former servers stored data in many separated documents,but the I/O resources of server is limited,with the increasing in customer's request,the server will become slower in reaction,or broken down in serious emergences.In order to resolve this problem,in this paper we use Berkeley DB and cluster technology in server system and this is cal Cluster server based on Berkeley DB.The BDB Cluster is through using Berkeley DB in cluster system to form a server system.The use of Cluster technology can get stable, reliable, expandable server system, and the use of Berkeley DB database system can get more effective data processing system.To get these characteristics together,the BDB Cluster system is reliable, stable, expandable , lower in cost and effective.This paper first introduced the current developing situation and the meaning of WEB server.Then studied the system structer,the working principle,the working ways of the BDB Cluster Server.In this paper we discussed function and design method of each part.Lastly,we introduced testing results of BDB Cluster Server.In this paper we wrote some Apache extensible modules to achieve some functions and then loaded them into Apache server under runtime.At the same time deleted some Apache modules that are not needed.So the system would not get any system resource for this modules when Apache server was started.In this way some system resource were saved.In this paper, Berkeley DB's replicarion technology, recovery mechanism, lock mechanism, memory sharing mechanism are used in this system.The use of these technologies can keep the integrality and thesecurity of data in system,at the same time improving the efficiency of data operating.In this paper we wrote communicating program between servers based on TCP/IP with C, simplifying formats of communicating data between can effectively improve the communication efficiency between servers. In this paper, we used Linux's muti-thread to design serving programs working on servers,wrote shell program to moniter serving program,wrote makefile to compile program, used ab testing tool and testing program to test the whole system. |