Ed Cryer <ed@somewhere.in.the.uk> wrote:

I have an external 2TB Seagate HD that just suddenly failed. I'd been
using it weekly for about 7 years: one day I plug it in and Windows
tells me it needs formatting. I tried it on another PC, same thing.
I did a quick NTFS format, which took just a few seconds. Then I loaded stuff onto it. It held ok, and looked perfect. But I was a little
anxious that maybe I should have done a full format; checked every sector.

It's been formatting now for 2 hours on a 16-core laptop, and it's only
half complete.
4 hours for 2TB; 20 for 10TB; sheesh!

A couple of questions.
Am I simply wasting time with a full format? Gaining nothing?
What causes a HD to suddenly fail? I suspect the flimsy USB plug-hole
they have. I think they call it Micro USB.

I haven't had errors that a drive needed formatting, but on one of my
1TB WD Elements external drives, I often get a 'not ready' error, which
is 'fixed' by reseating the USB plugs at both ends (USB 3.0 MicroUSB at
the drive end and USB-A at the computer end). It happens mostly when the
drive has been plugged in a while ago, but has not been accessed for a
while. I think it's some kind of 'sleeping and refusing to wake up'
problem, but don't know how to prevent/fix it.

[Rewind/repeat:]

Am I simply wasting time with a full format? Gaining nothing?

Yes and yes.

A drive needing a full format to be 'fixed' is something of the 80s.

But I'm sure Paul can, and probably will, write a couple of pages on
the subject! :-)

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: NOYB (3:633/280.2@fidonet)

Ed Cryer wrote:

[snip]

A couple of questions.
Am I simply wasting time with a full format? Gaining nothing?
What causes a HD to suddenly fail? I suspect the flimsy USB plug-hole
they have. I think they call it Micro USB.

The full format probably tries to write to each logical sector, which is
why it takes so much time.

If you have the time (about a week), it may be better to use a dedicated
disk tester program - I think Seagate offer one. This could give useful information about reliability and read re-tries, so might give you
confidence in the device.

The Micro USB connectors are a disaster. I suspect they are only
designed for less than 100 mating cycles, regardless of what the specifications say! Much better to have the disc built into a NAS
housing using an Ethernet - RJ45 - connector. This also moves the power
to either a separate wall wart or an integrated supply fed from the mains.


--
Graham J

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

On 2024-12-06 12:23, Graham J wrote:

Ed Cryer wrote:

[snip]

A couple of questions.
Am I simply wasting time with a full format? Gaining nothing?
What causes a HD to suddenly fail? I suspect the flimsy USB plug-hole
they have. I think they call it Micro USB.

The full format probably tries to write to each logical sector, which is
why it takes so much time.

If you have the time (about a week), it may be better to use a dedicated disk tester program - I think Seagate offer one.ÿ This could give useful information about reliability and read re-tries, so might give you confidence in the device.

The Micro USB connectors are a disaster.ÿ I suspect they are only
designed for less than 100 mating cycles, regardless of what the specifications say!ÿ Much better to have the disc built into a NAS
housing using an Ethernet - RJ45 - connector.ÿ This also moves the power
to either a separate wall wart or an integrated supply fed from the mains.

Yes, +1 for a NAS.

--

Fake news kills!

I may be contacted via the contact address given on my website: www.macfh.co.uk


--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

On 2024-12-06 12:35, Ed Cryer wrote:

I have an external 2TB Seagate HD that just suddenly failed. I'd been
using it weekly for about 7 years: one day I plug it in and Windows
tells me it needs formatting. I tried it on another PC, same thing.
I did a quick NTFS format, which took just a few seconds. Then I loaded stuff onto it. It held ok, and looked perfect. But I was a little
anxious that maybe I should have done a full format; checked every sector.

Seagate has a hard disk test that I recommend you use instead of a full format. Name Seatools. Google for it, because the link I have is in Spanish.

Otherwise, use a tool that can read the SMART database. This is
problematic to read on an external HD, the USB chip interferes.

The error in the hard disk could have been caused by software error
because of a sudden power down, for instance. Not damaged sectors. The
test software should tell you the problem.

