How to force copy retries until finally copied all due to a dying old HDD in Linux?

[Ant]

Hi!

I have a very old PATA HDD that is dying due to SMART errors and I am
trying to copy my data to another working drive, but it has problems
reading some data. Here is an example:

$ cp -R * ~/DebianPATAhddBackups/
cp: cannot stat `Files/AntsImages': Input/output error
cp: reading `Mac OS X.dmg': Input/output error
cp: failed to extend `Mac OS X.dmg': Input/output error

I assume my Debian stable was unable to copy those files. IIRC, there
was a way to retry over and over until copied fully. How do I do that?
It has been so many years I did this. :(

Thank you in advance. :)
--
"It is not enough to be busy, so are the ants. The question is: what are
we busy about?" --Henry David Thoreau
/\___/\ Ant(Dude) @ http://antfarm.ma.cx (Personal Web Site)
/ /\ /\ \ Ant's Quality Foraged Links: http://aqfl.net
| |o o| |
\ _ / If crediting, then use Ant nickname and AQFL URL/link.
( ) If e-mailing, then axe ANT from its address if needed.
Ant is currently not listening to any songs on this computer.

[Robert Heller]

At Sun, 04 Jan 2015 06:55:59 -0800 Ant <a...@zimage.comANT> wrote:

>
> Hi!
>
> I have a very old PATA HDD that is dying due to SMART errors and I am
> trying to copy my data to another working drive, but it has problems
> reading some data. Here is an example:
>
> $ cp -R * ~/DebianPATAhddBackups/
> cp: cannot stat `Files/AntsImages': Input/output error
> cp: reading `Mac OS X.dmg': Input/output error
> cp: failed to extend `Mac OS X.dmg': Input/output error
>
> I assume my Debian stable was unable to copy those files. IIRC, there
> was a way to retry over and over until copied fully. How do I do that?
> It has been so many years I did this. :(

dd can do that:

dd ... conv=noerror

man dd for full details.

Basically, you do a raw DD of the 'bad' disk's file system (eg create a copy
of the file system either onto a new disk partition (of the proper size) or to
an image file. Then you can mount the partition (or image via -o loop) and
then copy the files off. Note: files with bad blocks will likely be 'broken'
and may not be usable, but you should be able to retrieve other files.

>
> Thank you in advance. :)

--

Robert Heller -- 978-544-6933
Deepwoods Software -- Custom Software Services
http://www.deepsoft.com/ -- Linux Administration Services
hel...@deepsoft.com -- Webhosting Services

[Lusotec]

Ant wrote:
> Hi!
>
> I have a very old PATA HDD that is dying due to SMART errors and I am
> trying to copy my data to another working drive, but it has problems
> reading some data. Here is an example:
>
> $ cp -R * ~/DebianPATAhddBackups/
> cp: cannot stat `Files/AntsImages': Input/output error
> cp: reading `Mac OS X.dmg': Input/output error
> cp: failed to extend `Mac OS X.dmg': Input/output error
>
> I assume my Debian stable was unable to copy those files. IIRC, there
> was a way to retry over and over until copied fully. How do I do that?
> It has been so many years I did this. :(
>
> Thank you in advance. :)

https://www.gnu.org/software/ddrescue/
https://www.gnu.org/software/ddrescue/manual/ddrescue_manual.html

[Ant]

On 1/4/2015 7:23 AM, Robert Heller wrote:

> dd can do that:
>
> dd ... conv=noerror
>
> man dd for full details.
>
> Basically, you do a raw DD of the 'bad' disk's file system (eg create a copy
> of the file system either onto a new disk partition (of the proper size) or to
> an image file. Then you can mount the partition (or image via -o loop) and
> then copy the files off. Note: files with bad blocks will likely be 'broken'
> and may not be usable, but you should be able to retrieve other files.

This is more complex than I expected as a newbie. So, I can't copy files
directly like cp command? I do NOT need everything from the old dying
drive. I thought there was a way to do that to an existing HDD (don't
want to mess up my destination drive).

