Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I am currently testing many hard drives (with gsmartcontrol). My goto, rule of thumb of detecting a defective hard drive is to look at the Reallocated Sector Count and the Current Pending Sector Count.

However, with one hard drive, the Reallocated Sector Count and the Current Pending Sector Count read 0, but another metric, Soft Read Error Rate is highlighted and reads 4.

How do I interpret such data? What are soft read error rates, and with them, are hard drives considered defective?

share|improve this question
    
Have you rebooted since you saw it? Try shutting down (actually powering off) and then see what the value is the next time you start up. Depending on how your drive’s manufacturer implemented that value, it may be reset to zero. –  Synetech Aug 29 '12 at 0:09
    
@Synetech I ran a hard drive test, and it was reset to zero. –  edmastermind29 Aug 29 '12 at 1:57
    
That makes sense. It indicates that it received a request, tried to read, couldn’t, reported to the OS, and tried again. If there’s no change to the other values you listed, it obviously managed to read it when it tried again. That could mean that the sector is bad (especially if it happens several times—e.g. four), in which case it will eventually relocate it, but if it was once or twice, it could have been explained by other means and likely ignored safely. If you did a full test and it came up clean, then it was probably something like a dip in the power-supply or humidity or cosmic rays. –  Synetech Aug 29 '12 at 2:45
    
Other than suggesting cosmic rays ;), good info....thanks. –  edmastermind29 Aug 29 '12 at 3:32
    
Technically, cosmic rays are legitimate source of electronic errors, assuming of course that the errors are not consistent or reproducable. –  Synetech Aug 13 '13 at 19:26
add comment

2 Answers

up vote 1 down vote accepted

What are soft read error rates

Soft errors are errors that went away when your disk retried an operation. A small number might be caused by minor electrical glitches and is no direct cause for worry.

with them, are hard drives considered defective?

No. The 4 problems were transient, and probably did not last for more than a second combined.

share|improve this answer
    
+1 Short, to the point. Thanks. –  edmastermind29 Aug 29 '12 at 1:55
add comment

thats a vender specific name. my samsung has that as 0. take as look at raw read error rate, hardware ecc recovered and seek error rate. usually drives that are about to go have those numbers shooting up. I read the attributes listed a dir and listed the attributes again. on a questionable drive it often goes up by the million's of errors in just a few min.

do a benchmark on the drive when booted from another drive see if the read test has any dips in specific spots while the drive is not mounted.

its not unusual for a drive to have a stair stepped look where it starts at 50mb and dropps to about 30mb or 80 to 40. It should never have severe drops like 50 to 20 or less then back up if its not in use. I got the 80-40 mb from my 3.3 year old (29020 powered hours) drive. I have also seen bad hdd controllers and system board controllers cause it to start @5mb and occasionally spike up to 20mb

you may want to consider a full drive read scan. If you question it read test it. I have seen many a dell pass the smart and confidence test yet fail the read test. we often use smart details and benchmark results for when we know drives are going to fail but havent gotten an error code yet to get them to rma drives. Some of it is just getting to know whats normal for your drives so you can do trending.

share|improve this answer
    
+1 Good info. I did a full scan of the hard drive, and it succeeded. –  edmastermind29 Aug 29 '12 at 1:56
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.