谁能解决速联系我：[email protected] 钱不是问题，重酬，多谢！其中一题，例如：
Final Project: A Replicated Distributed File System with Transactional Semantics 210 points (of which 10 are for design document)due Wednesday, July 30Your goal is to extend your transactional file system that you implemented for Assignment 2 to use a replicated server. You must use a primary copy scheme: when we test your system, our test client will communicate with only a single server. How you implement your replication scheme is up to you. You will use the same semantics and wire protocol as you used in A2, but now your file system must survive server crashes. You will be required to handle only a single component failure . You may assume that all failures are failstop and no network partitions occur. That is, we will test your system by killing one of your servers and will expect that the system will continue responding to client requests. We will test correctness of your system by checking that if the primary crashes the files and data from committed transactions are present in the file directory of the new primary. (This testing method eliminiates the need to implement the READ method: for once I am having you do less work ;)). Coordinates of the primary serverWe require that you specify the IP address and the port of the primary in the file primary.txt. Location of the primary.txt will be given to you by startup arguments. Our client program will use the IP address and port specified in that file to talk to your server, so it is very important that these are correct and are updated whenever the identity of the primary changes. Startup Arguments and ProcedureWe give you flexibility with respect to startup arguments and procedure. Please specify them in the README file submitted with your system. We require that one of the arguments is the absolute path of the primary.txt file, such as /home/user/final-project-test/primary.txt. We guarantee that this path will refer to a network-mounted file system, so primary.txt will be readable and writable by all replicas. We also require that another argument is the file directory where the server keeps the files created by the client.
Changes to Client/Server InteractionIn Assignment 2 your server was required to continue processing a transaction that has begun but has not committed. For this assignment, we relax that assumption. Your system is required to reflect committed transactions, but not those transactions that failed to commit due to a crash. This means that if the primary fails, the backup must be able to re-send acknowledgements of previously committed transactions to the client. Testing Your SystemHere are some of the tests that we will run when grading your system. It's a good idea to run the same tests before submitting your assignment.
Normal case operation We will ensure that your system correctly processes transactions when no failures occur.Interacton when there is a failed secondary We will test that your system continues to function properly after a secondary replica is killed.Interaction when there is a failed primary We will test that your system is up-to-date after the primary has failed. That is, the server that assumes the identity of the primary must be able to (1) acknowledge all transactions that committed before the failure of the old primary, and (2) have all the data for transactions committed by the old primary in its file directory.Interaction when a failed server comes back up We will run a test where we will kill a primary P1, then execute transactions against the new primary P2, then restart the old primary P1 and then kill the P2, forcing the P1 to become the primary again. We will check that P1 successfully acknowledges transactions that were (1) committed before P1 was killed, (2) committed after P1 failed and before it came back up, and (3) committed after P1 became the primary once again. We will also check that the data for all those transactions is in P1's files.Interaction when a client exhibits incorrect behavior We will make sure that if the client sends requests to the secondary replica, i.e., to the IP address and port NOT specified in primary.txt, the replica returns an error or does not respond to the client. We will be updating the specification of this assignment, so do visit this website for updates. Test clientUse this test client to test your server. This program will generate valid A2 protocol messages and send them to your server based on the commands you supply. Run the test client as "java TCPClient [server] [port]". It will display a help message explaining how to use the program.
Final Project: A Replicated Distributed File System with Transactional Semantics 210 points (of which 10 are for design document)due Wednesday, July 30Your goal is to extend your transactional file system that you implemented for Assignment 2 to use a replicated server. You must use a primary copy scheme: when we test your system, our test client will communicate with only a single server. How you implement your replication scheme is up to you. You will use the same semantics and wire protocol as you used in A2, but now your file system must survive server crashes. You will be required to handle only a single component failure . You may assume that all failures are failstop and no network partitions occur. That is, we will test your system by killing one of your servers and will expect that the system will continue responding to client requests. We will test correctness of your system by checking that if the primary crashes the files and data from committed transactions are present in the file directory of the new primary. (This testing method eliminiates the need to implement the READ method: for once I am having you do less work ;)). Coordinates of the primary serverWe require that you specify the IP address and the port of the primary in the file primary.txt. Location of the primary.txt will be given to you by startup arguments. Our client program will use the IP address and port specified in that file to talk to your server, so it is very important that these are correct and are updated whenever the identity of the primary changes. Startup Arguments and ProcedureWe give you flexibility with respect to startup arguments and procedure. Please specify them in the README file submitted with your system. We require that one of the arguments is the absolute path of the primary.txt file, such as /home/user/final-project-test/primary.txt. We guarantee that this path will refer to a network-mounted file system, so primary.txt will be readable and writable by all replicas. We also require that another argument is the file directory where the server keeps the files created by the client.