It's been formatting now for 2 hours on a 16-core laptop, and it's only
half complete.
4 hours for 2TB; 20 for 10TB; sheesh!

A couple of questions.
Am I simply wasting time with a full format? Gaining nothing?
What causes a HD to suddenly fail? I suspect the flimsy USB plug-hole
they have. I think they call it Micro USB.

True damage, surface damage, can be caused by vibration. Don't handle
the hard disk while it is powered up.

--
Cheers, Carlos.

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: ---:- FTN<->UseNet Gate -:--- (3:633/280.2@fidonet)

On Fri, 12/6/2024 7:23 AM, Graham J wrote:

Ed Cryer wrote:

[snip]

A couple of questions.
Am I simply wasting time with a full format? Gaining nothing?
What causes a HD to suddenly fail? I suspect the flimsy USB plug-hole they have. I think they call it Micro USB.

The full format probably tries to write to each logical sector, which is why it takes so much time.

If you have the time (about a week), it may be better to use a dedicated disk tester program - I think Seagate offer one.ÿ This could give useful information about reliability and read re-tries, so might give you confidence in the device.

The Micro USB connectors are a disaster.ÿ I suspect they are only designed for less than 100 mating cycles, regardless of what the specifications say!ÿ Much better to have the disc built into a NAS housing using an Ethernet - RJ45 - connector.ÿ This also moves the power to either a separate wall wart or an integrated supply fed from the mains.

A non-quick "format" does this:

1) Writes a new $MFT, an empty one. All the files are gone.
The rest of the surface receives no writes. A program like
Photorec can recover files from the surface. So far, our
"format" is not a forensic quality cleaning.

2) The bad block scan is a *read* operation from one end of
the new partition, to the other end. Bad clusters are removed
from circulation, and added to $BADCLUS. The data on the sector
isn't even considered. Only the CRC error flag matters to a
$BADCLUS determination. It's a scan for CRC errors. Even though this
adds relatively little of value, it takes *hours* and uses zero CPU.

To actually "erase" a drive or "freshen the surface" or
"flush pending sectors that are unhealthy", you use dd.exe .
(When the popup appears, just reload the page using the reload icon.)

http://www.chrysocome.net/download

http://www.chrysocome.net/downloads/dd-0.6beta3.zip <=== 191KB

# Usage instructions

http://www.chrysocome.net/dd

Use dd.exe --list (as administrator, in administrator terminal). Partition 0 is a shorthand for "erase the whole drive".

\\?\Device\Harddisk1\Partition0 <=== This is the address of my 250GB WD Blue dud
link to \\?\Device\Harddisk1\DR1
Fixed hard disk media. Block size = 512
size is 250059350016 bytes <=== The legal definition of "at least 250 GB"

Now, we erase the 250GB disk drive.

dd.exe if=/dev/zero of=\\?\Device\Harddisk1\Partition0 bs=8192 # Good for modern HDD hard drives
# SSD ? Maybe not as good an idea.
# For an SSD, try Secure Erase.

That will erase the drive, and remove the partition table, so in disk management, it'll just be a black block. A check with HxD hex editor,
scrolling over the hard drive surface, will show lots of zeros now
occupy the surface.

Now, you can use the HDTune bad block scanner if you want, use
the SMART tab, look for Reallocated raw data to be non-zero.

If the surface is bad, attempts to write will be slow, and you'll
have a hint the drive is no longer good.

The free HDTune does not bench the entire surface of extremely large
drives. For the small 250GB drive I'm using for this right now,
the benchmark covers the entire surface. But the bad block scanner, does
scan the entire surface of the drive, for CRC errors. It does not
edit or modify any $BADCLUS while doing so. You can tell from the MB/sec
read number in the interface, whether it is "stuck" or not.

https://www.hdtune.com/download.html

HD Tune 2.55 12 February 2008 hdtune_255.exe 628 KB

Paul




--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

On Fri, 6 Dec 2024 11:35:00 +0000, Ed Cryer <ed@somewhere.in.the.uk>
wrote:

I have an external 2TB Seagate HD that just suddenly failed. I'd been
using it weekly for about 7 years:

If it's really seven years old I'd really think about replacing it.
They don't last for ever.

