[Server Storage Data Recovery] A data recovery case of a RAID 5 crash caused by the failure of a certain model of Huawei OceanStor storage RAID 5 hard disk and the failure to synchronize data with the

服务器存储raid数据恢复环境：

华为OceanStor某型号存储,16块FC硬盘（Contains a hot spare）组成RAID5.

服务器存储raid故障：

该存储RAID5One of the hard drives in is offline for unknown reasons,The hot spare disk goes online to synchronize data,About halfway through the synchronization, another hard disk goes offline for unknown reasons,数据同步失败,raid5崩溃,上层的lun失效.The administrator contacts our data recovery center for this in storageraid5进行数据恢复.

服务器存储raid数据恢复过程：

1、检测raid5All disks in ：

North Asia hardware engineers are first on the faultraidPhysical failure detection is performed on all disks in the system,The test result is that the first disconnected hard disk is physically faulty,Others, including the hard drives that were dropped later, have no physical failures.

2、备份raid5All disks in ：

将raid5All disks are mirrored backed up,Subsequent data recovery operations are performed in the image file,Avoid possible secondary damage to the original data.

3、分析raid5结构：

Data recovery engineer on failureraidAll disks in the file are analyzed,Find the hot spare（Hot spare disks are significantly different from other data disks）.由于raidis striped,raidThe data in is stored according to certain rules,通过分析raidThe distribution of database pages in each physical disk,North Asia data recovery engineers calculatedraiddisk order、数据走向、Basic information such as stripe size.

4、找出raid5The bad disk being synced in：

根据分析获取到的RAID信息,尝试通过北亚自主开发的RAIDThe virtual reorganizer will be originalRAIDVirtually reassembled.故障RAIDTwo disks are disconnected and the data in one of the hard disks is damaged due to synchronization failure,Data recovery engineers carefully analyze the data in each hard drive,发现有一块硬盘在同一个条带上的数据和其他硬盘明显不一样,Therefore, it is preliminarily judged that this hard disk is a hard disk damaged by synchronization.通过北亚自主开发的RAID校验程序对这个条带进行校验,Finalize the hard drive that was damaged by synchronization.

5、分析raid中的lun信息：

分析lun在故障raidAllocation status and in lun分配的数据块,然后根据数据MAP导出LUN的数据.

6、解析EXT3文件系统：

Because it is virtual reorganization using hot spare diskRAID,EXT3文件系统无法正常挂载,只能提取出oracle数据库文件.Use the file system parsing program independently developed by Beiya to parse and export the file systemoracle数据库文件,The database file is then handed over to a database recovery engineer for checksum verification.

7、修复数据库,Check data file integrity：

使用OracleThe database file detection tool checks the integrity of each database file but finds errors.So use the North Asia independent research and developmentOracleDatabase detection tool to detect,Some database file and log file errors were found, system和sysauxThere are many bad blocks in the tablespace,All control files are corrupted,eschoolspaceTablespace files have more bad blocks,undotbs02丢失;Beiya database data recovery engineers repair these files manually.

North Asia database data recovery engineers create control files、undo表空间,启动数据库到mount.由于systemThere are bad blocks in the data file,数据库不能open,Various implicit parameters cannot be bypassed eithersystem的坏块.Build a database environment and use itdmp文件还原数据库,最终导入10G左右的数据.

数据验证：

It is started with the cooperation of the user sideOracle数据库,Install on a local virtual machineOA客户端,通过OAThe client verifies the data record,And let users arrange personnel from different departments for remote verification.验证通过,数据恢复成功.