服务器ZFS文件系统故障后如何恢复数据

服务器ZFS文件系统故障后如何恢复数据？很多人都不太了解，今天小编为了让大家更加了解数据恢复的过程，所以给大家总结了以下内容，一起往下看吧。

How to recover data after server ZFS file system failure? Many people don't know much about it. Today, in order to let you know more about the process of data recovery, Xiaobian summarized the following contents for you, let's look down.

一、服务器数据恢复环境部署介绍：

1、 Introduction to server data recovery environment deployment:

今天为大家介绍的数据恢复成功案例服务器型号为：ORACLE-SUN-ZFS7320。服务器内涉及硬盘32块，服务器操作采用的是Windows操作系统。

The server model of the successful data recovery case introduced to you today is oracle-sun-zfs7320. 32 hard disks are involved in the server, and windows operating system is used for server operation.

二、服务器数据恢复故障情况

2、 Server data recovery failure

服务器在正常运行的时候突然崩溃，没有断电、进水、异常操作、机房不稳定等外部因素。服务器管理员对设备进行重启后发现无法进入系统，需要对服务器内的数据进行恢复。

When the server is in normal operation, it suddenly collapses without external factors such as power failure, water inflow, abnormal operation and instability of the computer room. After the server administrator restarts the device, it is found that the device cannot enter the system, and the data in the server needs to be recovered.

三、分析服务器磁盘底层数据

3、 Analyze server disk underlying data

服务器管理员对所有硬盘进行扇区级镜像后将镜像文件送到数据恢复中心进行数据恢复。服务器数据恢复工程师对客户的故障服务器进行底层数据分析得到如下信息：故障服务器采用zfs文件系统；所有磁盘被分为4个组，每组8块硬盘；热备盘全部启用。

The server administrator will send the mirror files to the data recovery center for data recovery after sector level mirror of all hard disks. The server data recovery engineer analyzes the underlying data of the client's fault server and obtains the following information: the fault server adopts ZFS file system; all disks are divided into four groups, each group has eight hard disks; all hot spare disks are enabled.

四、服务器故障情况分析

4、 Server failure analysis

在服务器ZFS文件系统中，池被称为ZPOOL。ZPOOL的子设备可以有很多种类，包括块设备、文件、磁盘等等，在本案例中所采用的是其中的一种------三组RAIDZ作为子设备。经过分析发现，三组RAIDZ内有两组分别启用热备盘个数为1和3。在启用热备盘后，第一组内仍出现一块离线盘，第二组内则出现两块。以此进行故障现场模拟：三组RAIDZ内第一二组分别出现离线盘，热备盘及时进行替换；热备盘无冗余状态下第一组出现一块离线盘，第二组出现两块离线盘，ZPOOL进入高负荷状态（每次读取数据都需要进行校验得到正确数据）；第二组内出现第三块离线盘，RAIDZ崩溃、ZPOOL下线、服务器崩溃。

In the server ZFS file system, the pool is called zpool. Zpool can have many kinds of sub devices, including block devices, files, disks, etc. in this case, three groups of raidz are used as sub devices. After analysis, it is found that there are two groups of raidz that enable the number of hot spares to be 1 and 3 respectively. After the hot spare is enabled, there is still one offline disk in the first group and two offline disks in the second group. In this way, the fault field simulation is carried out: the first and second groups of raidz have offline disks respectively, and the hot spare disks are replaced in time; when the hot spare disks have no redundancy, the first group has one offline disk, the second group has two offline disks, and zpool enters the high load state (each time reading data needs to be verified to get the correct data); the second group has the third offline disk, and raidz crashes , zpool offline, server crash.

五、重组ZPOOL，追踪数据入口

5、 Reorganize zpool, track data entry

ZFS管理的存储池与常规存储不同，所有磁盘都由ZFS进行管理。常规RAID在存储数据时，只按照特定的规则组建池，不关心文件在子设备上的位置。而ZFS在数据存储时会为每次写入的数据分配适当大小的空间，并计算得到指向子设备的数据指针。这种特性使得RAIDZ缺盘时无法直接进行校验得到数据，必须将整个ZPOOL作为一个整体进行解析。手工截取事务块数据，编写程序获取最大事务号入口：获取文件系统入口获取到文件系统入口后，编写数据指针解析程序进行地址解析：解析数据指针获取到文件系统入口点在各磁盘分布情况后，开始手工截取并分析文件系统内部结构，入口分布所在的磁盘组无缺失盘，可直接提取信息。根据ZFS文件系统的数据存储结构顺利找到客户映射的LUN名称，进而找到其节点。

The storage pool managed by ZFS is different from conventional storage, and all disks are managed by ZFS. When a regular raid stores data, it only builds a pool according to specific rules, and does not care about the location of files on the sub devices. When ZFS stores data, it allocates a proper amount of space for each write and calculates the data pointer to the sub device. This feature makes it impossible to check the data directly when raidz is short of disk, so the whole zpool must be analyzed as a whole. Manually intercept the transaction block data, and write a program to obtain the maximum transaction number entry: after obtaining the file system entry and obtaining the file system entry, write a data pointer resolver for address resolution: after parsing the data pointer to obtain the file system entry point in each disk distribution, start to manually intercept and analyze the internal structure of the file system, and the disk group where the entry distribution is located is intact Information can be extracted directly if the disk is lost. According to the data storage structure of ZFS file system, find the Lun name mapped by the customer and then find its node.