Changes to Client/Server InteractionIn Assignment 2 your server was required to continue processing a transaction that has begun but has not committed. For this assignment, we relax that assumption. Your system is required to reflect committed transactions, but not those transactions that failed to commit due to a crash. This means that if the primary fails, the backup must be able to re-send acknowledgements of previously committed transactions to the client. Testing Your SystemHere are some of the tests that we will run when grading your system. It's a good idea to run the same tests before submitting your assignment.
Normal case operation We will ensure that your system correctly processes transactions when no failures occur.Interacton when there is a failed secondary We will test that your system continues to function properly after a secondary replica is killed.Interaction when there is a failed primary We will test that your system is up-to-date after the primary has failed. That is, the server that assumes the identity of the primary must be able to (1) acknowledge all transactions that committed before the failure of the old primary, and (2) have all the data for transactions committed by the old primary in its file directory.Interaction when a failed server comes back up We will run a test where we will kill a primary P1, then execute transactions against the new primary P2, then restart the old primary P1 and then kill the P2, forcing the P1 to become the primary again. We will check that P1 successfully acknowledges transactions that were (1) committed before P1 was killed, (2) committed after P1 failed and before it came back up, and (3) committed after P1 became the primary once again. We will also check that the data for all those transactions is in P1's files.Interaction when a client exhibits incorrect behavior We will make sure that if the client sends requests to the secondary replica, i.e., to the IP address and port NOT specified in primary.txt, the replica returns an error or does not respond to the client. We will be updating the specification of this assignment, so do visit this website for updates. Test clientUse this test client to test your server. This program will generate valid A2 protocol messages and send them to your server based on the commands you supply. Run the test client as "java TCPClient [server] [port]". It will display a help message explaining how to use the program.
解决方案 »
- hibernate 数据字典取值问题
- org.hibernate.criterion.Criterion找不到
- 一个Struts2跳转不刷新的问题
- IE6下下拉框不能隐藏的问题
- soap 客户端，java.util.Date 类型，怎么序列化？请高手指点
- EJB的缺点
- 谁能帮忙写个傻瓜式简单的java+access数据库教程，让我java数据库入门。
- tomcat 的错误信息。。。。。急啊！tomcat 的错误信息。。。。。急啊！tomcat 的错误信息。。。。。急啊！
- 有用struts做过中英文系统的吗？能进来指点指点吗？
- 大家帮看一下，问题出在哪里，有关ejb的客户调用，很简单的原形，但通不过
- SQL问题
- 求<深入JAVA虚拟机>随书光盘， 源码！！
楼主【geminiiv】截止到2008-07-16 14:47:32的历史汇总数据(不包括此帖)：
发帖的总数量:3 发帖的总分数:180 每贴平均分数:60
回帖的总数量:5 得分贴总数量:4 回帖的得分率:80%
结贴的总数量:1 结贴的总分数:100
无满意结贴数:0 无满意结贴分:0
未结的帖子数:2 未结的总分数:80
结贴的百分比:33.33 % 结分的百分比:55.56 %
无满意结贴率:0.00 % 无满意结分率:0.00 %
楼主该结一些帖子了
It has the basic skills to be a distributed filesystem,supporting replication,failure detection and failure recovery.
It is implemented using Mysql,Ftp Server and Web Server.
Mysql as the file and directory tree;
Ftp Server as the file storage node;
Web Server as the client and the management node.And during the cource,I study some thesises about distributed filesystem.
According to your requirement,I introduce Hadoop to you.
Hadoop is a distributed application development architecture,and provides HDFS to supporting distributed filesystem.
I think you can refer to it. Good luck~~