A possible SpinRite purchase

[Info]

Hi,

Apologies if this isn't the correct section, I'm considering purchasing SpinRite and just wanted to ask a few questions.

my current configuration: Asus Dark Hero x570 + ryzen 9 5950x

C: is nvme - windows 11 - without encryption

D: HDD 22TB WD Gold - with BitLocker XTS-AES 256-bit encryption

E: HDD 22TB WD Gold - with BitLocker XTS-AES 256-bit encryption

F: HDD 24TB WD Gold - with BitLocker XTS-AES 256-bit encryption

G: HDD 26TB WD Gold - with BitLocker XTS-AES 256-bit encryption

My interest is in using SpinRite on the HDDs to check for any problems, to know if they are all 100% error-free.
Specifically, am I required to disable encryption? I wouldn't like to disable it because it takes a long time to encrypt the entire disk. (I have all the BitLocker keys saved, printed and on USB drives).
Is there any risk that SpinRite will erase any files?
Which SpinRite level should I run on the HDDs to be absolutely sure that there are no problems? And considering the size of the disks, would this take more than 24 hours?

Sorry my bad english.

Best Regards,

 

[ColbyBouma]

my current configuration: Asus Dark Hero x570 + ryzen 9 5950x

That is a modern motherboard. It may not be able to boot DOS. I recommend trying Bootable first. https://www.grc.com/bootable.htm

Specifically, am I required to disable encryption?

No. SpinRite is compatible with software encryption like BitLocker.

Is there any risk that SpinRite will erase any files?

Yes, but the risk is usually very low, and there are ways to reduce the risk if you need to. Always back up your data before performing any sort of maintenance operation.

Which SpinRite level should I run on the HDDs to be absolutely sure that there are no problems?

Level 5 is the most thorough, but takes the longest.

And considering the size of the disks, would this take more than 24 hours?

Yes, even on Level 1 or 2. The exact speed varies from drive to drive, but your drives will very likely take longer than 24 hours. Here are a couple of logs:
18 TB, 29.75 hours: https://gitlab.com/GRC-Community/Sp.../6.1/Peter_Blaise_LOGs_2023-04-20-to-28/7.LOG
20 TB, 25.5 hours: https://gitlab.com/GRC-Community/Sp.../Peter_Blaise_LOGs_2023-04-15-to-19/0 (9).LOG

 

[Info]

Yes, even on Level 1 or 2. The exact speed varies from drive to drive, but your drives will very likely take longer than 24 hours. Here are a couple of logs:
18 TB, 29.75 hours: https://gitlab.com/GRC-Community/Sp.../6.1/Peter_Blaise_LOGs_2023-04-20-to-28/7.LOG
20 TB, 25.5 hours: https://gitlab.com/GRC-Community/SpinRite-Logs/-/blob/main/6.1/Peter_Blaise_LOGs_2023-04-15-to-19/0 (9).LOG

Thanks very much for reply ColbyBouma,

One last question: does the program analyze one disk at a time, or can it analyze several disks simultaneously? I mean, is the processing time the same for one selected disk or for three selected disks?
I just want to know if the program works in parallel or if it always needs to finish processing one disk before starting another.

Best Regards,

​

 

[Bplayer]

It is one disk at a time and that is the only process that is running on the PC. As Colby mentioned that new PC will most likely not be able to boot DOS to then run SpinRite. The alternative is install a VM in a fast USB drive, boot from that and then run SR in the VM. See this thread for the details. But is is still the only process running so the PC will be out of service for other uses.

 

[peterblaise]

18 TB, 29.75 hours: https://gitlab.com/GRC-Community/Sp.../6.1/Peter_Blaise_LOGs_2023-04-20-to-28/7.LOG
20 TB, 25.5 hours: https://gitlab.com/GRC-Community/Sp.../Peter_Blaise_LOGs_2023-04-15-to-19/0 (9).LOG

;-)

Happy to be of service - and there are more on the way all the time as
I collect my SpinRite USB drives from all over.

A note of caution regarding time expectations - the 22 and 24 TB may
be SMR, taking longer still, not a 1:1 increase in times required
compared to the 19 and 20 TB CMR drives in the above SpinRite LOGs.

However, I testify that it is worth it, SpinRite having discovered that
one of my new-to-me 18 TB drives had unrecoverable surface errors,
and the vendor swapped it no question, preventing significant
headaches for me down the line.

So, fully test with SpinRite Level 5, which is 'only 3 reads and 2 writes
of every bit - if a drive cannot do that on demand, then it's not ready
for prime time..

With Data:

SPINRITE NORAMTEST LEVEL 5 DYNASTAT 1 NOREWRITE​

Without data:

SPINRITE NORAMTEST LEVEL 5 DYNASTAT 0​

Note, an old BIOS PC with SATA3 sockets is all we need, and used
one's are cheap - I throw nothing away.

Buy 2 old computers to test 2 drives at once - I use a KVM switch
between 'em.

I have dozens of old computers just for SpinRite - of course I do.

That NVMe can even be tested with SpinRite via USB2 on an old
BIOS computer using a USB-to-NVMe adapter.

