Question Different tests show different results for my HDD health

andrei1237

Prominent
Jun 9, 2019
7
0
510
I have external hard disk and i checked its health with CrystalDiskInfo, Windows 7 chkdsk, and HD Tune. CrystalDiskInfo showed that the health status is good, Windows 7 chkdsk didn't find any errors, but HD Tune showed (C4) Reallocated Event Count warning with these values: current=100, worst=100, threshold=0, data=4. Which program and test to trust?
 

Moribund

Distinguished
Feb 27, 2014
177
1
18,715
Generally the best thing is to run a manufacturer's long test, even though it will take some time. Quicker and a bit less reliable way comparing your drive to other hard drives of the same model is this: Download Speedfan here http://www.almico.com/sfdownload.php . Bring up your HD in its S.M.A.R.T. tab and post a screenshot on Imgur. 2 Run extended online scan option and post the screenshot as well. Post links to both here. You could also post links to images of your chkdsk log from Event Viewer here
 

andrei1237

Prominent
Jun 9, 2019
7
0
510
These are screenshots from Speed Fan, Windows chkdsk and HD Tune. My hard disk is Hitachi:

CBVED71.png


P99agKj.png


X3YaEDl.png
 

jason201

Prominent
Feb 20, 2018
231
8
765
Both apps are actually showing you the same results!
On speedfan, you have to look at the RAW VALUE (which also shows the "reallocated event count" attribute as 4), the current field means nothing. But to sum it up, a drive with bad/reallocated sectors should not be trusted, and I'd advise you to backup your data ASAP and get it replaced! (Also, do keep in mind you were lucky here, as in some cases, HDDs/SSDs can just go all out without warning!) Always keep regular backup, and don't take any drive's reliability for granted.
 

jason201

Prominent
Feb 20, 2018
231
8
765
Why the RAW value data is so important? What does it represent and what the attribute 4 means?
Because it represents the actual data. The attribute is not 4 (the value of it is).
Each drive has a limited number of spare sectors (data storage units) and a reallocated event count of 4 means 4 bad sectors were replaced with spare ones. Generally speaking, a drive with bad sectors is not to be trusted.
 

USAFRet

Titan
Moderator
How these 4 bad sectors are replaced with spare ones? Does that mean the data on these 4 bad sectors is moved to new health sectors?
Yes, probably.
The drive firmware has marked those as Do Not Use.
Whatever data may have been in those locations has been moved. The drive does this on its own.

However...the number of 'bad sectors' never goes down, always up. Eventually, there will be no more spares.
 

Moribund

Distinguished
Feb 27, 2014
177
1
18,715
This doesn't mean, of course, that your drive will imminently fail. I had many drives with a much higher raw count lasting years. Used them sparingly, never filling them up "to the brim". However, it's not so much the number of re-allocated sectors as the fact that they began to appear, that marks a serious issue, so, (and as others have pointed out), a backup of all your data is an absolute must, preferably to an SSD. This here is a very good option, I have built a number of systems with Samsung's EVO SSDs, and I haven't yet seen one of them fail https://www.newegg.ca/samsung-860-e...GmgJEWOr2Sif3HuUVXMya038JA0vBrrvbgaAnLM8P8HAQ. To monitor that drive - use a program called Samsung Magician.

2. As long as you have a copy of everything on that drive on an SSD or another drive, you can keep using it. But monitor the raw count regularly, and If you see it go up significantly over a short period of time, - replace it.
 
The drive hasn't yet replaced ("reallocated) any problematic sectors, nor are there any sectors which are "pending" reallocation. There is a record of 4 reallocation "events", but this probably means that 4 previously pending sectors were retested and returned to service. A drive's firmware may decide to preemptively replace a problematic sector, but only if it can reread it without error. This probably happens during background scanning (that's the noise you hear when the drive is idle). Otherwise the drive needs to wait for a signal from the host that a particular sector's contents are no longer needed. This happens when the host writes new data to the "pending" sector. The drive can then transparently reallocate a new, spare sector to the target LBA and retire the old sector.
 

Moribund

Distinguished
Feb 27, 2014
177
1
18,715
I would download Hitachi DFT (drive fitness test program) and run it, since the manufacturer may be aware of peculiarities of their drive other generic tests aren't even covering (currently owned by WD however ;-) ). S.M.A.R.T. scans are not entirely accurate and not all attributes have good predictive capacity for failure. Worst SMART attribute you should be concerned about is 0xC6 or 198 because it has great predictive capacity. Google study found that at the first 60 days the drive which had uncorrectable error was 39 times (!) more likely to fail than a drive without such error. Other attributes with good predictive capacity of failure is 196 (0xC4) and 0xC5 (197). You have to realise though, that the irony of this is that there are so many reasons why a drive can fail which aren't even reflected by these scans (such as certain mechanical or electronic failures which have nothing directly to do with the surface of the drive). So regardless of what the S.M.A.R.T. scan shows - ALWAYS BACK UP YOUR DATA AT ALL TIMES is always the most useful heuristic ;-)