Windows based hard drive scan tool

[JW]

Looking for a Windows based hard drive scan utility. The one that comes
with Windows is useless in that if it has difficulty reading a sector, it
keeps banging away until successful then never reports the problem!
Looking for a program that reports a single error and not keep re-trying.

Currently checking out http://hddguru.com/software/2006.01.22-HDDScan/
But I'm not sure if it will do the same stupid thing Windows scan does...
Don't like the fact that HDDscan doesn't give you a progress bar - what on
Earth were they thinking?

Thanks.

 

[JW]

Looking for a Windows based hard drive scan utility. The one that comes
with Windows is useless in that if it has difficulty reading a sector, it
keeps banging away until successful then never reports the problem!
Looking for a program that reports a single error and not keep re-trying.

Currently checking out http://hddguru.com/software/2006.01.22-HDDScan/
But I'm not sure if it will do the same stupid thing Windows scan does...
Don't like the fact that HDDscan doesn't give you a progress bar - what on
Earth were they thinking?

Oops, I stand corrected. Clicking on the task *will* bring up a status
window. RTFM moment.

 

[Man-wai Chang]

Currently checking out http://hddguru.com/software/2006.01.22-HDDScan/

But I'm not sure if it will do the same stupid thing Windows scan does...
Don't like the fact that HDDscan doesn't give you a progress bar - what on
Earth were they thinking?

Linux has a command called badblocks, but it doesn't respect file
systems. So you possibly can't use it to mark a bad sector in a NTFS
partition.

 

[Mike Tomlinson]

Looking for a program that reports a single error and not keep re-trying.

HDTune has a scan option with a visual display. You can do the full
scan, or a quick one.

 

[Arno]

Linux has a command called badblocks, but it doesn't respect file
systems. So you possibly can't use it to mark a bad sector in a NTFS
partition.

Today you "mark" bad blocks on HDD level by overwriting them (and
the disk then does reallocation). If you have to makr on filesystem
layer, then the disk has run out of spare sectors and is close
to dath.

As to the original question: Run a long SMART selftest, e.g. with
a tool like HDDsentinel (free edition will work fine). Scanning
defective sectors from the OS is for disks that do not have
automatic defect management, and AFAIK these are not on the market
for something like > 20 years now.

Arno

 

[Rod Speed]

Arno wrote

Today you "mark" bad blocks on HDD level by overwriting them (and
the disk then does reallocation). If you have to makr on filesystem
layer, then the disk has run out of spare sectors and is close to dath.

As to the original question: Run a long SMART selftest,

The main problem with that approach is that it doesnt tell
you what the flakey sector is part of file or directory wise.

e.g. with a tool like HDDsentinel (free edition will work fine).

Scanning defective sectors from the OS is for disks that do
not have automatic defect management, and AFAIK these
are not on the market for something like > 20 years now.

Correct.

 

[Man-wai Chang]

Today you "mark" bad blocks on HDD level by overwriting them (and

the disk then does reallocation). If you have to makr on filesystem
layer, then the disk has run out of spare sectors and is close
to dath.

Ever since introduction of IDE hard disks, we lost the ability to mark a
bad sector by low-level format. And Micro$oft's format program just
refuse to mark bad sectors at file-system level.

--
@~@ Might, Courage, Vision, SINCERITY.
/ v \ Simplicity is Beauty! May the Force and Farce be with you!
/( _ )\ (x86_64 Ubuntu 9.10) Linux 2.6.35.7
^ ^ 22:28:02 up 3 days 8:07 0 users load average: 2.56 2.40 2.28
ä¸å€Ÿè²¸! ä¸è©é¨™! ä¸æ´äº¤! ä¸æ‰“交! ä¸æ‰“劫! ä¸è‡ªæ®º! è«‹è€ƒæ…®ç¶œæ´ (CSSA):
http://www.swd.gov.hk/tc/index/site_pubsvc/page_socsecu/sub_addressesa

 

[Rod Speed]

Man-wai Chang wrote

Ever since introduction of IDE hard disks, we lost the ability to mark a bad sector by low-level format.