Has anyone had success using a virtual machine to test NVMe in their
native sockets at native speeds?

Let us know how it goes.

 

[Info]

18 TB, 29.75 hours: https://gitlab.com/GRC-Community/Sp.../6.1/Peter_Blaise_LOGs_2023-04-20-to-28/7.LOG
20 TB, 25.5 hours: https://gitlab.com/GRC-Community/Sp.../Peter_Blaise_LOGs_2023-04-15-to-19/0 (9).LOG

;-)

Happy to be of service - and there are more on the way all the time as
I collect my SpinRite USB drives from all over.

A note of caution regarding time expectations - the 22 and 24 TB may
be SMR, taking longer still, not a 1:1 increase in times required
compared to the 19 and 20 TB CMR drives in the above SpinRite LOGs.

However, I testify that it is worth it, SpinRite having discovered that
one of my new-to-me 18 TB drives had unrecoverable surface errors,
and the vendor swapped it no question, preventing significant
headaches for me down the line.

So, fully test with SpinRite Level 5, which is 'only 3 reads and 2 writes
of every bit - if a drive cannot do that on demand, then it's not ready
for prime time..

With Data:

SPINRITE NORAMTEST LEVEL 5 DYNASTAT 1 NOREWRITE​

Without data:

SPINRITE NORAMTEST LEVEL 5 DYNASTAT 0​

Note, an old BIOS PC with SATA3 sockets is all we need, and used
one's are cheap - I throw nothing away.
a
Buy 2 old computers to test 2 drives at once - I use a KVM switch
between 'em.

I have dozens of old computers just for SpinRite - of course I do.

That NVMe can even be tested with SpinRite via USB2 on an old
BIOS computer using a USB-to-NVMe adapter.

Has anyone had success using a virtual machine to test NVMe in their
native sockets at native speeds?

Let us know how it goes.

Thanks for reply,
Anyway, an old computer won't be a problem, my drives are CMR. I recently received a new WD drive that needed an RMA; I think it arrived with the head stuck.
Another detail: I'm not interested in NVMe drives, as it's only the OS drive for me, so it's not important.

My concern is my storage drives. Can I connect them to an old computer and leave them running the test? If the power goes out in the middle of the test, does that pose any risk?

SPINRITE NORAMTEST LEVEL 5 DYNASTAT 1 NOREWRITE

I just want to know if there are any defective sectors; all of them are working.
A log is generated at the end, right? If all of them pass, it means there are no errors. How often should I repeat this test? Every year?

 

[peterblaise]

Since we can lose 32,768 sectors if the electricity flickers, I'd suggest a
long-term power supply, not cheap.

Alternatively, backup first, then we can run without an uninterruptible
power supply, and possibly go faster with:

SPINRITE NORAMTEST LEVEL 5 DYNASTAT 0​

Check the SpinRite menus to make sure a LOG is on, I use the option
for brief descriptions.

Specific sectors that required data remediation will be listed, along
with the remedy applied, or all will be clear.

Over the years, if the same sectors require attention, a high-level
set-aside such as:

CHKDSK x: /R​

... may let the operating system mark them as unusable.

Otherwise, we'd need to translate the drive sectors to partition
sectors, and manually mark them unusable in a hidden Windows
NTFS $BadClus file.

Bad or slow sectors cause interminable delays when reading and
writing, and so long as the data eventually gets read or written, the
drive and the operating system are happy, even if we are not.

I wish there was a user-adjustable threshold on this, but there is not.

No worries . . . until a drive starts to 'feel' slow and
uncharacteristically unresponsive.

Hence periodic retesting to compare against the as-new-to-us test
makes sense.

Keep our original LOGS, and compare.

- - - - -

Also, we can get read-responsiveness graphs from free HDDScan
- long and precise - and especially free HD Tune - brief and
imprecise, but brief - and both work on a live drive, in case we
want to check without taking the drive offline, since both work
under Windows.

- - - - -

I'm curious if you partitioned the drives as ~16 TB - 16,777,200 MB -
plus the remainder as second partitions, both partitions compressed,
or if you partitioned them as full 22 TB and 24 TB partitions without
compression.

16,777,200 is the maximum size that NTFS will apply compression
too, and I wanted to store the maximum, hence using NTFS
compression which only works up to 16 MB less than 16,777,216 MB
( which equals 16 TB ).

I hate Microsoft and their arbitrary unanticipatory thoughtlessness.

Let us know how it goes - share LOGS!

Thanks.

 

[peterblaise]

Also, are the drives in separate servers, or in the daily use computer?

Separated on their own, they are immune to the challenges of a live
under-our-fingertips daily-use computer, frequent reboots, updated,
driver clashes, and such.

Constant fear:

"... I just formated a drive ... what letter? I though it was a thumb
drive. Did I just reformat my 20 TB backup drive?!? OH NO! ..."

For me, when my daily-use computer got a glitch, the gazillion-TB
drives in the same system took a hit, munging their partition tables,
and they then needed their data recovered, a long process,
considering their size.

