Read Verify Sector Ext error!

[Doc Savage]

I just formatted a 1 TB WD Green SATA drive and ran, or tried to run WD LifeGuard diagnostics tests on it.

It passes the SMART test but the quick test and extended test both fail quickly with this message: Read Verify Sector Ext error!

After this happened the first time I tried a chkdsk, but am still getting this message.

What does this mean?

Thanks.

 

[Paul]

I just formatted a 1 TB WD Green SATA drive and ran, or tried to run WD LifeGuard diagnostics tests on it.

It passes the SMART test but the quick test and extended test both fail quickly with this message: Read Verify Sector Ext error!

After this happened the first time I tried a chkdsk, but am still getting this message.

What does this mean?

Thanks.

That's actually the name of a single disk drive command.

"The Read Verify Sector(s) Ext verifies one or more sectors
on the device. No data is transferred to the host.

The difference between the Read Sector(s) Ext command and
the Read Verify Sector(s) Ext command is whether the data
is transferred to the host or not.

If an uncorrectable error occurs, the Read Verify Sector(s) Ext
will be terminated at the failing sector."

The command would accept a length, and all the sectors starting
at a certain LBA would have their CRC status checked.

*******

Being a green drive, it probably has aggressive power reduction
strategies. Including spinning down or doing other stupid
things, while a test is running. Perhaps the WD utility
does not take the aggressive behavior into account. Maybe the
drive needs to be "tickled" until the command comes back ?

*******

You can run your own test. Using HDTune, read the disk
from end to end while in Windows ("Error scan"). The disk
does not need a partition on it, as this kind of testing
considers the drive at the block level.

http://www.hdtune.com/files/hdtune_255.exe

I've had one Seagate drive here, that behaved better
after only one end-to-end error scan. You can alternate
an end-to-end error scan, a write operation from end
to end, and then go back and run the WD Test.

But if the problem is a conflict between the WD
test program, and the behavior of the drive, then
it may continue to report errors.

I suppose the fact you can't format it, is also
a bad sign.

The HDTune has a benchmark tab, which does a read operation
at incrementing offsets. In the interest of time, it doesn't
read the entire disk. If you see "sustained" performance
dips, like ten percent of the drive surface reports 1MB/sec
transfer rate, that means the drive is pretty sick. It's OK
for the transfer rate curve to have the odd spike in it,
because the OS can interfere. In particular, the latest OS,
Windows 8, can start doing absolutely stupid things in the
middle of your HDTune run, and spoil the test results. Earlier
OSes like WinXP, are a little better in that regard. In Windows 8,
towards the end of the test, my disk drive reported 1MB/sec transfer
rate, but it was because one of the Windows 8 maintenance tasks
started processing the drive at the same time I was benchmarking
it (causing the heads to move back and forth to support
both programs). I had to think about it a bit, before deciding not to panic.
That it was the OS screwing me, and not the disk drive company :-(

If you need to write the disk from end to end, you can
try "dd". For information sources, it has a "bucket of
zeros" in the form of /dev/zero and a "bucket of random
numbers" in the form of /dev/random. You can use either
of those to fill the disk with data. The command will promptly
stop on the first error it encounters. You won't get
to the end of a sick disk. One would hope, the power
management features don't cause the stupid green drive,
to spin down while it is being written.

http://www.chrysocome.net/dd

The following command over-writes all of the second disk
in Disk Management. The second disk is Harddisk1. And
Partition0 references the entire disk starting at sector 0.
With no additional parameters, such as bs (block size) and
count, the command runs at roughly 13MB/sec. If you specify
bs=262144 or a quarter megabyte, it'll run faster. For
example, this command should run around 39MB/sec.

dd if=/dev/random of=\\?\Device\Harddisk1\Partition0 bs=262144

It's not exactly SpinRite, but you may find the drive
behaves a bit different after it's had a read from
end to end and a write from end to end. Maybe there
is really nothing wrong with it, and the WD test
just blew it (didn't notice the drive had gone to sleep).

But your failing format attempt, that can't be good.

Paul

 

[Doc Savage]

I suppose the fact you can't format it, is also
a bad sign.

No, it formats - shows up in windows and in Disk manager. What I was sayingis I tried the WD diagnostic are doing a format and then again after trying a CHKDSK after doing the format. It's my understanding a full format doesa chkdsk by default but figured I'd try it. I haven't written any files toit since formatting since I don't trust it at this time.

This WD Green used to be an external drive but after the interface failed Ipulled it out and have been using it as a regular SATA drive for storage.

 

[Doc Savage]

That's actually the name of a single disk drive command.

Dunno, that's the message/error I'm getting. I tried the same test on the WD 320gig C: drive and it went fine.

Being a green drive, it probably has aggressive power reduction
strategies. Including spinning down or doing other stupid
things, while a test is running. Perhaps the WD utility
does not take the aggressive behavior into account.

I'm not sure if this drive spins down when unused but I don't think that's the problem since the tests both failed shortly after starting, like maybe 10 seconds.