Nope, the earliest IDEs could still do that.

And Micro$oft's format program just refuse to mark bad sectors at file-system level.

Wrong again, it will mark them as bad if they are bad enough.

 

[Arno]

Ever since introduction of IDE hard disks, we lost the ability to mark a
bad sector by low-level format. And Micro$oft's format program just
refuse to mark bad sectors at file-system level.

This does make sense, as the disk is in the process of dying
anyways if defect reallocation stops working. For Linux, you
can still manually pass a list of bad blocks to mke2fs and
e2fsck. mkdosfs allso supports this, so for FAT12/16/32 created
with Linux (they work without problem with at least XP and W7),
you can also specify bad blocks manually. Linux mkntfs does
not seem to have that option anymore, but it does call badblocks,
which in turn could be manipulated to give the list you want to
mkntfs. Seems this functionality is not irrelevant enough that
nobody biotherd to add the (very simple to add) commanfline
option to read from file instead from badblocks output.

But all that said, with Linux you get bad blocks marking
in filesystem level for ext2/3/4, FAT and NTFS.

Arno

 

[Yousuf Khan]

As to the original question: Run a long SMART selftest, e.g. with
a tool like HDDsentinel (free edition will work fine). Scanning
defective sectors from the OS is for disks that do not have
automatic defect management, and AFAIK these are not on the market
for something like> 20 years now.

I don't think the free version supports the long self-test, just the
short one.

Yousuf Khan

 

[Arno]

I don't think the free version supports the long self-test, just the
short one.

Then use the smartmontools, they are unrestricted. Or any other
free SMART tool that can run long selftests.

Arno

 

[Man-wai Chang]

But all that said, with Linux you get bad blocks marking

in filesystem level for ext2/3/4, FAT and NTFS.

Great...

 

[Man-wai Chang]

And Micro$oft's format program just refuse to mark bad sectors at file-system level.

Wrong again, it will mark them as bad if they are bad enough.

Bad enough means "minutes of failures", rather than "a few tries"?

 

[Rod Speed]

Man-wai Chang wrote

Bad enough means "minutes of failures", rather than "a few tries"?

Nope, nothing like minutes of failures.

 

[Man-wai Chang]

Bad enough means "minutes of failures", rather than "a few tries"?

Nope, nothing like minutes of failures.

Any hack to change the time from minutes to seconds?

And why did Micro$oft smart asses do this?

--
@~@ Might, Courage, Vision, SINCERITY.
/ v \ Simplicity is Beauty! May the Force and Farce be with you!
/( _ )\ (x86_64 Ubuntu 9.10) Linux 2.6.35.7
^ ^ 18:49:01 up 3:50 0 users load average: 1.06 1.11 1.12
ä¸å€Ÿè²¸! ä¸è©é¨™! ä¸æ´äº¤! ä¸æ‰“交! ä¸æ‰“劫! ä¸è‡ªæ®º! è«‹è€ƒæ…®ç¶œæ´ (CSSA):
http://www.swd.gov.hk/tc/index/site_pubsvc/page_socsecu/sub_addressesa

 

[Rod Speed]

Any hack to change the time from minutes to seconds?

Nope, because most of the retrying is the drive doing it, not the OS.

And why did Micro$oft smart asses do this?

Its just a hangover from the past when drives didnt handle bad sectors themselves.

There is no point in marking bads at the OS level anymore,
its best to let the drive handle bads auto instead.

 

[Man-wai Chang]

Any hack to change the time from minutes to seconds?

Nope, because most of the retrying is the drive doing it, not the OS.

Really? So the drive was trying to remap the sector?