六、编写数据提取程序并运行

6、 Write data extraction program and run it

经过仔细分析，发现在此存储中的ZFS版本与开源版本有较大差别，无法使用公司原先开发的解析程序进行解析，所以重新编写了数据提取程序。提取数据程序由于磁盘组内缺盘个数较多，每个IO流都需要通过校验得到，提取进度极为缓慢。与客户沟通后得知，此ZVOL卷映射到XenServer作为存储设备，客户所需的文件在其中一个大小约为2T的vhd内。提取ZVOL卷头部信息，按照XenStore卷存储结构进行分析，发现2T vhd在整个卷的尾部，计算得到其起始位置后从此位置开始提取数据。

After careful analysis, it is found that the ZFS version in this storage is quite different from the open source version, so it can't be parsed by the original resolver developed by the company, so the data extraction program is rewritten. Due to the large number of missing disks in the disk group, each IO stream needs to be verified, and the extraction progress is very slow. After communicating with the customer, we know that this zvol volume is mapped to XenServer as a storage device, and the file required by the customer is in one of the VHD with a size of about 2T. Extract the head information of zvol volume and analyze it according to the xenstore volume storage structure. It is found that 2T VHD is at the end of the whole volume. After calculating its starting position, extract the data from this position.

七、验证数据完整性，服务器数据恢复成功

7、 Verify the data integrity and recover the server data successfully

Vhd提取完毕后，对其内部的压缩包及图片、视频等文件进行验证，均可正常打开。联系客户验证数据，确定文件数量与系统自动记录的文件个数一致。验证文件可用性，文件全部可正常打开，服务器数据恢复成功。

After VHD extraction, the internal compressed package, pictures, videos and other files are verified and can be opened normally. Contact the customer to verify the data and make sure that the number of files is consistent with the number of files automatically recorded by the system. Verify the availability of the file. All the files can be opened normally. The data of the server is recovered successfully.

看完上诉内容，你们对服务器ZFS文件系统故障后如何恢复数据大概了解了吗？如果想了解更多，欢迎关注主机参考行业资讯频道哦！

After reading the contents of the appeal, do you have a general understanding of how to recover the data after the server ZFS file system fails? If you want to learn more, please pay attention to the host reference industry information channel!

主机参考，作为业内资深的专业云计算服务提供商、云安全服务提供商，致力于为广大互联网企业用户和传统行业的企业用户提供“云服务器、裸金属服务器、高防服务器、香港服务器、美国服务器”等云主机租用服务以及企业上云的综合解决方案，具有“安全稳定、简单易用、服务可用性高、性价比高”等特点与优势，专为企业上云打造定制，能够满足用户丰富、多元化的应用场景需求。

Host reference, as a senior professional cloud computing service provider and cloud security service provider in the industry, is committed to providing "cloud server, bare metal server, high defense server, Hong Kong server, US server" and other cloud host rental services as well as integrated solutions for Enterprise Cloud access for Internet enterprise users and traditional industry enterprise users, with "security and stability" Features and advantages such as fixed, simple and easy to use, high service availability and high cost performance "are specially designed for enterprises to build customized cloud applications, which can meet the needs of users' rich and diversified application scenarios.

--------------------------------------------------------------
主机参考，收集国内外VPS，VPS测评，主机测评，云服务器，虚拟主机，独立服务器，国内外服务器，高性价比建站主机相关优惠信息@zhujicankao.com
详细介绍和测评国外VPS主机,云服务器,国外服务器,国外主机的相关优惠信息,商家背景,网络带宽等等,也是目前国内最好的主机云服务器VPS参考测评资讯优惠信息分享平台

这几篇文章你可能也喜欢：

• DesiVPS，美国VPS特价8折循环优惠低至$15/年，美国圣何塞，1Gbps带宽/不限流量，高性能AMD系列VPS新品上线
• RackNerd：15美元/年/1核@Ryzen 7950X/512MB内存/6GB NVMe空间/400GB流量/1Gbps端口/KVM/洛杉矶/圣何塞/芝加哥/纽约/达拉斯等。
• DesiVPS 推出全新 AMD Ryzen9+NVMe 系列 VPS，起价每年 24 美元，圣何塞机房
• Sharktech，美国高防御大带宽不限流量服务器秋季特惠，60G防御/10Gbps带宽/不限流量，洛杉矶/拉斯维加斯/芝加哥/丹佛/阿姆斯特丹
• Mondoze：住宅IP/原生IP/IDC IP，VPS低至$8.33/马来西亚服务器/AS152742/11.11促销

本文由主机参考刊发，转载请注明：服务器ZFS文件系统故障后如何恢复数据 https://zhujicankao.com/14148.html