EXT3 File system failure

A Web server. Data volume is a 1TB hard disk, did not do Raid,ext3 file system, unknown reason, remote inaccessible, to the engine room query failure found file system can not mount, did not do fsck. Administrators for security, trying to the entire hard drive DD to another piece of the same hard disk for analysis and data recovery work, but the DD backup can only back up the first 320M data, the subsequent unreadable, trying to adjust the starting position is still useless. Because the machine room environment is very noisy, cannot distinguish the hard disk to have the dissimilar sound.

The user asked a few questions:

1, what is the initial estimate of the failure? How likely is the data recovery?

Answer: There are two most likely cases: bad or head instability. If it is a bad reason, it can be read by a professional reading device such as pc3k DE, which is usually not readable except for the absolutely physically damaged sector, which has a great opportunity to read. If the head is unstable, you need to replace the head to read the data, the success rate of the opening change head depends on the ability of the engineer, the physical environment of the opening, the matching degree of the backup, the mirror mode of the data after the opening (high success rate of the device reading), but the decisive factor is whether the disk surface is scratched, if there is a scratch, Even if the other way is the best, can not restore success.

2, my data is very small, about dozens of G, according to EXT3 format to see, will not be stored in the front of the hard disk. So, if it's bad, can you pick a recovery, like just restoring the previous data?

A: The format of the ext3 is really to store the data and nodes in an incremental way, in terms of block groups, which are usually stored in the past, but where the data is deleted and added to the stored procedure, its storage location may change.

For bad, if the file system structure can be read at one level, it can be recovered by choice, or it can be determined that the data is in the front dozens of G, or read in this section.

3. How is this situation produced? How should we avoid it?

A: Even with a SAS drive, it is not safe to make a single disk, let alone a SATA hard drive (1TB hard drive is currently not a SAS or IDE or SCSI). A single hard drive is difficult to guarantee against any failure in a long time use environment. If you need to avoid it, you can choose to use RAID for data protection, and you need to do backup work (backup strategy, etc.).