Bad Disk block - but where?

My backup program reports a I/O error. In the System Log I
see: The device, \Device\Harddisk0\DR0, has a bad block.

Is Harddisk0 the same as Disk 0 in Disk Management? And
what does DR0 mean?

On 9/6/19 7:14 PM, Jason wrote:
My backup program reports a I/O error. In the System Log I
see: The device, \Device\Harddisk0\DR0, has a bad block.

Is Harddisk0 the same as Disk 0 in Disk Management? And
what does DR0 mean?


Hi Jason,

Not what you asked, but that won't stop me from commenting.
Lucky you!

Immediate fix:
chkdsk c: /r (presuming HardDisk0 is C

And, you have the start of a bad drive. Plan on replacing
and cloning it to a new drive ASAP before it becomes unreadable.

Clonezilla, in its advanced options, has a switch to "skip bad
sectors". I use the option frequently.

https://clonezilla.org/clonezilla-live.php

You can use dd with the iso to copy to a flash drive.

-T

In article , lid says...
chkdsk c: /r (presuming HardDisk0 is C

Thanks. Every variation on chkdsk fails to find anything wrong...

In my case, if I'm reading the numbering in Disk Management correctly,
HardDisk0 is not C: (fortunately). I moved things around when I added a
SSD. Best I can figure is that Disk 0 is the one I've added that just
holds %temp% and the paging file (as well as temp space for some apps
like Audition).

Jason wrote:
My backup program reports a I/O error. In the System Log I
see: The device, \Device\Harddisk0\DR0, has a bad block.

Is Harddisk0 the same as Disk 0 in Disk Management? And
what does DR0 mean?


That's raw disk I/O at a low level.

\Device\Harddisk0 # the whole disk, offset starts at 0

\Device\Harddisk0\Partition0 # the partition, starting at, say, sector 63
# This style addressing is only sufficient for
# a single partition (address starts at zero,
# applied to the partition start address).

\Device\Harddisk0\DR0 # Microsoft refuses to parse this for us.
# The digit on the end tends to be the same
# as the hard disk number before it.
# Harddisk2\DR2
# Other letters besides DR are possible, which
# is why a secret decoder ring would be nice to
# have, but I haven't seen one.

If the storage device is small enough, HDTune health scan can give you
a review of the disk, by reading the blocks for you.

This program cannot decode SSD SMART fields for you,
but it can do the Health Scan. A newer version of the
program, is of course "full featured". This is free,
and it's a ten year old program.

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

Unlike on hard drives, the physical location of the bad block
is meaningless. The drive works from a "pool of free sectors",
and which sector it uses next can be a "jumble". Consequently,
the LBA that has failed is just "some LBA". We don't care where
it is. Failed means that more than 10% of the sector is
messed up and the powerful error correcting code cannot fix it.
The storage area will be up for replacement on the next attempted
write to that location.

On a TLC or QLC drive, this can happen for a file that was
written "years ago", and the charge is draining off a bit.
There was never an intention that the flash storage be
"sloppy enough", that entire sectors drop out. The weakened
cells were only intended to require correction on every read,
with sufficient strength (Hamming distance) for the code
to correct the sector up to the end of the device life.
So what's happened to you is not "desirable" by any stretch
of the imagination. It's not supposed to do that.

This could happen if the device was exposed to ionizing radiation.
(Time to dig out that Geiger counter.)

If the ToolKit for your drive says the drive is near end of
life and has done 3000 writes per location, you might expect
this sort of behavior. It's just as likely the drive could
go Read Only (some brands) or wink out entirely (Intel!).

Paul

On 9/6/19 7:59 PM, Jason wrote:
In article , lid says...
chkdsk c: /r (presuming HardDisk0 is C

Thanks. Every variation on chkdsk fails to find anything wrong...

In my case, if I'm reading the numbering in Disk Management correctly,
HardDisk0 is not C: (fortunately). I moved things around when I added a
SSD. Best I can figure is that Disk 0 is the one I've added that just
holds %temp% and the paging file (as well as temp space for some apps
like Audition).


Be careful, the only SSD drive I have found that are
not crap are Samsung's:

el-cheap-o SSD's + Intel SSD's = Door Stop

My experience is that it is best to size your hard
drive large enough that everything can go on your
C: drive. In theory you can use a second D: drive,
but that goes wrong a lot when programmers forget to
target the D: drive and write half their stuff to C:
and half to D:

Use a second drive just as back up.

