Shulou(Shulou.com)06/03 Report--

[server raid5 disk Array Overview]

A MS SQL SERVER server used by a government agency of Xizang is a raid5 disk array composed of four hard drives, which is divided into a logical volume, and a single disk is a 73GSCSI hard disk, which is built on an IBM X225 server. The operating system is WINDOWS 2003.

[server disk array failure]

A, the server in the normal use of raid suddenly paralyzed, check the server found that there are 3 hard drives are offline. B, the administrator chose 2 hard drives at random to force the online operation, but the operating system could not start. C. After booting the operating system with WINPE CD, you can see the data. D, copy the backed-up database files to the removable hard disk using zip file format. After testing on other server machines, it is found that the backup files cannot be decompressed correctly, but the file size and name are correct. E. hire a third-party maintenance company to replace the raid card on the server and rebuild a set of raid5 arrays. F, the administrator worked normally after reinstalling the system on raid, while trying to repair the database in the zip file, but it was still unsuccessful after a day.

[server disk array data recovery conclusion]

The result of this case is completely different from the server data recovery case I shared before, because the data was seriously damaged and could not be recovered in accordance with the customer's requirements.

[cause analysis of raid5 data recovery failure and raid5 data recovery recommendations]

A, when the server uses the RAID5 array for storage, it must maintain the raid in time to ensure that the raid is normal. Once it is found that a hard disk is offline, even if the server can run, it must back up the data in time, and rebuild the disk array that fails. Waiting for the server to crash before checking the raid disk array has been remedied afterwards, and the difficulty of data recovery has also increased accordingly.

B, if more than 2 hard drives of the raid5 array are offline due to untimely operation and maintenance work or other reasons, it is important to remember not to choose the hard disk online at will. If you boot the system in the case of the wrong disk, it may cause the raid state to change, destroy the important data in the disk array, and lead to data recovery failure. Specific countermeasures can be referred to engineer Zhang Yu's "after RAID damage, how should we deal with it urgently?" "this article.

C, as long as the directory area is normal or partially normal, you can use PE to see the directory, but this does not mean that the data area must be normal. In fact, in the second step, the system cannot start because the strong online operation is wrong, and you should stop the operation at this time. When the administrator reads the directory in PE, the file system has actually been loaded, destroying the metadata area of the normal file system (but it is possible to destroy the data to be recovered).

The reason why the DJR ZIP file cannot be unzipped is due to the wrong structure of RAID. The administrator put two hard drives online by forcing the online operation, and there are three hard drives online at this time, but one of the three hard drives is offline first, which leads to the mixing of the latest data with the old data in the first offline disk. At this time, the directory is correct, but the data area is chaotic, and the data can be completely recovered if the three hard drives are not fully synchronized.

E, if you are unable to solve the failure of the raid5 array, you should hire a maintenance company to pay attention to the items identified in the contract for data recovery, which can be used for data recovery processing (but it is best to consult several professional data recovery companies to determine the data recovery plan). If the maintenance company does not have the scope of data recovery services, then it is best to give priority to or directly select the data recovery company. Because in most cases, maintenance companies have no data recovery services to carry out data recovery, they have to find data recovery companies again, which not only has no direct control over data security, process norms, and other aspects, but also increases the difficulty of data recovery, reduces the success rate of data recovery, and leads to an increase in costs (sometimes terrible).

F, the most lethal action that makes it impossible to recover in this case is to rebuild the RAID5. When IBM X225 rebuilds RAID using SERVER RAID SUPPORT CD, all data is cleared by default. Other types of servers usually re-synchronize the check when rebuilding the raid, disrupting the original data structure, but the process of complete synchronization takes a certain amount of time. If the synchronization is not completed, you can still restore the remaining data that is not synchronized.

G, the administrator has synchronized for a day, and the 73G RAID member disks have been synchronized. The data has been completely destroyed.

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.