I needed 2 identical of each drive, or larger, to facilitate a 1:1 recovery
by copying out what was recovered to a drive ready to receive the
massive amount of files.

But if the drives were in a networked 'server', even a cheap old BIOS
PC, then they would have been 'saved' and unharmed.

Even USB docks are safer than internal, by some degree of risk, and
if backups are slow, overnight, then that's actually OK, and we can
turn them off when not needed.

I have supplemental cooling fans all around, keeping them at room
temperature.

I'm just curious about sharing our experiences with the new
gazillion-TB drives - how are we all using them, and most
importantly, what do we do when they fail?

Thanks.

 

[Info]

Since we can lose 32,768 sectors if the electricity flickers, I'd suggest a
long-term power supply, not cheap.

Alternatively, backup first, then we can run without an uninterruptible
power supply, and possibly go faster with:

SPINRITE NORAMTEST LEVEL 5 DYNASTAT 0​

Check the SpinRite menus to make sure a LOG is on, I use the option
for brief descriptions.

Specific sectors that required data remediation will be listed, along
with the remedy applied, or all will be clear.

Over the years, if the same sectors require attention, a high-level
set-aside such as:

CHKDSK x: /R​

... may let the operating system mark them as unusable.

Otherwise, we'd need to translate the drive sectors to partition
sectors, and manually mark them unusable in a hidden Windows
NTFS $BadClus file.

Bad or slow sectors cause interminable delays when reading and
writing, and so long as the data eventually gets read or written, the
drive and the operating system are happy, even if we are not.

I wish there was a user-adjustable threshold on this, but there is not.

No worries . . . until a drive starts to 'feel' slow and
uncharacteristically unresponsive.

Hence periodic retesting to compare against the as-new-to-us test
makes sense.

Keep our original LOGS, and compare.

- - - - -

Also, we can get read-responsiveness graphs from free HDDScan
- long and precise - and especially free HD Tune - brief and
imprecise, but brief - and both work on a live drive, in case we
want to check without taking the drive offline, since both work
under Windows.

- - - - -

I'm curious if you partitioned the drives as ~16 TB - 16,777,200 MB -
plus the remainder as second partitions, both partitions compressed,
or if you partitioned them as full 22 TB and 24 TB partitions without
compression.

16,777,200 is the maximum size that NTFS will apply compression
too, and I wanted to store the maximum, hence using NTFS
compression which only works up to 16 MB less than 16,777,216 MB
( which equals 16 TB ).

I hate Microsoft and their arbitrary unanticipatory thoughtlessness.

Let us know how it goes - share LOGS!

Thanks.

Thank you very much for all the explanation,

Well, I think the safest method for me will be to make a 1:1 backup before starting. Of course, I have daily backups for critical files, but even so, I prefer to avoid any problems and make a 1:1 backup before running SpinRite.

After selecting SPINRITE NORAMTEST LEVEL 5 DYNASTAT 0 and verifying that logging is enabled, I just have to wait for it to finish, right?

Regarding partitioning, I don't partition any drives, and I'm using it on a normal computer, not a server.

I think it's safer this way: I'll remove one drive, make a 1:1 backup, and test it. I can start with the 26TB drive that I use specifically for backups. Once the 26TB backup is complete and verified, I'll use it to make the backup of the other drives, and I'll test them one by one.

I don't think I'll have any problems, but I want to be 100% sure that all my drives are in good condition and 100% healthy.

Thank you very much for all the help.

Best Regards,

 

[peterblaise]

By 'server' I mean NOT the computer under our fingertips and
keyboard daily use.

I have some 'servers' that are twin processor, duplicate power supplies,
secure everything, robust as can be, take forever to boot, have total
remote capabilities via terminal sockets, multiple networks on the
system board, the multiple fans sound like a helicopter when first
turned on, and they weight more than I do.

That's not what I mean by 'server' in this context.

I mean simple Windows network shares and use from any computer
to any other computer.

Even an old 386 with SATA adapters to PATA can be a useful 'server'.

I have a bank of USB drives on an old computer in a closet, and it's a
great and safe place for me to drag-and-drop files from my daily
beaters under my finder tips.

My daily beaters crash all the time, but the USB-drive 'server' in the
closet just trudges on, unaffected, reliable as can be.

Those 'server' computers never reboot, never update, never crash.

They are stable and 'finished' computers, built for a purpose, and
never changing.

When my main computer goes blue screen of death and whacks the
boot sector or MFT, I am so grateful that my backups reside in
another computer altogether, and are sacrosanct and 'out of the way'.

The 'server' computer's only goal is to NOT be hammered by daily
rebooting, updating, installing and removing stuff, especially DRIVER
UPDATES under Microsoft's whimsical policies of 'let the end users
test it' update Tuesday and apology Thursday.

So long as we network and have access to the drives, we get the
benefits.

But if they are in another computer, then they are safe from whatever
goes wrong in our daily boot computer.

Just explaining 'server' as a verb, not a noun in this context.

- - - - -

That said, pulling out one drive at a time for long-term testing makes
sense.

We look forward to the months ahead as the results trickle in!

Good luck.

Thanks.