Data Recovery: Data Loss and Drive Failures

S.M.A.R.T Failure Detection

S.M.A.R.T. drive failure detection is supposed to detect a possible drive failure, which works by running constant diagnostics on the status of the hard drive’s mechanical and electronic components. S.M.A.R.T. is an acronym for Self-Monitoring, Analysis, and Reporting Technology, and it was developed to help forestall a sudden drive failure for which you were totally unprepared; more or less an early warning system. Unfortunately, it is far from foolproof. Many times, S.M.A.R.T. detection puts out false alarms. Nevertheless, it is a good idea to run the drive manufacturer’s diagnostic tools, just to be on the safe side. On the other hand, many times S.M.A.R.T. detection does not see a hard drive failure coming, and you are caught off guard with a sudden failure. It is not a good idea to place all of your data at the mercy of the S.M.A.R.T. system, for it really isn’t all that smart. Still, it doesn’t hurt to have it on.

S.M.A.R.T. detection works by checking the hard drive’s components and makes sure they are operating within a predefined margin of tolerance. An increase of error corrections, which may point to problems with the drive heads, an increase in drive spin-up time or an increase in the number of spin-up tries, which may point to mechanical failure, can all raise red flags in the S.M.A.R.T. system. Of course, sudden and unexpected problems do arise where the S.M.A.R.T. system cannot give an advance warning, such as in the case of electronics failure. Again, it is not wise to rely solely on the S.M.A.R.T. system for warnings on when to spring into action and backup your data. It’s much better to do regular backups all along so it won’t be much of a hardship if you do experience a drive failure.

RAID is Not as Secure as You Thought

This brings the topic of discussion around to the issue of RAID arrays, and the failure of a drive in an array. If you have a RAID array set up, the following information is assuming that you are familiar with RAID. If you are not familiar with it, take a gander at the PCMech article, RAID – Your Guide. With IDE and SATA RAID, there are 0, 1, 0+1, and JBOD arrays available, depending on what the RAID controllers on your motherboard or PCI RAID card will support.

RAID 0 offers speed in what is called striping. Since the data is divided between drives, the loss of one drive will result in the loss of all data between the two drives. Nevertheless, if you find yourself in this situation, mentioned earlier was a program called File Scavenger, where the website claimed that the software was able to recover data over striped arrays. Unfortunately, I did not have any free drives to test this feature to see how good it was on recovering data, if at all.

RAID 1, a.k.a. mirroring, does not offer the security that you may think it does. Exact duplication of a drive instantaneously does not protect you against bad sectors, data corruption, data deletion, and usually doesn’t protect you against a hard drive failure of the master drive. All these imperfections in the data are carried over to the mirrored drive, so anything that happens to the master drive, usually happens to the mirrored drive. It would be better to use the extra drive for real backup space.

RAID 0+1 is better than the first two forms of RAID, for it offers redundancy as well as striping. Depending on your setup, it may be possible to recover data from this sort of array.

JBOD arrays, which really don’t fall under the RAID category, can make multiple drives appear as one. Personally, I haven’t found much use for this other than under RAID 5. Like striping, if you lose one drive of this array, you lose everything.

Last but not least, RAID 5, which is widely available with SCSI and just starting to appear in the IDE/SATA RAID market, can be any sort of redundancy striping and JBOD array, complete with a swappable drive should one go bad. Mainly used in server applications, this is the most effective form of RAID, however, somewhat costly.

So, why is it worse to use RAID 0 and JBOD arrays than to just leave the drives as stand-alones? The odds are higher with two drives, rather than one, that your data can be lost. If either drive is lost, and you lose all the data that was in the array stored on both drives. With independent drives, if only one drive is lost, only the data on that drive is lost and not on any other. Therefore, it may be more productive to use one drive as normal, and the other to keep backups. The chances of both drives going bad at the same time are slim, yet if the computer gets hit by an outside force (electrical storm, fire, water damage, etc.), keeping backups on other media may be an addition form of security. This is just one good practice of data loss prevention. There are still others that you may want to put into use.