Anyways, I tried that one of the unreadable file as a quick test:
$ cp Mac\ OS\ X.dmg ~/Downloads/test

cp: reading `Mac OS X.dmg': Input/output error

cp: failed to extend `/home/ant/Downloads/test/Mac OS X.dmg':
Input/output error
--
"Where the sugar is, there will the ant be also." --Philippines

[Ant]

On 1/4/2015 7:39 AM, Lusotec wrote:

> https://www.gnu.org/software/ddrescue/manual/ddrescue_manual.html

"If you use a device or a partition as destination, any data stored
there will be overwritten." I don't want to erase anything on my
destinated drive (/dev/sdb*) though. I just wanted to copy a few
unreadable files. I do NOT want to copy everything from the very old
dying HDD. This is a lot more complex than I expected as a newbie. So, I
can't copy files directly like cp command? I thought there was a way to
do that to an existing HDD (don't want to mess up my destination drive). :(
--
"God is a mean kid sitting on an ant-hill with a magnifying glass, and
I'm the ant." --Bruce Nolan (Bruce Almighty movie)

[Lusotec]

Ant wrote:
> Lusotec wrote:
>> https://www.gnu.org/software/ddrescue/manual/ddrescue_manual.html
>
> "If you use a device or a partition as destination, any data stored
> there will be overwritten." I don't want to erase anything on my
> destinated drive (/dev/sdb*) though.

Then just use a plain file instead of a device or a partition as
destination.

> I just wanted to copy a few
> unreadable files. I do NOT want to copy everything from the very old
> dying HDD. This is a lot more complex than I expected as a newbie. So, I
> can't copy files directly like cp command? I thought there was a way to
> do that to an existing HDD (don't want to mess up my destination drive).
> :(

There is a GUI for it. May be easier to use.
https://launchpad.net/ddrescue-gui

[William Unruh]

On 2015-01-04, Ant <a...@zimage.comANT> wrote:
> Hi!
>
> I have a very old PATA HDD that is dying due to SMART errors and I am
> trying to copy my data to another working drive, but it has problems
> reading some data. Here is an example:
>
> $ cp -R * ~/DebianPATAhddBackups/
> cp: cannot stat `Files/AntsImages': Input/output error
> cp: reading `Mac OS X.dmg': Input/output error
> cp: failed to extend `Mac OS X.dmg': Input/output error
>
> I assume my Debian stable was unable to copy those files. IIRC, there
> was a way to retry over and over until copied fully. How do I do that?
> It has been so many years I did this. :(

If the rust has flaked off the drive, there is nothing to copy, no
matter how many times you try.

[William Unruh]

On 2015-01-04, Ant <a...@zimage.comANT> wrote:

> On 1/4/2015 7:39 AM, Lusotec wrote:
>
>> https://www.gnu.org/software/ddrescue/manual/ddrescue_manual.html
>
> "If you use a device or a partition as destination, any data stored
> there will be overwritten." I don't want to erase anything on my
> destinated drive (/dev/sdb*) though. I just wanted to copy a few
> unreadable files. I do NOT want to copy everything from the very old
> dying HDD. This is a lot more complex than I expected as a newbie. So, I
> can't copy files directly like cp command? I thought there was a way to
> do that to an existing HDD (don't want to mess up my destination drive). :(

If the data is important to you, go out and buy a new drive and use dd
from that broken drive to the new one (nothing there to erase). Then
afterwards reformat/partion the new drive and use it as spare storage
space (eg backups!).

When a hard disk fails, it fails. The recording medium is scratched, the
rust (iron oxide or ferrite) flakes off, etc. It is IMPOSSIBLE to
recover the data there. dd can be told to ignore errors, on blocks, and
just go on to the next one. cp cannot.

[Robert Heller]

Right. Using 'dd conf=noerrors ...' lets you copy what you can to 'clean'
media (or an image file). Then you can mount the clean media (or an image
file) and then recover what you can.

The *best* option is to *always* make regular backups in the first place. Then
when the drive fails (and it will, sooner or later), all of the data is on the
backup media (and it is best to make multiple backup copies on multiple
instantances of backup media of all of your important data).

