Samsung F1 failed.... Or maybe not ?

[nick_h]

This is odd. Vista reported a Bad Block on my system drive last Thursday.

I whipped all the data off, dropped the Samsung's Diagnostic program onto a DVD and ran it.

All passes until it started the Surface Scan which reported a Bad Block at around 2%. Stopped the scan and restarted and this time it only got as far as the random read which failed reporting Sector Errors.

Fine I thinks. I generate an RMA with Samsung and run the Samsungs Low Level Format utility on the disk to wipe it.

Decided this morning to give the diagnostic program one last run before it the disk goes and so far it's got to 57% with no errors !

Whats gives ? In my experience if a disk has a Bad Block then it has a Bad Block !

[LuckyNV]

Generally diagnostic tools have the ability to examine the surface, if any bad sectors are found, it will mark them as bad and then the HDD will ignore and never write to that sector.

I had issues with my WD Raptor, I only did a quick format because I'm such a lazy sod, anyway, installed Windows and every start up Windows kept running the disk check, after 3-4 times I decided that something was definitely wrong, ran the WD diagnostic tool, said it found errors and corrected them, bang, no more issues (drive is about 2yo now, near 24/7 use)

[Agent]

Something that people often forget is that bad sectors are *normal* on drives. No manufacturing process is perfect.
The issue is when bad sectors start rising. A hard disk has so many sectors 'spare' so that data can be stored there and not in the bad area, without altering the size of the drive.

Bad sectors after manufacture are normally marked as such, thus the user never sees them. Sometimes they can slip through though.

In short: Run a full scan (a few times if you like) and see what it reports. If the errors 'vanish' it's just the drive allocating the spare sectors to the bad ones. If they rise, RMA the drive.

Having said that if you hear anything abnormal or you just don't trust the drive, RMA it

[nick_h]

Thanks guys. In my experience when a drive starts to get bad blocks it's usually the start of the end. Acording to my Vista logs it was only one bad block .....

I've now run the full Samsung Diagnostics again including the surface scan and it say the drive has no errors.

I hope I'm not in a Catch 22 situation now ....

I've already raised the RMA on the results of the first run of Samsungs diagnostic prog. I dont want to send the drive back to them for them to do the same check and say the drive is fine.

On the otherhand I dont want them to say I cant raise another RMA later if the drive shows more symptoms of failing.

According to KillDisk the drive has only had a total of 3 months of use in the last 18 or so.

[Agent]

GRab something like Speedfan and view the SMART status of the drive. That'll contain all sorts of useful info

[Lourdes]

Did you do the full surface scan after using the low level format tool? If that's the case then the drive is fine and the bad block was probably caused by an interrupted write due to a crash etc.

[speardane]

try soak testing it whilst it has no valuable data on it.

I use an old utility that keeps looping disk tests (unfortunately no longer produced).
But various security companies (try Steganos) also have utilities for secure over-writes that will overwrite the entire disk (rather than files) with sequences of all 0 and then all 1s - pick a high value of repeats... leave it running for a while and then retest - if the number of bad blocks rises - as Agent says, then it needs RMAing...