 

[Doc Savage]

The HDTune has a benchmark tab, which does a read operation
at incrementing offsets.

Benchmark failed immediately, test wouldn't run.

You can run your own test. Using HDTune, read the disk
from end to end while in Windows ("Error scan"). The disk
does not need a partition on it, as this kind of testing
considers the drive at the block level.

So far it's about halfway through, the first block shows as bad and then another a few down. Anything that can be done about that?

 

[Flasherly]

So far

it's about halfway through, the first block shows as bad and then
another a few down. Anything that can be done about that?

--

Maybe you want as close to the WD factory ROM methodology/routine for
reinitializing sectors (correctional identification, realignment,
however accounted). When I've run these sorts of things in the past,
it's always been a major pain, takes forever, for something just as
likely to continue along, reporting symptomatic errors if not eminent
fault failure behavior indications.

Usually, too, on somebody else's and not my computer. Not that I
haven't lost a drive, maybe two, though I can't recall a second one,
offhand. (Only "bad" drives I've had I found on CompUSA/BestBuy
sales.) Researched drives, invariably from NewEgg, from Seagate to a
spate of larger Samsung, I take care of -- defragmenting, structurally
for accounting partitions and usage, along with adequate cooling. I
have, I think, only one decent Western Digital now (600M class and
another NewEgg item). So happens those "discounted" failed HD were
Western Digital -- and both failed like clockwork within nearly
eclipsing a year and their warranty period.

Total friggin' nightmare.

Not that I liked my last failure, from 3, I guess, 200/250 class
Seagates I bought at 7200RPM;- consider, though, they're 10+
year's-old, simply run hot, but I nevertheless run them hard for
boot-class and not storage drives. Joyous beaters for the most -- 111F
on the Seagate 250G across the room with the side cover off, case
laying on its side, a 120mm front-panel fan and an empty space between
the 250 and a couple terabyte+ -sized Samsungs. (The one that failed
was up to 120F+ in an enclosed, assembled case, hence without a side
panel pulled off.)

Not sure why another, similar 250G Seagate model in this Antec
aluminum Lanboy, standing upright though with side panel removed, runs
cooler, closer to 105. Weird.

Drives in ABS-plastic boxes, sold as such, for external "storage"
purposes give me shivers up the spine -- but, hey, may as well check
it out, anyway...

http://www.ehow.com/how_5356208_perform-digital-ide-hard-drive.html

 

[Paul]

Benchmark failed immediately, test wouldn't run.




So far it's about halfway through, the first block shows as bad and then another a few down. Anything that can be done about that?

OK. Since formatting works, the MBR (sector 0) is
probably good. Even though the first block in your
scan is red, maybe the MBR isn't the sector that
is affected. If the MBR was bad, the drive would
only be useful as a door stop.

Let the scan continue, until it finishes.

Then, look to see how large of a chunk is available,
with no errors in it. Record the percentages of
disk before and after that area. You'll then want
to arrange partitions (both fake and real partitions),
to make a real partition overlap the good area. Once
the real partition is in place, you can remove the fake
partitions (without formatting them, RAW mode). That
will leave unallocated areas on either side of the
good area.

In other words, if you make a partition in a bad
section of the disk, don't format it. Disk management
should say it is RAW. When you place the real partition
over top of the "green blocks" area, you'll be formatting
that, as that is supposed to work. Later, you go to
Disk Management again, and you can remove the fake
(RAW) partitions, as they're not needed. And you
really don't want the disk going there if possible.

And that's strictly a desperation move. Until you can
buy a replacement. I don't trust a drive that far gone,
for anything. Might be suited to usage as a scratch
drive, and then I lose at most a day's work.

*******

Reallocation on IDE disks is automatic, and can't be
reset by the user. The factory can reset the spares
table, but I can't. When the drive runs out of spares
to use in a bad area, that's when CRC errors finally
show up (blocks that can't be repaired by the usage
of a spare).

Using HDTune, you can examine the SMART data from
the drive, and see what percentage of life is left
in the "Reallocated Sector" line in the table.

This is data I collected over three successive days.
The drive showed symptoms of having a problem, so I checked
SMART. In the last line there, the reallocations
have finally affected the percentage of life left.
That drive was taken out of service at that point.
Every time I did a major set of writes to it, the
"Data" value would grow. While the drive still works,
I have no way to predict when it might not start.

Current Worst Threshold Data Status
Reallocated Sector Count 100 100 36 0 OK
Reallocated Sector Count 100 100 36 57 OK
Reallocated Sector Count 98 98 36 104 OK

I've also had drive problems, where "Data" stayed
at zero (good). SMART is not guaranteed to catch everything.
I expect when you check the SMART tab (after allowing
the currently running scan to finish), that row
in the table will be in horrible shape With the
conditions above, I was still "100% green blocks" in the
error scan. You've got red blocks showing, so you
must be a bit low on available spares to use. Maybe
that line in the table will be yellow or red in color.

Paul