Current Pending Sector Count

[Joe Kotroczo]

Hello,

I assume when it says "Warning" in "Current Pending Sector Count",
that's a bad sign, right?

It's a WD30EZRX, the actual values are: Normalized: 200, Worst: 66,
Threshold: 0, Value: 63 Sectors.

Reason enough to send it in? It's under warranty until mid 2014.

Thanks.

 

[VanguardLH]

I assume when it says

And "it" is WHAT?

"Warning" in "Current Pending Sector Count",
that's a bad sign, right?

Is the pending count staying the same, going up, or going down (after
you've used the disk for awhile to generate writes)? What *is* the
count? What does your unidentified SMART tool say about the status of
this attribute?

Controller timeouts can cause errors and you'll have pending errors
until the bad sectors get remapped. The pending bad sector doesn't get
remapped until the next time it gets written.

http://en.wikipedia.org/wiki/Self-M...ng_Technology#Known_ATA_S.M.A.R.T._attributes
See ID = 197, Current Pending Sector Count

It's a WD30EZRX, the actual values are: Normalized: 200, Worst: 66,
Threshold: 0, Value: 63 Sectors.

Does the value of 63 remain constant? Is it going up? Is it going
down? Have you defragged all the drives on the hard disk to see if
later reads (that would retest if the sectors are still considered bad)
and writes (at which point the remaps are done) make the value go down?

http://kb.acronis.com/content/9133
"current count of unstable sectors (waiting for remapping)"

The remapping isn't done until the next time that sector is written. So
you need to do something to write that sector to see if it still tests
bad and to write to it to initiate the remapping to the spare sectors.
A defrag might touch (refresh by write) that suspect sector but it might
not. A disk utility, like SpinRite or HDD Regenerator (both payware),
that has a refresh function where it reads the sectors, temporarily
moves then (and updates the file table), tests the sector, and then
writes the data back to that sector (if it tested okay), will result in
writing to that sector. Restoring a partition or disk image will write
all those sectors. All magnetic media experiences dipole stress so the
field strength wanes over time. Old hard disks used iron oxide (rust)
but newer disks use cobalt alloy to reduce or slow demagnetization. As
the dipoles relax, they rotate and this means it is more difficult to
differentiate them. Refreshing realigns the dipoles. When this pending
sector remapping count spikes up, it should eventually go down if the
remapping is working. If the count stays the same over time or walks
upward, the remapping isn't working because there are no more spare
sectors.

It's a *pending* count. If later reads of the suspect sector are okay
then this count goes down. If still bad, the count goes down when the
sector is written since that's when the bad sector gets remapped.

While many users will remake that any non-zero value indicates the
device is failing, that's not true. All mechanical devices will wear so
remapping compensates for that. ALso, rare few users do a refresh
(read/write) on every allocated sector in every partition on their hard
disk(s). They may not touch some files for many years during which
dipole stress results in reduced field strength that can cause read
errors. There are many files, even system (OS) files, that only get
read and never rewritten.

The pending remap count is listed read errors. It doesn't result in
testing the sector by reading, writing to elsewhere, and writing back to
refresh that sector. Unless you use utilities to force a refresh of
each sector (or, at least, the bad ones), the count won't go down until
sometime later when you or some process happens to write to that
previously flagged suspect sector.

Reason enough to send it in? It's under warranty until mid 2014.

Have you yet ran the disk maker's diagnostic utility on the device?

 

[Rod Speed]

I assume when it says "Warning" in "Current Pending Sector Count", that's
a bad sign, right?

Yes, those are sectors that are a problem.

It's a WD30EZRX, the actual values are: Normalized: 200, Worst: 66,
Threshold: 0, Value: 63 Sectors.

It's the last one that matters and that's a very bad number of pending
sectors.

 

[Rod Speed]

Joe Kotroczo wrote

And "it" is WHAT?

Is the pending count staying the same, going up, or going down
(after you've used the disk for awhile to generate writes)?

The drive has a problem regardless of whats happening there.

What *is* the count?

He told you.

What does your unidentified SMART tool say about the status of this
attribute?

Doesn't matter, the raw numbers are what matter.

Controller timeouts can cause errors

Not pending sectors they cant.

and you'll have pending errors until the bad sectors get remapped.

And if they have to be remapped, the drive has a problem.

The pending bad sector doesn't get remapped
until the next time it gets written.

http://en.wikipedia.org/wiki/Self-M...ng_Technology#Known_ATA_S.M.A.R.T._attributes
See ID = 197, Current Pending Sector Count

Does the value of 63 remain constant? Is it going up? Is it going down?

