磁盘阵列系统掉电保护技术研究与实现

发布时间：2018-03-11 08:03

本文选题：磁盘阵列　切入点：掉电保护　出处：《华中科技大学》2012年硕士论文　论文类型：学位论文

【摘要】：数据重要性越来越受到企业和用户的关注，系统可靠性也是企业在选购存储设备时首要考虑的因素。RAID（Redundant Array of Independent Disk独立冗余磁盘阵列）存储系统的可靠性可以通过冗余机制来保证，容忍一定程度的磁盘故障，但是当发生较大的故障，比如系统供电中断等情况时，系统中的运行数据和状态将无法得到保护，造成数据不一致，甚至带来更严重的经济损失。针对上述问题，研究分析了现有系统掉电保护设计实现一般方案的优缺点，分析了RAID控制器系统运行流程，提出一种适用RAID的掉电保护硬件和软件实施方案，可有效处理供电中断的异常，提高RAID系统的可靠性。该方案在硬件上集成具有掉电保护功能的电源模块，软件上结合程序模块实现关键数据掉电保护和加电恢复功能实现，并根据系统中配置信息的管理方式、CACHE数据组织方式和请求处理流程，设计相应的数据保护方案和数据恢复策略。系统采用软件模拟掉电信号实现电源模块的掉电保护功能，，并在系统中实现相应的软件模块，使系统达到以下的目标：可以对全局配置信息进行掉电保护、加电恢复，保证系统的配置信息能恢复到掉电前的最新状态；可以按设计的转存格式，进行快速转存；可以掉电保护未完成的写请求；可以在系统加电时恢复所有转存的数据。采用模拟掉电信号的方式进行测试，测试项目包括：在拷贝数据结束时（请求均已发送，数据在内存），采用模拟掉电信号的方式产生掉电中断，触发掉电保护，在系统加电恢复后，计算并对比原测试文件和拷贝到系统的文件的MD5，测试模块功能的正确性，测试结果表明开启掉电保护，数据对比正确；使用IOMETER工具，对比测试系统在增加掉电保护模块前后的读写性能，测试结果表明增加掉电保护模块对系统的读写性能基本没有影响。
[Abstract]:The importance of data is paid more and more attention by enterprises and users. The reliability of storage system can be guaranteed by redundancy mechanism. Some disk failures are tolerated, but when larger failures occur, such as system power outages, the running data and states in the system will not be protected, resulting in inconsistent data and even more serious economic losses. In view of the above problems, this paper analyzes the advantages and disadvantages of the design and implementation of the current system power-off protection, analyzes the running process of the RAID controller system, and puts forward a hardware and software implementation scheme for the power loss protection for RAID. It can effectively deal with the abnormal power supply interruption and improve the reliability of the RAID system. This scheme integrates the power supply module with the power loss protection function on the hardware, and realizes the key data power-off protection and power recovery function in the software combined with the program module. According to the management mode of configuration information in the system, data organization mode and request processing flow of CACHE, the corresponding data protection scheme and data recovery strategy are designed. The system uses software to simulate the electric signal to realize the power loss protection function of the power supply module. And the corresponding software module is implemented in the system, which makes the system achieve the following goals: the global configuration information can be power-down protection, power-on recovery, to ensure that the system configuration information can be restored to the latest state before power down; According to the design of the transfer format, can be quickly transferred, can be power-off protection of unfinished write requests, can be power on the system when the recovery of all the data transferred. At the end of the copy data (the request has been sent, the data is in the memory), the power off is generated by using the analog electric signal, and the protection of power down is triggered, the test items are as follows: at the end of the copy data, the data is in memory. After the power recovery of the system, the MD5 of the original test files and the files copied to the system are calculated and compared, and the function of the test module is correct. The test results show that the power off protection is turned on, the data contrast is correct, and the IOMETER tool is used. Comparing the read and write performance of the test system before and after adding the power off protection module, the test results show that the increase of the power loss protection module has no effect on the reading and writing performance of the system.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP333.35;TP309.3

【参考文献】