HDD fails S.M.A.R.T. short test, but passes long test?
Note this is a >1 year old thread that was brought up by a first-post...
Well my plan to wear out that F@H inefficient computer of mine (see second one in my sig) finally paid off. One of the two hard drives, a Seagate 160GB that's easily the loudest hard disk I have used in the last 8 years, is showing errors. (I back up everything on that untrustworthy rig so I lost nothing.) Since it's only a 160GB IDE disk, I'm not going to try to save it, so it's upgrade time for the whole rig (post in "New Builds" to come).
However, it's in a strange state in that the only thing that fails is a S.M.A.R.T. short self test. The long test passes, and I was able to use badblocks to write 0x55, 0xAA, 0xFF, 0x00 to the drive without it saying any of the sectors were bad. Linux's smartctl and SeaTools also me the same result.
Here's the lengthy S.M.A.R.T. log for anyone interested:
Showing errors during use or just SMART errors? I've had a bazillion drives toss errors in SMART and never ever have problems. I disable SMART or ignore it if you can't disable it.
No errors during yet; I have to explicitly view the log to see the errors (e.g. OS and BIOS are not yelling at me).
Psssh. SMART fails. I have 3 clients whose SMART yells at them every time they boot, even with SMART turned off. Zero problems for years now.
Disable smart stuff in cmos.
I have seen smart drive show up supposed stuff that never transpired.
I also did check though with appropriate drive test software from the manufactuer of the drive.
Download the drive manufacturers specific test software
See what that discloses.
No problems with the manufacturers software? then no issues.
While you CAN ignore SMART warnings they are there for a reason and it is letting you know something isn't 100% right.
By ignoring these warnings you are taking chances with the data and the drive.
@tyreman: The manufacturer's test (Seatools) gives me the same result. I tried a short test which failed 2 / 2 times, then a long test which passed. Note that SMART was NOT tripped on the drive; I only discovered the issue because I set Ubuntu to run short tests on a schedule.
Anyway I had a 250 GB SATA drive lying around so I put that in and let the computer rebuild the raid 1 array on it. (resulting in a very strange software RAID with 1 IDE and 1 SATA drive). I didn't want to take chances with a drive containing a Subversion repo, since I read those can self-corrupt on a working disk.
Thanks for the advice / interesting to hear stories of some users tripping SMART and still losing nothing. The last two times my drive tripped a SMART warning (to the point where the BIOS is yelling), I had to reinstall my OS. ...then there was the Fujitsu drive which was eating my OS yet passing the manufacturer's test :doh:
SMART Short Test Failure
I have experienced this on my HP Pavillion All-In-One IQ846 Desktop. The log shows it is on my D (backup drive) and a search of the Web indicates that Hitachi is known to have physical failures of their product, so I assumed it was the hard drive. I am in the process of Cloning it, out of the unit, because the 846 went black a few days ago and I have to ship it back for a new motherboard. All the other tests passed. Just thought everyone would like to know they are not alone and it is not one specific drive. And this unit was a replacement for an 816 that had many problems. Good thing I bought the Extra Care Package, which allows for replacement; however, I had to go through h.... :ph34r:to get one. My advice...clone it on another drive after backup and trash the hard drive!
Wow, I'm surprised to see this thread pop up again...
I'm replying not because of thread necro (bumping up a >1 year old thread by replying to it), but because you seem to have jumped to conclusions very quickly about your HP machine and Hitachi hard drives. ANY hard drive can fail or last a decade without failing, depending on many factors other than the manufacture.
Seeing as how this thread is directly related to my issue, I will post here as well since I came upon it through the search.
I recently bought a WD2002FAEX (WD 2TB Caviar Black HDD) and the results that I have gotten are as follows when running tests with WD Data LifeGuard, HD Tune, and HD Sentinel. Should I be worried about my drive potentially dying? It looks like I am unable to retrieve any relevant SMART info at all. The extended tests on all 3 programs passed.
The status keeps alternating between fail and pass on WD Data LifeGuard.
HD Tune is not much different
And HD Sentinel is unable to get any readings from SMART.
Gigabyte X58A-UD5 Mobo
Core I7 930
Windows 7 Ultimate
Bios set to AHCI mode (same results when set to IDE)
Would these results justify returning / exchanging the drive for a new one?
|All times are GMT -7. The time now is 06:00 AM.|