The drive still has a problem whichever its doing in that regard.

Have you defragged all the drives on the hard disk to see if later
reads (that would retest if the sectors are still considered bad) and
writes (at which point the remaps are done) make the value go down?

That's a stupid way using a drive with that many pending sectors.

And even if it does make it go down, they just become remapped
sectors. The drive still has a problem needing that many remapped.

http://kb.acronis.com/content/9133
"current count of unstable sectors (waiting for remapping)"

The remapping isn't done until the next time that sector is written.
So you need to do something to write that sector to see if it still tests
bad and to write to it to initiate the remapping to the spare sectors.

The drive clearly still has a problem if that many are pending.

A defrag might touch (refresh by write) that suspect sector but it might
not.

In fact **** all files will be moved so it's a silly approach to take.

A disk utility, like SpinRite or HDD Regenerator (both payware),
that has a refresh function where it reads the sectors, temporarily
moves then (and updates the file table), tests the sector, and then
writes the data back to that sector (if it tested okay), will result in
writing to that sector.

But all that does is remap the pending sectors. The
drive still has a problem with that many remapped.

Restoring a partition or disk image will write all those sectors.

Not necessarily. The reorganisation involved in
restoring may well see the later sectors not written
to at all depending on how full the drive is.

All magnetic media experiences dipole
stress so the field strength wanes over time.

But you don't get anything like that many
pending sectors on drives that old normally.

Old hard disks used iron oxide (rust) but newer disks
use cobalt alloy to reduce or slow demagnetization.

So that wont be the problem.

As the dipoles relax, they rotate and this means it is more difficult to
differentiate them. Refreshing realigns the dipoles. When this pending
sector remapping count spikes up, it should eventually go down if the
remapping is working. If the count stays the same over time or walks
upward, the remapping isn't working because there are no more spare
sectors.

Sectors arent regularly remapped with modern drives that don't have a
problem.

It's a *pending* count. If later reads of the suspect sector are okay
then this count goes down. If still bad, the count goes down when
the sector is written since that's when the bad sector gets remapped.

And he shouldn't be seeing that many pending sectors with a modern drive.

While many users will remake that any non-zero
value indicates the device is failing, that's not true.

Yes, but that's FAR more pending sectors than normal and so there is a
problem.

All mechanical devices will wear so remapping compensates for that.

Pending sectors arent due to wear.

Also, rare few users do a refresh (read/write) on every
allocated sector in every partition on their hard disk(s).

Because that isnt necessary with modern drives.

They may not touch some files for many years

That wont be the problem in his case because the drive is still in warranty.

during which dipole stress results in reduced
field strength that can cause read errors.

That wont be the problem in his case because the drive is still in warranty.

There are many files, even system (OS) files, that only get read and never
rewritten.

That wont be the problem in his case because the drive is still in warranty.

The pending remap count is listed read errors. It doesn't result in
testing the sector by reading, writing to elsewhere, and writing back
to refresh that sector. Unless you use utilities to force a refresh of
each sector (or, at least, the bad ones), the count won't go down
until sometime later when you or some process happens to write
to that previously flagged suspect sector.

And a modern drive with that many pending sectors has a problem.

Have you yet ran the disk maker's diagnostic utility on the device?

No point, he knows its got far too many pending sectors.

 

[VanguardLH]

The drive has a problem regardless of whats happening there.


He told you.


Doesn't matter, the raw numbers are what matter.


Not pending sectors they cant.


And if they have to be remapped, the drive has a problem.




The drive still has a problem whichever its doing in that regard.


That's a stupid way using a drive with that many pending sectors.

And even if it does make it go down, they just become remapped
sectors. The drive still has a problem needing that many remapped.



The drive clearly still has a problem if that many are pending.


In fact **** all files will be moved so it's a silly approach to take.


But all that does is remap the pending sectors. The
drive still has a problem with that many remapped.


Not necessarily. The reorganisation involved in
restoring may well see the later sectors not written
to at all depending on how full the drive is.


But you don't get anything like that many
pending sectors on drives that old normally.


So that wont be the problem.


Sectors arent regularly remapped with modern drives that don't have a
problem.


And he shouldn't be seeing that many pending sectors with a modern drive.


Yes, but that's FAR more pending sectors than normal and so there is a
problem.


Pending sectors arent due to wear.


