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

Principle of Raid 5 data recovery algorithm

To understand the principle of raid5 data recovery, we must first understand raid5. The independent disk structure of distributed parity (that is, we call it raid5) data recovery needs to understand the concept of "parity". We can simply think of it as an "XOR operation" in binary operations, and the commonly used logo is xor. The rule of operation is that if the values of the two are the same, the result is 0, and if the results are different, the result is 1.

For example, if 0101 xor 0010 is calculated according to the above operation rules, the first bit of both is 0, and the result is 0. If the values of the second, third and fourth digits are different, the result is 1, so the final result is 0111, which is expressed by the formula as: 0101 xor 0010 = 0111, so if one of them is missing in a xor breadc, we can calculate it from other data. This is the whole secret of raid5 data recovery. After knowing this secret, it will not be a problem for raid5 data to lose your own recovery data.

Raid5 data recovery success story 1:

Customers use Dell 2850 server to build a raid5 disk array, which contains 6 hard disks (SCSI hard disk, single disk capacity 300G), and the server operating system is linux Redhat4; file system and ext3 file system. The server was paralyzed in the process of use, and the administrator checked the array and found that two hard drives were offline and one of them was forced to go online, but the operating system started abnormally. At this time, the administrator realized the seriousness of the matter, immediately shut down the server, and then contacted the data recovery center for data recovery.

Conclusion of initial inspection of data recovery center:

In the server data recovery work, it is very common for two raid5 hard drives to go offline. Because the raid5 disk array supports the honor protection of one hard disk offline, once multiple hard drives are offline, the server will be paralyzed and will not go online automatically. As the raid controller has a certain surface sensitivity, most of the hard disk drop is closely due to power fluctuations, controller bug and other random reasons, so the disconnect disk may not have serious physical failure. This is the case. But at this time, the risk of forced online operation by the administrator is very great, once the online error will cause the controller to produce some irreversible damage to the data, and when the administrator enters the operating system, it will be repaired because the file system is inconsistent, all the hard disk data in the server is inconsistent, and the data recovery is very difficult.

Server data recovery process:

First of all, make a full backup of all the hard drives in the server, and during the backup process, it is found that several hard drives have bad paths but are not offline, because raid did not read the bad paths of the hard drives. After the backup is completed, the raid structure of the original server is analyzed, and then the raid environment is simulated to verify the raid structure malformation. the structure destroyed in the later stage of the server is manually modified, and the modified data is exported to an intermediate storage for temporary storage. The last step of data recovery is to use the intact hard disk to build a new raid5 disk array on the server and migrate the recovered data to the new raid. Verify that all the recovered data is normal and the data is recovered successfully.

Server data recovery conclusion:

Once multiple hard drives are offline and the server is paralyzed in the raid5 disk array, don't force the online operation blindly. If there is enough spare space, you can mirror all the source hard drives. There are two ways (under WINDOWS2003 or DOS, other operating systems are at risk):

1. You can use the hard disk that is the same or larger than the capacity of the source disk as the target disk, and CLONE all sectors of the source disk to the target disk. Do the same operation on all disks.

2. Each source disk can completely output files to a mass storage space in sector mode (such as high-capacity hard disk, NAS, SAN, DAS, etc.)

Raid 5 data recovery case 2:

Raid disk array in a server, there are 12 disks in the disk array, the single disk capacity is 500G file system, and the system platform is Linux platform. Disk 2 and disk 6 in Raid report yellow lights. After checking, the manager made a mandatory online operation on disk 6. After being forced to go online, it was found that most directories could not be opened, and some files were lost in some directories that could be opened. The customer contacted the North Asia data recovery Center for data recovery of the raid disk array.

1. First of all, analyze the original raid structure, remove the No. 6 disk that is forced to go online, add disk No. 2, and build the raid platform through the virtual environment.

two。 Then the software is used to extract the auxiliary data, and the irregular bad track is found in the No. 2 plate after extracting the data.

3. We chose a commonly used mirroring tool to fully mirror the No. 2 hard disk, and more than 90% of the bad channels were read successfully.

4. Add the mirrored data to the virtual raid environment, extract the data again, and more than 99% of the data has been completely recovered.

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.