When I downloaded the program from Hitachi called "WinDFT" that scans the HDD for errors, after running both the long and short tests, it tells me that the HDD has no error so I cannot RMA the drive.
I read somewhere that it is possible that a one time error occur and SMART records and remembers it. So even though the error doesnt occur anymore, SMART still think the drive going to fail because it remembers the error last time. Is this true?
Based on the information provided, is my drive going to fail? Thank you for any advice
SMART data is very accurate in my opinion so you can believe the statistics it's giving you. However, it's just an indication that the drive is starting to wear out. How the various disk-testing utilities are programmed to treat that data when reporting a test result is another matter.
Another thing to remember is that if the drive in question is the system drive (ie you run Windows from it) you cannot properly test it anyway with any Windows-based testing software such as WinDFT.
The system drive should only be tested with DOS-based diagnostic software, so you should be using the original Hitachi DFT for DOS which loads from a bootable CD, not Hitach WinDFT which is for testing non-system Hitachi hard drives and Hitachi external hard drives only. You will get inaccurate results using WinDFT to test your system drive.
I see.. so what you're saying is that the Hitachi WinDFT could be interpreting the data as normal wear and not an imminent possible failure whereas windows tries to be on the safe side and warns me that the drive is about to fail even though it may take a few months or years for the problems detected by SMART to progressively become more serious?
So far the drive hasn't shown any increase in the values of SMART attributes.
I made a backup of the drive using windows backup and formatted the drive using windows (I did not select quick format). "Current Pending Errors Count" have returned to normal but the other 3 attributes mentioned earlier are still showing values above the threshold.
My system drive is a Corsair 120GB SSD, the Hitachi drive in question is my secondary drive. Based on what you have said, WinDFT should be accurate. However it cannot replicate the error results from HDDScan 3.3
What should be my next course of action? Is the drive safe to use?
All those SMART values are above their thresholds. That's good. An attribute is considered to have failed when its normalised value falls BELOW the threshold. In fact the normalised values (eg 86) are "health" scores. They usually begin at 100 and decline as the number of errors increases. The raw value of the attribute usually indicates the actual error count.
That said, the fact that the Reallocation (sic) Sectors Count has declined from 100 to 86 does suggest that your drive is not healthy. Therefore I wouldn't trust it. My advice would be to backup your data and keep an eye on the SMART report.
BTW, the reason that the Current Pending Errors Count attribute may have cleared up is that it reports those bad sectors that are waiting to be replaced. When you formatted the drive, you rewrote the data in those sectors. The drive then knew that the original data in those sectors were no longer of any importance. These pending sectors would then have been retested and returned to service if good, or replaced with spares if confirmed bad. Any sectors that were replaced would then have appeared in the Reallocated Sectors Count.
ic... but a score of 86/100 is way above the threshold of 5. So why is it a sign that the drive is unhealthy if the threshold is still far away. Why did SMART decide that 5 is the correct threshold value anyway?
I just ran a extended surface test using WinDFT and it reported that the drive failed the extended surface test. However it passed the extended SMART test. May I know what is the difference between the surface test and the SMART test. Does surface mean that the drive surface is wearing out and is unrepairable by software means?
I've never used Hitachi's WinDFT diagnostic, but generally speaking, a SMART test is run internally by the drive. That is, the software application (host) will use a single ATA command to initiate a particular SMART selftest, whether short or long. The drive will then process this command without further intervention by the host. The nature of these tests is defined by each HDD manufacturer in their own particular way. I presume that the long test scans the entire surface of the drive, but I'm not certain.
As for the surface test, it would read every sector on the drive. It would do so by sending a continuous stream of ATA read commands over the SATA interface. I don't know if the surface scan does anything more than a long self test, so I can't answer your question as to why one passed and the other failed.
Your question regarding the SMART thresholds is difficult to answer. That's because the significance of the Reallocated Sector Count is a matter of opinion. Each HDD manufacturer allows for a certain number of reallocations before a drive is deemed to have failed. For example, some Seagate models allow for 2500 bad sectors, while some WD models allow 500. Unfortunately each HDD manufacturer does not assess a drive as being "healthy" or "unhealthy", per se. Instead, their SMART diagnostics merely tell you whether the attributes are above their thresholds, ie they are pass-or-fail tests. In your case, a drive with a score of 6 would be considered to be OK whereas a drive with a score of 5 would be deemed to have failed. Most users would have become concerned long before then, and rightfully so. In fact some users become concerned when they see even 1 reallocation.