Shulou(Shulou.com)06/02 Report--

Server failure description:

Customer built a raid5 disk array with 6 hard disks (SCSI hard disks, 300 gigabytes per disk) on Dell 2850 server with linux Redhat4 operating system and ext3 file system. The server crashed during use. After checking the array, the administrator found that two hard disks were offline. One of the hard disks was forcibly brought online, but the operating system started abnormally. The administrator realized the seriousness of the matter at this time, immediately shut down the server, and then contacted the data recovery center for data recovery.

Data Recovery Center Initial Inspection Conclusion:

In the server data recovery work, it is very common for raid5 two hard disks to be offline. Since raid5 disk array supports honor protection when a hard disk is offline, once multiple hard disks are offline, the server will be in a paralyzed state and will not be automatically online. Since the RAID controller has a certain surface sensitivity, most hard disk drops are caused by random reasons such as power fluctuations and controller bugs, so the dropped disk may not have serious physical failures. This is the case. However, at this time, the administrator's forced online operation is very risky. Once the online error occurs, the controller will cause some irreversible damage to the data. When the administrator enters the operating system, the file system is inconsistent and repaired. All the hard disk data in the server is inconsistent, and the data recovery is very difficult.

Server data recovery process:

First of all, a complete backup of all hard disks in the server was carried out. During the backup process, it was found that many hard disks already had bad sectors but did not go offline. The reason was that raid did not read bad sectors of hard disks. After the backup is completed, analyze the RAID composition structure of the original server, and then virtualize the RAID environment to verify the abnormal structure of the RAID, manually correct the structure of the server that was destroyed in the later period, and export the corrected data to an intermediate storage for temporary storage. The last step of data recovery is to build a new RAID 5 disk array on the server using a good hard disk and migrate the recovered data to the new RAID. Check that the recovered data is normal, and the data is successfully recovered.

Server Data Recovery Conclusion:

Once the RAID 5 disk array has multiple hard disks offline and the server is paralyzed, remember not to blindly force the online operation. If there is enough spare space, you can mirror all the source hard disks. There are two ways (Windows2003 or DOS, other operating systems are at risk):

1, can be the same as or larger than the source disk capacity of the hard disk as the target disk, the source disk all sectors CLONE mode to the target disk. Do the same for all disks.

2. Each source disk can be completely exported to a large capacity storage space (such as large capacity hard disk, NAS, SAN, DAS, etc.) in sector mode.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.