Discussion:
freebsd-questions Digest, Vol 723, Issue 2
(too old to reply)
Dave B via freebsd-questions
2018-04-17 12:25:43 UTC
Permalink
ntfs HD with io errors
16/04/18 15:07
Have a 3TB ntfs HD with i/o errors.
ntfs-3g does read it, but a ls command of 2 different directories get
error msg about i/o errors and nothing is listed.
Is there a way to fsck this HD to fix it?
There is only 600GB of data on the 3TB HD. Looking for way to copy the
data to a 1TB ufs HD.
What will dump command do with i/o errors?
Can dd command skip over unused space?
Any other ideas?
Check the power supply, for stability, voltage level, noise etc.
(especially if the errors have a some randomness about them.)

I've missed a few exchanges on this thread, but what (if known) happened
to the drive immediately before the trouble started?  Or, between it's
last known good state, and the state it's in now.   ANY unscheduled
"events"?

Is it making any "unusual" sounds, frequent servo resets, squealing or
ticking noises.  Does it vibrate?  (They all do to an extent) but if
it's significant, like you can feel it when holding it in your hand
while it is spinning, the drive may have been dropped and a platter has
moved off centre on the spindle.  I've a portable NAS drive that's
happened to, it runs hotter as a result of the extra work the servo has
to do to track the off centre cylinders!.

Try it in different positions, laying flat, vertical etc, at an angle
all orientations.   But don't move it while it is spinning.

If the errors seem to be getting more and more frequent, it may be on
the verge of dying anyway.

I'd still go for the Spinrite option.  If it works, great, read the
contents and pension off the drive for a new one (but Spinrite that
before use!)  If it doesn’t help, contact GRC it is likely they will
refund you, I know of someone personally who had that happen.

Regards.

Dave B.
Dave B via freebsd-questions
2018-04-17 12:30:21 UTC
Permalink
This is a 3 TB disk and spinrite can only scan the first 2 TB's. Usually
that would be enough if the product can fix it, but it is worth noting.
If you do go with Spinrite, then OK the current version may limit at 2Tb
(not sure of that myself) but there is a newer version in the pipeline
that will be a major advance in the size of drive it can work with, and
the amount of data it transfers in one hit, greatly speeding up the test
times.   Current version 6 owners (no doubt that will be verified) are
said to be able to get the new version for free, when it is released.

Regards.

Dave B.

NO affiliation, for info only.
--
Created on and sent from a Unix like PC running and using free and open source software:
Polytropon
2018-04-17 15:30:02 UTC
Permalink
Post by Dave B via freebsd-questions
Check the power supply, for stability, voltage level, noise etc.
(especially if the errors have a some randomness about them.)
The disk currently is attached via a SATA->USB adapter,
if I remember correctly.
Post by Dave B via freebsd-questions
I've missed a few exchanges on this thread, but what (if known) happened
to the drive immediately before the trouble started?  Or, between it's
last known good state, and the state it's in now.   ANY unscheduled
"events"?
That kind of information has not been provided yet. The
assumption "a virus" and "no hardware defercts", but "with
dd, I/O errors"...
Post by Dave B via freebsd-questions
Try it in different positions, laying flat, vertical etc, at an angle
all orientations.   But don't move it while it is spinning.
Yes, this sometimes helps when dealing with read errors
(sounds strange, but sometimes _really_ works).
Post by Dave B via freebsd-questions
If the errors seem to be getting more and more frequent, it may be on
the verge of dying anyway.
Performing a check with smartctl should show.
Post by Dave B via freebsd-questions
I'd still go for the Spinrite option.  If it works, great, read the
contents and pension off the drive for a new one (but Spinrite that
before use!)  If it doesn’t help, contact GRC it is likely they will
refund you, I know of someone personally who had that happen.
Definitely worth a try. Also, checking in what state
the FS (here: NTFS) is and maybe repairing it (with
the 1:1 copy, just in case!) could enable the missing
data for proper reading. If not, magicrescue can at
least get the data, while file information (directory
and file names) cannot be determined.
--
Polytropon
Magdeburg, Germany
Happy FreeBSD user since 4.0
Andra moi ennepe, Mousa, ...
Loading...