[Rich]

In comp.os.linux.misc Ant <a...@zimage.comant> wrote:
> On 1/4/2015 7:23 AM, Robert Heller wrote:

> > dd can do that:
> >
> > dd ... conv=noerror
> >
> > man dd for full details.
> >
> > Basically, you do a raw DD of the 'bad' disk's file system (eg create a copy
> > of the file system either onto a new disk partition (of the proper size) or to
> > an image file. Then you can mount the partition (or image via -o loop) and
> > then copy the files off. Note: files with bad blocks will likely be 'broken'
> > and may not be usable, but you should be able to retrieve other files.

> This is more complex than I expected as a newbie. So, I can't copy files
> directly like cp command? I do NOT need everything from the old dying
> drive. I thought there was a way to do that to an existing HDD (don't
> want to mess up my destination drive).

dd in="/old/drive/Mac OS X.dmg" out="/new/drive/that/is/not/broken/Mac OS X.dmg" conv=noerror

should work to "copy" a single file. Change "in=" and "out=" files to
point to the ones you are trying to copy.

Note, that given the "errors" when reading the files, and the fact that
the example you gave is a mac dmg file, the result you get from the
"copy" process will likely be useless anyway.

[Robert Heller]

Yes indeed. The 'best' option is to copy the whole file system (using dd
conv=noerror) to a scratch disk or to a file image. Then go though the copy to
pick out the files that are still usable. The OP seemed to want to copy
everything still usable. Doing dd on each file is very tedious, since dd does
not work with wildcards or multiple files.

[Ant]

On 1/4/2015 10:07 AM, William Unruh wrote:
>>
>> I have a very old PATA HDD that is dying due to SMART errors and I am
>> trying to copy my data to another working drive, but it has problems
>> reading some data. Here is an example:
>>
>> $ cp -R * ~/DebianPATAhddBackups/
>> cp: cannot stat `Files/AntsImages': Input/output error
>> cp: reading `Mac OS X.dmg': Input/output error
>> cp: failed to extend `Mac OS X.dmg': Input/output error
>>
>> I assume my Debian stable was unable to copy those files. IIRC, there
>> was a way to retry over and over until copied fully. How do I do that?
>> It has been so many years I did this. :(
>
> If the rust has flaked off the drive, there is nothing to copy, no
> matter how many times you try.

Even with ddrescue? I thought it retries until it finally copies.
--
"The great companies did not know that the line between hunger and anger
is a thin line. And money that might have gone to wages went for gas,
for guns, for agents and spies, for blacklists, for drilling. On the
highways the people moved like ants and searched for work, for food. And
the anger began to ferment." --John Steinbeck's The Grapes of Wrath

[William Unruh]

On 2015-01-04, Ant <a...@zimage.comANT> wrote:

> On 1/4/2015 10:07 AM, William Unruh wrote:
>>>
>>> I have a very old PATA HDD that is dying due to SMART errors and I am
>>> trying to copy my data to another working drive, but it has problems
>>> reading some data. Here is an example:
>>>
>>> $ cp -R * ~/DebianPATAhddBackups/
>>> cp: cannot stat `Files/AntsImages': Input/output error
>>> cp: reading `Mac OS X.dmg': Input/output error
>>> cp: failed to extend `Mac OS X.dmg': Input/output error
>>>
>>> I assume my Debian stable was unable to copy those files. IIRC, there
>>> was a way to retry over and over until copied fully. How do I do that?
>>> It has been so many years I did this. :(
>>
>> If the rust has flaked off the drive, there is nothing to copy, no
>> matter how many times you try.
>
> Even with ddrescue? I thought it retries until it finally copies.

ddrescue just allows you to conitue if there are errors (it essentially
retries once) . It does NOT
keep retrying until it reads it (well, you can tell it to try, but as
they say, that could damage the disk so that subsequent reads of
formerly good areas could go bad-- rust flakes between the drive head
and the platter do not do good things to the platter).
To try forever would be stupid, because it can be
impossible to read it. As I said, data that is not there cannot be read,
no matter how hard you try.