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

On 12/6/2024 10:33 AM, Peter Johnson wrote:

On Fri, 6 Dec 2024 11:35:00 +0000, Ed Cryer <ed@somewhere.in.the.uk>
wrote:

I have an external 2TB Seagate HD that just suddenly failed. I'd been
using it weekly for about 7 years:

If it's really seven years old I'd really think about replacing it.
They don't last for ever.

I suggest using a low cost USB to HDD docking station to eliminate
possible connection wear problems. Also if alternating HDD for backup
purposes have two docks to eliminate frequent exchanges. See:

<https://www.amazon.com/gp/product/B08P1539VD/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1>


--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

Java Jive <java@evij.com.invalid> wrote:

On 2024-12-06 12:23, Graham J wrote:

Ed Cryer wrote:

[snip]

A couple of questions.
Am I simply wasting time with a full format? Gaining nothing?
What causes a HD to suddenly fail? I suspect the flimsy USB plug-hole
they have. I think they call it Micro USB.

The full format probably tries to write to each logical sector, which is why it takes so much time.

If you have the time (about a week), it may be better to use a dedicated disk tester program - I think Seagate offer one.ÿ This could give useful information about reliability and read re-tries, so might give you confidence in the device.

The Micro USB connectors are a disaster.ÿ I suspect they are only
designed for less than 100 mating cycles, regardless of what the specifications say!ÿ Much better to have the disc built into a NAS
housing using an Ethernet - RJ45 - connector.ÿ This also moves the power to either a separate wall wart or an integrated supply fed from the mains.

Yes, +1 for a NAS.

I don't know what type of computer Ed has, but I have a laptop and use
an external USB HDD (also) while traveling, so I have external USB HDDs
*and* a (Synology) NAS. (And the external USB HDDs are also used for
off-site backup.)

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: NOYB (3:633/280.2@fidonet)

Ed Cryer wrote:

I have an external 2TB Seagate HD that just suddenly failed. I'd been
using it weekly for about 7 years: one day I plug it in and Windows
tells me it needs formatting. I tried it on another PC, same thing.
I did a quick NTFS format, which took just a few seconds. Then I loaded stuff onto it. It held ok, and looked perfect. But I was a little
anxious that maybe I should have done a full format; checked every sector.

It's been formatting now for 2 hours on a 16-core laptop, and it's only
half complete.
4 hours for 2TB; 20 for 10TB; sheesh!

A couple of questions.
Am I simply wasting time with a full format? Gaining nothing?
What causes a HD to suddenly fail? I suspect the flimsy USB plug-hole
they have. I think they call it Micro USB.

Ed

I am a "hardware" type person so one of the first things I would do
would be to remove the drive from the case and plug it into a sata cable
to my test / gamer machine. Of course that assumes that the drive has a
sata jack, unlike some external drives that I have seen that don't have
a sata jack and the USB jack is part of the drive itself.

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

Paul in Houston TX wrote:

Ed Cryer wrote:

I have an external 2TB Seagate HD that just suddenly failed. I'd been
using it weekly for about 7 years: one day I plug it in and Windows
tells me it needs formatting. I tried it on another PC, same thing.
I did a quick NTFS format, which took just a few seconds. Then I
loaded stuff onto it. It held ok, and looked perfect. But I was a
little anxious that maybe I should have done a full format; checked
every sector.

It's been formatting now for 2 hours on a 16-core laptop, and it's
only half complete.
4 hours for 2TB; 20 for 10TB; sheesh!

A couple of questions.
Am I simply wasting time with a full format? Gaining nothing?
What causes a HD to suddenly fail? I suspect the flimsy USB plug-hole
they have. I think they call it Micro USB.

Ed

I am a "hardware" type person so one of the first things I would do
would be to remove the drive from the case and plug it into a sata cable
to my test / gamer machine. Of course that assumes that the drive has a
sata jack, unlike some external drives that I have seen that don't have
a sata jack and the USB jack is part of the drive itself.

Absolutely! Also a good idea to change the cable. Many times they go bad
and can cause some strange problems.