On 9/6/19 7:59 PM, Jason wrote:
HardDisk0 is not C: (fortunately)

It might be the same physical drive though. Windows
Nein (W10) likes lots of partitions.

Jason wrote:

lid says...

chkdsk c: /r (presuming HardDisk0 is C

Thanks. Every variation on chkdsk fails to find anything wrong...

In my case, if I'm reading the numbering in Disk Management correctly,
HardDisk0 is not C: (fortunately). I moved things around when I added a
SSD. Best I can figure is that Disk 0 is the one I've added that just
holds %temp% and the paging file (as well as temp space for some apps
like Audition).

Disks start from 0 (zero).
Partititions start with 1.

In an elevated command shell, run:

diskpart
list disk

That lists the disks. If they are differently sized, you can match the
"Disk n" number to the matching physical disk. To see the partitions
in a disk, run:

select disk n
list partition

Notice the disks started with zero but partitions start with one. To
exit diskpart and also the command shell, run:

exit
exit

The numbering is in the order the BIOS/UEFI discovers the disks, and
that depends into which onboard SATA port you plug in the disks. You
added an SSD, but did you unplug your old HDD from the SATA1 port (into
another SATA port) and plug the SSD into the SATA1 port? Or did you
simply plug the SSD disk into whatever was an unused SATA port?

You can use SysInternals' DiskExt to list the mount ID, drive letter,
disk number, and partition (extent) of each drive in the OS.

https://docs.microsoft.com/en-us/sys...nloads/diskext

You can also run the following command in a command shell:

powershell
get-disk
exit

The physical disk number is under the "Number" column. For me, disk0 is
not the bootable one for the OS. Instead, the m.2 NVM3 SSD is seen as
disk number 1. The internal HDD is disk 0, my printer's USB drive port
is disk 3. And the USB-attached HDD is disk 2.

You can also select a disk in Disk Management (diskmgmt.msc). I don't
mean to select a drive from the top-pane list. Instead, in the bottom
pane, right-click on the "Disk n" entry which shows the partitioning
on the right side. Select to view Properties and go to the Volumes tab.

HardDisk0\DR0 both reference the first drive. HardDisk# matches the
"Disk #" shown in Disk Management. The # is assigned as the devices are
detected upon boot. The DR# follows different rules: numbers are
sequentially assigned and are not reused while Windows is booting. Each
time a removable drive is disconnected and reconnected, the DR#
increments. Initially HardDisk# and DR# will match (and remain fixed
for internal disks); however, use of removable drives causes the DR# to
increase (but # gets reset to 0 upon restart/boot of Windows, not when
you logout and log back in). For internal HDDs and SSDs, DR# doesn't
change. For removable media, DR# increments on each remove and reinsert
but gets reset to 0 on a restart (not on a logoff).

Which backup program? I've seen problems like this with some backup
programs, especially if the HDD is a "green" drive. Could've been
caused by a power outage flicker: often just barely long enough to
visually notice and hopefully longer than the drain time for the
capacitors in the PSU, but the surge can effect defect in behavior.
Gamma radiation does, too. Also, HDDs and SSDs have reserve space to
mask out bad sectors. It's when you see a pending reallocation count
that doesn't diminish to zero that the disk is going bad: there are no
more reserved sectors to mask the bad ones. There might've been a bad
sector at one time, but it got remapped, so now any access to the bad
sector gets redirected to one in the remap block. The redirection slows
access to the disk but the bad physical sector isn't accessible anymore,
just the remapped sector is available. That's why looking at the SMART
data will tell you how many sectors have been reallocated (remapped).
Remapping of bad sectors is normal operation. No disk is perfect,
defects occur, and they get masked out. There's no way to find out from
the specs how large is the remap area. You check if the pending
reallocation count has gone up and doesn't go down.

Jason wrote:
In article , lid says...
chkdsk c: /r (presuming HardDisk0 is C

Thanks. Every variation on chkdsk fails to find anything wrong...

In my case, if I'm reading the numbering in Disk Management correctly,
HardDisk0 is not C: (fortunately). I moved things around when I added a
SSD. Best I can figure is that Disk 0 is the one I've added that just
holds %temp% and the paging file (as well as temp space for some apps
like Audition).

Diskpart (command line) and Disk Management, number
the disks in the same order. The zeroth disk, you'd expect
it to be at the top of the chart.

https://i.postimg.cc/y6MNC0y0/disk-management.gif

Paul

In article , lid

Be careful, the only SSD drive I have found that are
not crap are Samsung's:

el-cheap-o SSD's + Intel SSD's = Door Stop
If I'm decoding the disk # correctly, system is
complaining about a Samsung 128GB SSD that I have devoted
to %temp% as well as temp space for some apps. That's all
it's for, so losing it is not a big problem.