Because that isnt necessary with modern drives.


That wont be the problem in his case because the drive is still in warranty.


That wont be the problem in his case because the drive is still in warranty.


That wont be the problem in his case because the drive is still in warranty.


And a modern drive with that many pending sectors has a problem.



No point, he knows its got far too many pending sectors.

Yes, the disk has bad sectors. Once they become mapped, they're not bad
sectors anymore. So if you get them remapped and no more show up then
how is the disk still going bad if it is NOT showing more bad sectors?

63 sectors are pending remapping (for when next write occurs on them).
32KB total disk space to get remapped on a 3TB hard disk.
66 was the worst so it has already gone down.
The threshold is 200. That threshold has not yet been met.
But you're the expect so the 200 threshold must be excessive.

I've been using hard disks that have reported pending remaps for over 8
years - as long as the remap count went down and eventually zeroed out
(but that requires actually *writing* to the suspect sector at some
time).

You're just spreading FUD to prod the OP to make a possibly non-required
hardware purchase. You sound like the FUD purveyor telling someone that
their CPU must run under 60C despite that the device has been rated for
a *continual* operating temperature of 80C.

The OP should be doing backups whether the hard disk is good, bad, or
going bad. If you don't backup then you deem your data as worthless or
reproducible - and a bad hard disk is not the only cause for losing your
data. Then the OP can continue using his hard disk despite your
doom-and-gloom declaration until the hard disk actually goes bad.

The SMART tools are only showing you the values. They are not making
any recommendations. That's up to the hardware owner to determine.
Remember that a pending sector count (to remap) of zero doesn't mean
there haven't been any remaps. It doesn't tell you how many have
already been remapped. So your declaration that the value must be zero
still doesn't say how many have already occurred. How many have already
been reallocated (to provide a sense of historical remapping but not how
many at a time or a statiscal graph) is listed as the raw value under
the "Reallocated Sectors Count" attribute.

Rather than rely on you to figure out when the hard disk is really bad,
the OP should go by what the device manufacturer encoded into their
SMART table. A threshold of 0 (zero), as the OP stated for attribute
197 (Current Pending Sector Count) means this attribute is for
information only, not to determine when the device is bad. Now look at
attribute 1 (Read Error Rate) which has, say, a threshold of 51. When
THAT value gets exceeded as per what the mfr said (not you) is when the
device is considered compromised. That's why a non-changing pending
sector remap or increasing value over time indicates a problem. The
remapping isn't happening. If the sector gets written to and the remap
count doesn't go down then it cannot be remapped. Yeah, god forbid the
device is allowed to repair itself to remain in service without further
error.

The normalized value (or just the "value") ranges from 1 to 253, so it
cannot be a larger value. Often makers set it to 100 or 200 as the
initial normalized value. Higher is better than lower. If the
[normalized] value falls below the non-zero threshold. called TEC
(Threshold Exceeded Condition), then the disk is considered defective.
Since Current Pending Sector Count has a zero threhold then this
attribute is merely a measurement and not an indicator of disk health.
The worst value is the lowest recorded normalized value.

The makers would love if every user that had a single or dozen or
umpteen sectors get remapped got scared and bought a new hard disk.
Since remapping eliminates the problem providing there isn't future
growth (that is also not reduced through remapping) is when there really
is a problem.

Current Pending Sector Count are those that still need further
investigation which is performed on a later write. Some devices provide
an offline scan procedure as part of its self-diagnostics to test the
suspect sectors after the device is idle for awhile some period of time.
The device may get triggered to run various self-diagnostic procedures.
You can hear the self-diags wandering around the surfaces by letting
your host sit idle (and with no processes doing background requests on
disk access) to know when an offline scan is running (and why some users
wonder why their hard disk is "running" when they're not doing anything
with it). The Offline Uncorrectable Sector Count, if available, is
similar to the Reallocated Sector Count attribute but shows how many
defective sectors were found during the offline scan.

The Current Pending Sector Count does *not* declare the sectors are bad.
The suspect sectors may not be found bad at all upon further testing.
Some transient condition generated a read error. The raw value of this
attribute shows how many such sectors are suspect, NOT necessarily how
many are condemned to remapping.

Until tested, the sectors tracked by Current Pending Sector Count have
NOT yet been condemned as bad sectors. They are pending further
testing. You could wait until you, some process, or the OS happens to
write to the suspect sector or when the offline scan, if your disk has
one, happens to test the bad sector or you could force a test on it now.