--
@~@ Might, Courage, Vision, SINCERITY.
/ v \ Simplicity is Beauty! May the Force and Farce be with you!
/( _ )\ (x86_64 Ubuntu 9.10) Linux 2.6.35.7
^ ^ 19:32:01 up 1 day 4:33 0 users load average: 1.30 1.26 1.18
ä¸å€Ÿè²¸! ä¸è©é¨™! ä¸æ´äº¤! ä¸æ‰“交! ä¸æ‰“劫! ä¸è‡ªæ®º! è«‹è€ƒæ…®ç¶œæ´ (CSSA):
http://www.swd.gov.hk/tc/index/site_pubsvc/page_socsecu/sub_addressesa

 

[Arno]

Really? So the drive was trying to remap the sector?

The way this works is as follows:

1. The drive reads the sector and gets a checksum error
2. It tries with the ECC (Error Correction Code) data that
is part of the sector to correct it. It that succeeds and
the error looks serious, it will look more closely at the
sector and if it really has a problem, it will reallocate.
(Reads start before the heads are fully settled, small,
easily ECC correctible errors are common after seeks.)
3. If ECC fails, it tries to read the sector to get the data.
This may take a long time (minutes). Incidentally, that is
where "RAID drives" are different, they only try a short
time and than report an error. Trying a long time can make
a RAID controller to drop the whole drive from the Array.
4. If the read succeeds eventually (and often it does,
"retry" is a pretty successful strategy for read errors
from magnetic media), the drive re-tests the sector
to find out whether the disk surface has a problem or
whethert his was just a botched write (mechanical shock,
power spike or droput, EMV, etc.). If the sector tests
fine, it gets rewritten, if not reallocated. In both
cases the data is recovered.
5. If the reads do not eventually succeed, an error is
reported to the OS and the sector marked as potentially bad.
6. If a new read to this sector is done, goto 1.
7. If a write to this secotr is done, the drive will
test the sector (it knows there was a read problem)
and will basically do the same as under 4.

The best thing here (unless you are a data-recovery outfit and
know how to mess with firmware setings) is to just let the drive
do its thing. If you have a "RAID drive" with an unreadable sector
and need the data in there because you operated it as non-RAID
(not recommended) or RAID-0 (even less recommended), you can
do repeated read attempts from the software side and hope for
the best. You can also try to set the drive to non-RAID mode,
byt that is decidedly an "experts only" option.

Arno

 

[Man-wai Chang]

The best thing here (unless you are a data-recovery outfit and

know how to mess with firmware setings) is to just let the drive
do its thing. If you have a "RAID drive" with an unreadable sector
and need the data in there because you operated it as non-RAID
(not recommended) or RAID-0 (even less recommended), you can
do repeated read attempts from the software side and hope for
the best. You can also try to set the drive to non-RAID mode,
byt that is decidedly an "experts only" option.

Oh my cow.. the device should let me choose how to deal with potentially
bad sectors!!!!! Ask me about the number of retries.... With all these
stupid attempts to save a bad sector, it meant we should just throw away
the disk once you heard the VERY FIRST retry sound (click...click...click).

--
@~@ Might, Courage, Vision, SINCERITY.
/ v \ Simplicity is Beauty! May the Force and Farce be with you!
/( _ )\ (x86_64 Ubuntu 9.10) Linux 2.6.35.7
^ ^ 22:07:01 up 1 day 7:08 0 users load average: 1.02 1.13 1.10
ä¸å€Ÿè²¸! ä¸è©é¨™! ä¸æ´äº¤! ä¸æ‰“交! ä¸æ‰“劫! ä¸è‡ªæ®º! è«‹è€ƒæ…®ç¶œæ´ (CSSA):
http://www.swd.gov.hk/tc/index/site_pubsvc/page_socsecu/sub_addressesa

 

[GMAN]

Oh my cow.. the device should let me choose how to deal with potentially
bad sectors!!!!! Ask me about the number of retries.... With all these
stupid attempts to save a bad sector, it meant we should just throw away
the disk once you heard the VERY FIRST retry sound (click...click...click).

I am mad that its not easy to move the G-list over to the P-list with SMART.