[Rich]

In comp.os.linux.misc Ant <a...@zimage.comant> wrote:

> On 1/4/2015 10:07 AM, William Unruh wrote:
> >>
> >> I have a very old PATA HDD that is dying due to SMART errors and I am
> >> trying to copy my data to another working drive, but it has problems
> >> reading some data. Here is an example:
> >>
> >> $ cp -R * ~/DebianPATAhddBackups/
> >> cp: cannot stat `Files/AntsImages': Input/output error
> >> cp: reading `Mac OS X.dmg': Input/output error
> >> cp: failed to extend `Mac OS X.dmg': Input/output error
> >>
> >> I assume my Debian stable was unable to copy those files. IIRC, there
> >> was a way to retry over and over until copied fully. How do I do that?
> >> It has been so many years I did this. :(
> >
> > If the rust has flaked off the drive, there is nothing to copy, no
> > matter how many times you try.

> Even with ddrescue? I thought it retries until it finally copies.

If the disk sector is corrupted, no amount of retries will make it
"finally copy". It is simply gone. Retries are only useful when there
is an intermittent issue that occurs sometimes, but not other times.

This is why having backups on separate media is so very important.

[The Natural Philosopher]

On 04/01/15 17:25, Ant wrote:
> I just wanted to copy a few unreadable files

Good luck with that one...

--
Everything you read in newspapers is absolutely true, except for the
rare story of which you happen to have first-hand knowledge. – Erwin Knoll

[Ant]

Ah OK. I was reading
http://www.kalysto.org/utilities/dd_rhelp/index.en.html , and it sounded
interesting. I am currently running ddrescue right now from the old
drive to another old Seagate 80 GB PATA/IDE HDD that I found in one of
my moving boxes.
--
"I am afraid that I am convinced that a community of human beings is a
far more useful thing than a community of ants, and that if the human
being is condemned and restricted to perform the same functions over and
over again, he will not even be a good ant, not to mention a good human
being." --Norbert Wiener

[Ant]

On 1/4/2015 2:12 PM, Rich wrote:

>> Even with ddrescue? I thought it retries until it finally copies.
>
> If the disk sector is corrupted, no amount of retries will make it
> "finally copy". It is simply gone. Retries are only useful when there
> is an intermittent issue that occurs sometimes, but not other times.

Ah, so http://www.kalysto.org/utilities/dd_rhelp/index.en.html is
useless in this scenario. Interesting. Thanks.

[Rich]

In comp.os.linux.misc Ant <a...@zimage.comant> wrote:

> On 1/4/2015 2:12 PM, Rich wrote:

> >> Even with ddrescue? I thought it retries until it finally copies.
> >
> > If the disk sector is corrupted, no amount of retries will make it
> > "finally copy". It is simply gone. Retries are only useful when there
> > is an intermittent issue that occurs sometimes, but not other times.

> Ah, so http://www.kalysto.org/utilities/dd_rhelp/index.en.html is
> useless in this scenario. Interesting. Thanks.

That simply runs ddrescue on the good parts of the disk first
(therefore you get the maximum amount of readable data copied over in
the shortest time span) before trying to read the bad parts multiple
times.

But nothing will change the physical fact that data which is unreadable
(true unreadable, not intermittent) is just that, unreadable, and no
amount of attempting to read it over and over will magically make it
readable.

[Ant]

Thank you for explaining that. It was technical to read the details. :)

FYI. I was running dd_rescue which was the wrong command from Reddit
comments and here. So, I aborted it and saw:
...
Bad block: 29983
^C
dd_rescue: (fatal): Caught signal 2 "Interrupt". Exiting!
Summary for /dev/sda5 -> /dev/sdg1:
dd_rescue: (info): ipos: 15004.0k, opos: 15004.0k, xferd:
15004.0k
errs: 336, errxfer: 168.0k, succxfer:
14836.0k
+curr.rate: 2kB/s, avg.rate: 2kB/s,
avg.load: 0.0%


real 166m1.617s
user 0m0.004s
sys 0m0.084s

Ouch, 2 kB/sec! Slower than my crappy 3 kB/sec dial-up Internet
connections! :( So, I started "ddrescue --force /dev/sda5 /dev/sdg1
logfile" and recently saw:

# time ddrescue --force /dev/sda5 /dev/sdg1 logfile
GNU ddrescue 1.16
Press Ctrl-C to interrupt
rescued: 2955 MB, errsize: 109 MB, current rate: 1703 kB/s
rescued: 41020 MB, errsize: 109 MB, current rate: 22675 kB/s
ipos: 41130 MB, errors: 47, average rate: 8898 kB/s
opos: 41130 MB, time since last successful read: 0 s
Copying non-tried blocks...