Whether the OP wants to believe you or me, he could rely on the disk
manufacturer's own free diag tools to determine the health of their
disk. The testing may not be as rigorous as with payware solutions
(SpinRite or HDD Regenerator) but is probably sufficient. You and I
could argue ad infinitum about what is a "bad" value for Current Pending
Sectors and why it's called a PENDING attribute and the actual status of
those tracked sectors but I'm sure the OP would probably believe the
maker's own diagnostics.

http://support.wdc.com/product/download.asp?wdc_lang=en&fid=wdsfCaviar_Green

You can download the Windows version (i.e., runs as a Windows app) or
get the bootable DOS version (and burn the .iso file to a CD-R). Run
the extended test. Could take hours, especially since it's a 3TB hard
disk.

 

[Rod Speed]

Rod Speed wrote

Yes, the disk has bad sectors.

And so many that there is certainly a problem with the drive.

Once they become mapped, they're not bad sectors anymore.

But it shouldn't have anything like that many
remapped sectors unless there is a problem.

So if you get them remapped and no more show up then how
is the disk still going bad if it is NOT showing more bad sectors?

It shouldn't have anything like that number of remapped sectors.

A few can happen with a badly designed drive that is writing
when the power fails, but that cant be the reason for that 63.

63 sectors are pending remapping (for when next write occurs on them).

And that's a hell of a lot of bad sectors.

32KB total disk space to get remapped on a 3TB hard disk.

Irrelevant to the fact that it shouldn't have anything like that many.

66 was the worst so it has already gone down.

Because pending sectors are remapped when they are written to.

The threshold is 200.

That's not a sector count.

That threshold has not yet been met. But you're
the expect so the 200 threshold must be excessive.

That isnt a sector count.

I've been using hard disks that have reported pending remaps for over 8
years - as long as the remap count went down and eventually zeroed out
(but that requires actually *writing* to the suspect sector at some time).

66 bad sectors shows that there is a problem.

You're just spreading FUD to prod the OP to make
a possibly non-required hardware purchase.

THE DRIVE IS UNDER WARRANTY.

You sound like the FUD purveyor telling someone that
their CPU must run under 60C despite that the device has
been rated for a *continual* operating temperature of 80C.

Then you need to get your ears tested, BAD.

The OP should be doing backups whether
the hard disk is good, bad, or going bad.

Irrelevant to whether the drive has a problem.

If you don't backup then you deem your data as worthless or
reproducible - and a bad hard disk is not the only cause for losing your
data. Then the OP can continue using his hard disk despite your
doom-and-gloom declaration until the hard disk actually goes bad.

Its already gone bad and it makes sense to claim on the warranty.

The SMART tools are only showing you the values. They are not making
any recommendations. That's up to the hardware owner to determine.

And he asked for comments on what to do.

Remember that a pending sector count (to remap) of zero doesn't mean
there haven't been any remaps. It doesn't tell you how many have
already been remapped. So your declaration that the value must be zero

I never ever said anything like that.

still doesn't say how many have already occurred. How many have
already been reallocated (to provide a sense of historical remapping
but not how many at a time or a statiscal graph) is listed as
the raw value under the "Reallocated Sectors Count" attribute.

66 is a hell of a lot of bad sectors and indicates that there is a problem.

Rather than rely on you to figure out when the hard disk is really bad,
the OP should go by what the device manufacturer encoded into their
SMART table. A threshold of 0 (zero), as the OP stated for attribute
197 (Current Pending Sector Count) means this attribute is for
information only, not to determine when the device is bad.

Irrelevant the FACT that 66 is a hell of a lot of bad
sectors and indicates that there is a problem.

Now look at attribute 1 (Read Error Rate) which has, say, a threshold of
51.
When THAT value gets exceeded as per what the mfr said (not you) is when
the device is considered compromised. That's why a non-changing pending
sector remap or increasing value over time indicates a problem.

The fact that it doesn't change JUST indicates that those sectors havent
been written to.

The remapping isn't happening.

Because the sectors arent being written to.

If the sector gets written to and the remap count doesn't go
down then it cannot be remapped. Yeah, god forbid the device
is allowed to repair itself to remain in service without further error.

Completely off with the ****ing fairys, as always.

The normalized value (or just the "value") ranges from 1 to 253,
so it cannot be a larger value. Often makers set it to 100 or 200
as the initial normalized value. Higher is better than lower.

Irrelevant the FACT that 66 is a hell of a lot of bad
sectors and indicates that there is a problem.

