Current pending sector count reset

[Elixer]

After using my Samsung HD for ~ 60 hours, I was getting a crapload of 'Current Pending sector count' errors, and this was throwing SMART errors. It was over 30 as listed by the raw values.

Since I didn't want to RMA the HD with all my data on it, I decided I should just wipe it.
Once that was done, it corrected all the current pending sector count.
I now know that current pending error sector count means there was a read error, however, it don't remap these until a write is done, which the wipe did.

The question is, since it don't throw SMART errors anymore, since all the bad blocks were remapped would you still RMA the HD ?

 

[HellHawk]

I was faced with a similar situation a month or two ago. In short, yes. The peace of mind is worth it. Otherwise you will forever be wondering if there was really an issue and if it will come back.

 

[code65536]

What's the reallocated (not pending) sector count?

When a sector is "unreadable", it just means that the bits coming back from the head don't match what was expected (CRC is wrong). The drive has no way of knowing why there was a read error (at least not yet). It might be that part of the platter is physically damaged and that the head is actually trying to read a scratch. Or it might be that some random junk had been written to that part of the platter, thus causing a CRC mismatch (one common way this could happen is if there was a brief surge of power to the head as it passed over a part of the platter).

That's where writing comes in. The only way for the drive to answer that "why?" and to know if a sector is really bad (e.g., physical damage) or if it was just some junk bits and the sector is actually fine is by writing to the sector and then reading that sector back to make sure that what it read is what it wrote.

That's why pending sector counts don't go away until there is a write operation to the sector in question (which doesn't happen often, since the OS often just freaks out when there is an "unreadable" sector and doesn't try to write to it). If, after a write, the sector still looks bad, the drive will reallocate the sector, and the reallocated sector count will go up. Otherwise, the drive just shrugs and puts that sector back to duty.

Either way, the pending sector count will go down (usually to zero) if you do a drive-wide write (e.g., a low-level format). What matters is whether the actual reallocation sector count goes up or not. If it does, you've got bad sectors. If it doesn't, you might want to make sure that your system's power is clean.

 

[razel]

Awesome info! For historical sakes, I guess the definition of low-level format has changed. Not just in your post, but everywhere on the Internet.

Low-level format used to mean actually changing underlying 'structure' below the filesystem. This was back in the late 80's. By changing the interleave of an HDD you can increase it's performance. Back then if you upgraded your CPU or changed the refresh speed of your RAM, it's good idea to test for a new interleave.

 

[Elixer]

Thanks code65536, some interesting info you got there.
Trying to do a SMART test, I got

SMART Self-test log structure revision number 1
Num Test_Description Status Remaining LifeTime(hours) LBA_of_first_error
# 1 Short offline Completed: read failure 90% 76 620996481

The Reallocated Sector Count is a one now.