My experience is that it is best to size your hard
drive large enough that everything can go on your
C: drive. In theory you can use a second D: drive,
but that goes wrong a lot when programmers forget to
target the D: drive and write half their stuff to C:
and half to D:

Use a second drive just as back up.

In article ,
says...
....

Which backup program? I've seen problems like this with some backup
programs, especially if the HDD is a "green" drive.

I'm still digesting all this, but the answer to your
question: Acronis. I've used it for many years but am
running a new release as of a week or so ago. Perhaps
that's a contributor.

Thank you all for the helpful information.
Assuming I'm testing the correct disk --pretty sure I am--
then it has passed all the chkdsk variants I tried. Just
now, after a fresh boot, there are log entries (type NTFS)
for each HD and show all of them to be healthy. The big
test will be a full backup run which I'll try later today.

In article ,
says...

Thank you all for the helpful information.
Assuming I'm testing the correct disk --pretty sure I am--
then it has passed all the chkdsk variants I tried. Just
now, after a fresh boot, there are log entries (type NTFS)
for each HD and show all of them to be healthy. The big
test will be a full backup run which I'll try later today.

Be sure to use some utility that can give you an indication of the SMART
status of the disk. Something like Acronis Drive Monitor (free) or Hard
Disk Sentinel (not).

--

Phil, London

On 2019-09-06 19:14, Jason wrote:
My backup program reports a I/O error. In the System Log I
see: The device, \Device\Harddisk0\DR0, has a bad block.

Is Harddisk0 the same as Disk 0 in Disk Management? And
what does DR0 mean?


you used to be able to correct for this with a full disk scan with bad
block relocation via 'chkdsk'. I don't know what has been done with the
disk check menu recently, but last I looked, these options for "check on
next boot" and things like that don't seem to be there for some
reason... maybe I missed them.

Anyway, it used to be (as an admin in CMD) "chkdsk C: /F /R" and it
would prompt you to do it on the next bootup... (since it can't do that
to a live file system). Then just let it run until it's fixed
everything, a few hours probably. Pick a good time to do it, maybe do a
backup first.

you can try 'chkdsk /?' to see what options you still have available.


--
(aka 'Bombastic Bob' in case you wondered)

'Feeling with my fingers, and thinking with my brain' - me

'your story is so touching, but it sounds just like a lie'
"Straighten up and fly right"

On 2019-09-06 19:30, T wrote:
And, you have the start of a bad drive.Â* Plan on replacing
and cloning it to a new drive ASAP before it becomes unreadable.

yeah there are a few "drive-cloning" "ghost backup" etc. utilities out
there. Usually the SMART firmware will tell ya what the deal is on
newer drives, which he's probably using. You'd need a utility to read
the SMART info though. They're command line utilities in FreeBSD and
Linux, which I've used before when hard drives start to go bad [ZFS is a
LIFE SAVER for preventing data loss]. NTFS is not as reliable but
should be able to recover data when you run 'chkdsk /F /R'.

And then the SMART utility will tell ya if it's a one time fail or if
you're getting a growing cancer of bad blocks, or [worse, I've had this
happen] the electronics are starting to fail and you get read errors
that happen during heavy activity, but otherwise not. At least in this
last case data recovery is more likely if you back up files a few at a time.


--
(aka 'Bombastic Bob' in case you wondered)

'Feeling with my fingers, and thinking with my brain' - me

'your story is so touching, but it sounds just like a lie'
"Straighten up and fly right"

On 2019-09-06 20:44, T wrote:
Use a second drive just as back up.

or move your default user dir to D:\WhateverYouWant

Unfortunately I'm not 100% sure that windows applications, particularly
Micro-shaft ones, even 20 years following the possibility of doing this
with NT 4, can even handle this possility withotu causing problems...

(especially with SOME recent win-10-nic updates that reportedly kill off
your user files if not in the default location on C drive!!!)

Whereas, when I installed Linux on a formerly Win-10-nic Dell machine
(has a 256Mb SSD and 1T hard disk) I put the OS (minus /home) on the
SSD, and then put /home on the spinny drive.

works perfectly that way. Then the SSD does what it does best, and
reboots are EXTREMELY FAST.



--
(aka 'Bombastic Bob' in case you wondered)

'Feeling with my fingers, and thinking with my brain' - me

'your story is so touching, but it sounds just like a lie'
"Straighten up and fly right"