If the [normalized] value falls below the non-zero threshold. called TEC
(Threshold Exceeded Condition), then the disk is considered defective.
Since Current Pending Sector Count has a zero threhold then this
attribute is merely a measurement and not an indicator of disk health.

Irrelevant the FACT that 66 is a hell of a lot of bad
sectors and indicates that there is a problem.

The worst value is the lowest recorded normalized value.

Irrelevant the FACT that 66 is a hell of a lot of bad
sectors and indicates that there is a problem.

The makers would love if every user that had a single or dozen or
umpteen sectors get remapped got scared and bought a new hard disk.

THE DRIVE IS UNDER WARRANTY.

Since remapping eliminates the problem providing there isn't
future growth (that is also not reduced through remapping) is
when there really is a problem.

The FACT is that 66 is a hell of a lot of bad
sectors and indicates that there is a problem.

Current Pending Sector Count are those that still need
further investigation which is performed on a later write.

Irrelevant the FACT that 66 is a hell of a lot of bad
sectors and indicates that there is a problem.

Some devices provide an offline scan procedure as part of its
self-diagnostics
to test the suspect sectors after the device is idle for awhile some
period of time.

That drive doesn't.

The device may get triggered to run various self-diagnostic procedures.
You can hear the self-diags wandering around the surfaces by letting
your host sit idle (and with no processes doing background requests on
disk access) to know when an offline scan is running (and why some users
wonder why their hard disk is "running" when they're not doing anything
with it).

That drive doesn't.

The Offline Uncorrectable Sector Count, if available, is similar
to the Reallocated Sector Count attribute but shows how many
defective sectors were found during the offline scan.

Utterly mangled all over again.

The Current Pending Sector Count does *not* declare the sectors are bad.
The suspect sectors may not be found bad at all upon further testing.
Some transient condition generated a read error. The raw value of this
attribute shows how many such sectors are suspect, NOT necessarily how
many are condemned to remapping.

Irrelevant the FACT that 66 is a hell of a lot of bad
sectors and indicates that there is a problem.

Until tested, the sectors tracked by Current Pending Sector
Count have NOT yet been condemned as bad sectors.

But most of them will be.

They are pending further testing. You could wait until you, some
process, or the OS happens to write to the suspect sector or when
the offline scan, if your disk has one, happens to test the bad sector
or you could force a test on it now.

Or you could have enough of a clue to claim on the warranty.

Whether the OP wants to believe you or me, he could rely on the disk
manufacturer's own free diag tools to determine the health of their disk.

Or he could just claim on the warranty.

The testing may not be as rigorous as with payware solutions
(SpinRite or HDD Regenerator) but is probably sufficient.

Or its in the manufacturers interest to minimise warranty claims.

You and I could argue ad infinitum about what is a "bad" value
for Current Pending Sectors and why it's called a PENDING
attribute and the actual status of those tracked sectors but I'm
sure the OP would probably believe the maker's own diagnostics.

You have no basis for that surety.

http://support.wdc.com/product/download.asp?wdc_lang=en&fid=wdsfCaviar_Green

You can download the Windows version (i.e., runs as a Windows app)
or get the bootable DOS version (and burn the .iso file to a CD-R). Run
the extended test. Could take hours, especially since it's a 3TB hard
disk.

Waste of time. The FACT is that 66 is a hell of a lot
of bad sectors and indicates that there is a problem.

 

[Joe Kotroczo]

And "it" is WHAT?

"It" is the Disk Utility in Ubuntu, which is what I had lying around on
a USB stick. I assume it's just a GUI for smartmontools.

Is the pending count staying the same, going up, or going down (after
you've used the disk for awhile to generate writes)? What *is* the
count?

63, as I wrote. Uncorrectable Sector Count is 53, Write Error Rate is
117, Read Error Rate is 1863

Controller timeouts can cause errors

I think I can safely rule out the controller, the HD is one of 4
identical ones bought at the same time, and the other 3 all have 0
Pending Sectors and 0 Uncorrectable Sectors.

(...)

Does the value of 63 remain constant? Is it going up? Is it going
down?

I'll keep an eye on it. I assume I would have to actually use that disk
though to see any changes.

Have you defragged all the drives on the hard disk to see if
later reads (that would retest if the sectors are still considered bad)
and writes (at which point the remaps are done) make the value go down?