FYI, sdg(1=ext4 storage2 partition) is my old external USB2 HDD. Both
failing drive and external USB2 HDD are not mounted right now for dd
rescues.
--
"Have I told you how much I like ants, huh? Especially fried in a subtle
blend of mech fluid and grated gears?" --Rampage to Inferno,
"Transmutate" in Transformers (Beast Wars)

/\___/\ Ant(Dude) @ http://antfarm.ma.cx (Personal Web Site)
/ /\ /\ \ Ant's Quality Foraged Links: http://aqfl.net
| |o o| |
\ _ / If crediting, then use Ant nickname and AQFL URL/link.
( ) If e-mailing, then axe ANT from its address if needed.

A song (i/wa)s playing on this computer: Don McLean - American Pie
(Mastermix Classic Cuts 99 - Party)

[Ant]

> FYI, sdg(1=ext4 storage2 partition) is my old external USB2 HDD. Both
> failing drive and external USB2 HDD are not mounted right now for dd
> rescues.

My ddrescue aborted or maybe it is done? I am not sure if it was caused
by me while disabling Bluetooth (don't have any!) in KDE4's settings or
not, so I looked around:

# time ddrescue --force /dev/sda5 /dev/sdg1 logfile
GNU ddrescue 1.16
Press Ctrl-C to interrupt
rescued: 2955 MB, errsize: 109 MB, current rate: 1703 kB/s

rescued: 79615 MB, errsize: 109 MB, current rate: 630 B/s
ipos: 2147 MB, errors: 47, average rate: 11651 kB/s
opos: 2147 MB, time since last successful read: 0 s
Trimming failed blocks...
ddrescue: write error: Input/output error