--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

On Fri, 12/6/2024 10:33 AM, Peter Johnson wrote:

On Fri, 6 Dec 2024 11:35:00 +0000, Ed Cryer <ed@somewhere.in.the.uk>
wrote:

I have an external 2TB Seagate HD that just suddenly failed. I'd been
using it weekly for about 7 years:

If it's really seven years old I'd really think about replacing it.
They don't last for ever.

That depends on the power-on-hours and the
operating policies of the drive (head parking,
excessively low flying height, "polish marks"
on the platter surface where the heads haven't
moved for a while).

Some modern drives, like the 20TB ones, they now
have "dither" programmed in, so when the drive is
idle, the heads do not stay on the same track for
hours on end. The heads are moved around even when
the drive has no work. In the same way the image on a
plasma TV was moved around, to prevent burn-in.

I have a drive, where the SMART indicates 55,000 power on hours.
Another poster once reported they had a drive with 70,000
hours of time on the clock. On my drive, if you bench it,
the platter has a "nearly new" rating -- very few apparent
re-allocations.

I have seen a significant reallocation count, on a drive
that had 5,000 power-on-hours. All five identical model
numbers of drive, show similar "flaky" behavior. None of the
drives have died, but the surface of the platters is obviously
seriously compromised ("rusty").

Not all drives are champs. But for a few of them, the
staff washed their hands before doing assembly :-)

You can see in my examples, that power-on-hours alone is
not the only determinant. Some of the drives I own,
are engineering failures and the staff at Seagate likely
knew every one of those shipped... was rubbish. You can also
see this in BackBlaze, where a few specific models have
terrible statistics.

All you can do, is check your SMART, check your read benchmarks,
and determine when it is time to retire a drive. The 55,000 hour
drive, is still perfectly fit for usage. No adverse SMART
indicators, good read benchmark (no dips or spikes).

It's actually pretty hard to bench drives on windows now.
I spotted some anomalous behavior, for a picture I was
going to make for Ed, so I had to bail on the mission
and toss the results. Win2K is a good OS for doing
read benchmark runs. Now, where did I leave my Win2K...

Paul

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

Paul wrote:

I have an external 2TB Seagate HD that just suddenly failed. I'd been
using it weekly for about 7 years:

If it's really seven years old I'd really think about replacing it.
They don't last for ever.

That depends on the power-on-hours and the
operating policies of the drive

Not only. I had a couple of drives dying with very small numbers of power-on-hours (less than 700), probably simply because of aging
electronic components, which age even when not being used. So replacing
a 7yo drive absolutely makes sense to me.

-jw-
--
And now for something completely different...

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: Easynews - www.easynews.com (3:633/280.2@fidonet)

Ed Cryer <ed@somewhere.in.the.uk> wrote:
[...]

The drive is now working beautifully, but I've replaced it with a brand
new 4TB one.
I have a policy which I call "exorcising failures"; when something
breaks, replace it with something better.

However, in this case I'm dithering to just throw the old one away. It
has only 23,500 usage hours, and looks good good good.

A new theory has occurred to me. Maybe it failed because of a badly
seated cable; perhaps only partially inserted on the occasion when it
blew up.
Here's a dilemma for a nervous man. It passes all the scientific tests
for reliability, but it DID once fail.

Are you sure the error message said it needed to be *formatted*, not
just to be *checked*?

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: NOYB (3:633/280.2@fidonet)

On Sat, 12/7/2024 1:23 PM, Frank Slootweg wrote:

Ed Cryer <ed@somewhere.in.the.uk> wrote:
[...]

The drive is now working beautifully, but I've replaced it with a brand
new 4TB one.
I have a policy which I call "exorcising failures"; when something
breaks, replace it with something better.

However, in this case I'm dithering to just throw the old one away. It
has only 23,500 usage hours, and looks good good good.

A new theory has occurred to me. Maybe it failed because of a badly
seated cable; perhaps only partially inserted on the occasion when it
blew up.
Here's a dilemma for a nervous man. It passes all the scientific tests
for reliability, but it DID once fail.

Are you sure the error message said it needed to be *formatted*, not
just to be *checked*?

I wouldn't toss it.

I like hard drives :-)

Paul

--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)