What do you mean by "drives on the hard disk"? I've erased the
partitions that were on the disk and repartitioned it for testing
purposes. And written about 260 GB of data onto, again as a test. Took 7
hours. I guess it would take me 3 days to fill the entire HD with data
at the current speed. Which is what pushed me to look into this HD in
the first pace, in my opinion it should take about 1 hour to copy 260
GB, not 7 hours.

(...)

Have you yet ran the disk maker's diagnostic utility on the device?

Tried putting WDs .iso onto a USB-stick, it does boot but does not
recognize any HDs. Don't have a DVD or CDROM drive. Might have to buy
one. Or give putting DOS on a stick another try. Weird, I haven't had to
touch DOS in over 15 years. Why is it that manufacturers only release
their diagnostic tools for DOS? I would have thought it was dead and
buried by now.

 

[Keeper]

I couldn't resist to what i was reading from VanguardLH and as so i've decided to register in here just to prove that almost everything that you've said (aswell as your advices) are a complete bs.

I've got a Samsung Spinpoint F4; Model: HD322GJ - which if you read the reviews about it, it's a really good hard disk (great performance; single plated to minimize issues; etc) - and this is my story:

1) I started to notice some strange slowdowns and freezes to which i've decided to run chkdsk with /f so to see what was going on...and there was: there where a lot of orphaned files. Anyway, i booted Windows and continued to do what i was doing, thinking that everything was sorted out. But after a few (30mins maybe), my problem returned - and now i had an EXTREME SLOWDOWN on Windows with all sort of issues: freezes; not waking from sleep; you name it.

2) So, i've ran chkdsk with the /f /r flags and more corrupted files appeared...but it continued to report 0 bad sectors...i've additionally ran an sfc /scannow so to correct any issues that windows might had...but it didn't find anything. Anyway, after all this and after a few minutes working on my PC, the extreme slowdown returned.

3) Then i've ran HDD Tune and it reported 64 current pending sectors and 4 Ultra DMA CRC Errors. The UMDA CRC errors where probably due to some OC and/or something that i did in the past since this value never increased anymore. Temp: 28ºC (82ºF). Surface test: 0 bad blocks.

4)Being puzzled by all this and tired for not being able to work decently on my PC, i decided that it was enough and ran MHDD to test the surface: had only about 200 sectors with "<10 ms" and around 12 with "<50ms"...every other sectors were "<3ms". Additionally i've ran some smart tests and all passed. Finally, i've checked the sata cable and it was in perfect conditions (it is a L shaped connector).

5) Not happy with the results, i've made a backup to some other HDD and i've low level formatted the Samsung HDD. checked for bad sectors and or anything else: The only result that stood still was the 4 UDMA CRC errors, which is normal since it's saved on the HDD firmware for historical MFG analysis, i believe.

6) Finally, i've restored the backup (cloned it) to the Samsung HDD. It worked flawlessly for 3 days...and then...the pending sectors returned!...even worse then before: now i had 92...and what about bad sectors? nothing, zero, nada, niente... and what about the surface scan? result: 0 bad blocks...and slowdown? ha! that returned and it was terrible. Firefox crashing /IE crashing; blue screens; idk...so many stuff was breaking apart that i lost count.

7) I've ran HDDTUNE benchmark test and it was giving me some really wierd results: throught the test there were some peaks as low as 1,8 MB /sec - and yes, i disabled everything: AV; search indexer; closed all apps.

Anyway, i got tired of all this and i've prepared an RMA (it will go on this next Monday to Samsung) because this ain't normal...at ALL. Sectors are pending but none is being relocated...and the surface scan is clean...smart error count continues to be 4...low temps...no way...no way your explanation makes sense. What makes sense is what Rod Speed said.

I had to post this because you are encouraging people to stick with their defective drives when you should be doing the opposite: Advice people to RMA ASAP, because your drive/info on it, will "die" soon, for sure.

 

[westcoastguy]

Old thread, nonetheless, most if not all manufacturers regularly take RMA drives, zero them out a few times and ship them back out as re-certified drives that will have the same problems of reappearing pending sectors. The OP could certainly sent his drive out for warranty buy has very high chances of receiving a replacement drive in same or worse condition.
All the shipping at the OP's expense.

This is why when I purchase a new hdd, I purchase from a shop that offers a full NEW replacement warranty for a modest price. I just picked up an external 4TB Seagate that if it went bad under the warranty period I'd send it jn at my expense and play Russian roulette, or shell out an extra $23 (NCIX) for a 3 year no hassle NEW replacement drive should this one fail. Well worth the money.