real 113m52.396s
user 0m2.428s
sys 23m27.752s

--

dmesg:
...
[2015-01-04 19:04:59] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0
action 0x0
[2015-01-04 19:04:59] ata1.00: BMDMA stat 0x25
[2015-01-04 19:04:59] ata1.00: failed command: READ DMA
[2015-01-04 19:04:59] ata1.00: cmd c8/00:08:18:ca:4c/00:00:00:00:00/e0
tag 0 dma 4096 in
[2015-01-04 19:04:59] res 51/40:00:1c:ca:4c/00:00:00:00:00/e0
Emask 0x9 (media error)
[2015-01-04 19:04:59] ata1.00: status: { DRDY ERR }
[2015-01-04 19:04:59] ata1.00: error: { UNC }
[2015-01-04 19:04:59] ata1.00: configured for UDMA/100
[2015-01-04 19:04:59] ata1: EH complete
[2015-01-04 19:05:03] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0
action 0x0
[2015-01-04 19:05:03] ata1.00: BMDMA stat 0x25
[2015-01-04 19:05:03] ata1.00: failed command: READ DMA
[2015-01-04 19:05:03] ata1.00: cmd c8/00:08:18:ca:4c/00:00:00:00:00/e0
tag 0 dma 4096 in
[2015-01-04 19:05:03] res 51/40:00:1c:ca:4c/00:00:00:00:00/e0
Emask 0x9 (media error)
[2015-01-04 19:05:03] ata1.00: status: { DRDY ERR }
[2015-01-04 19:05:03] ata1.00: error: { UNC }
[2015-01-04 19:05:04] ata1.00: configured for UDMA/100
[2015-01-04 19:05:04] ata1: EH complete
[2015-01-04 19:05:08] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0
action 0x0
[2015-01-04 19:05:08] ata1.00: BMDMA stat 0x25
[2015-01-04 19:05:08] ata1.00: failed command: READ DMA
[2015-01-04 19:05:08] ata1.00: cmd c8/00:08:18:ca:4c/00:00:00:00:00/e0
tag 0 dma 4096 in
[2015-01-04 19:05:08] res 51/40:00:1c:ca:4c/00:00:00:00:00/e0
Emask 0x9 (media error)
[2015-01-04 19:05:08] ata1.00: status: { DRDY ERR }
[2015-01-04 19:05:08] ata1.00: error: { UNC }
[2015-01-04 19:05:08] ata1.00: configured for UDMA/100
[2015-01-04 19:05:08] ata1: EH complete
[2015-01-04 19:05:13] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0
action 0x0
[2015-01-04 19:05:13] ata1.00: BMDMA stat 0x25
[2015-01-04 19:05:13] ata1.00: failed command: READ DMA
[2015-01-04 19:05:13] ata1.00: cmd c8/00:08:18:ca:4c/00:00:00:00:00/e0
tag 0 dma 4096 in
[2015-01-04 19:05:13] res 51/40:00:1c:ca:4c/00:00:00:00:00/e0
Emask 0x9 (media error)
[2015-01-04 19:05:13] ata1.00: status: { DRDY ERR }
[2015-01-04 19:05:13] ata1.00: error: { UNC }
[2015-01-04 19:05:13] ata1.00: configured for UDMA/100
[2015-01-04 19:05:13] ata1: EH complete
[2015-01-04 19:05:18] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0
action 0x0
[2015-01-04 19:05:18] ata1.00: BMDMA stat 0x25
[2015-01-04 19:05:18] ata1.00: failed command: READ DMA
[2015-01-04 19:05:18] ata1.00: cmd c8/00:08:18:ca:4c/00:00:00:00:00/e0
tag 0 dma 4096 in
[2015-01-04 19:05:18] res 51/40:00:1c:ca:4c/00:00:00:00:00/e0
Emask 0x9 (media error)
[2015-01-04 19:05:18] ata1.00: status: { DRDY ERR }
[2015-01-04 19:05:18] ata1.00: error: { UNC }
[2015-01-04 19:05:18] ata1.00: configured for UDMA/100
[2015-01-04 19:05:18] ata1: EH complete
[2015-01-04 19:05:22] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0
action 0x0
[2015-01-04 19:05:22] ata1.00: BMDMA stat 0x25
[2015-01-04 19:05:22] ata1.00: failed command: READ DMA
[2015-01-04 19:05:22] ata1.00: cmd c8/00:08:18:ca:4c/00:00:00:00:00/e0
tag 0 dma 4096 in
[2015-01-04 19:05:22] res 51/40:00:1c:ca:4c/00:00:00:00:00/e0
Emask 0x9 (media error)
[2015-01-04 19:05:22] ata1.00: status: { DRDY ERR }
[2015-01-04 19:05:22] ata1.00: error: { UNC }
[2015-01-04 19:05:22] ata1.00: configured for UDMA/100
[2015-01-04 19:05:22] sd 0:0:0:0: [sda] Unhandled sense code
[2015-01-04 19:05:22] sd 0:0:0:0: [sda] Result: hostbyte=DID_OK
driverbyte=DRIVER_SENSE
[2015-01-04 19:05:22] sd 0:0:0:0: [sda] Sense Key : Medium Error
[current] [descriptor]
[2015-01-04 19:05:22] Descriptor sense data with sense descriptors (in hex):
[2015-01-04 19:05:22] 72 03 11 04 00 00 00 0c 00 0a 80 00 00 00
00 00
[2015-01-04 19:05:22] 00 4c ca 1c
[2015-01-04 19:05:22] sd 0:0:0:0: [sda] Add. Sense: Unrecovered read
error - auto reallocate failed
[2015-01-04 19:05:22] sd 0:0:0:0: [sda] CDB: Read(10): 28 00 00 4c ca 18
00 00 08 00
[2015-01-04 19:05:22] end_request: I/O error, dev sda, sector 5032476
[2015-01-04 19:05:22] Buffer I/O error on device sda5, logical block 556355
[2015-01-04 19:05:22] ata1: EH complete
[2015-01-04 19:05:48] sd 11:0:0:0: [sdg] Unhandled sense code
[2015-01-04 19:05:48] sd 11:0:0:0: [sdg] Result: hostbyte=DID_ERROR
driverbyte=DRIVER_SENSE
[2015-01-04 19:05:48] sd 11:0:0:0: [sdg] Sense Key : Hardware Error
[current]
[2015-01-04 19:05:48] sd 11:0:0:0: [sdg] Add. Sense: No additional
sense information
[2015-01-04 19:05:48] sd 11:0:0:0: [sdg] CDB: Read(10): 28 00 00 40 08
58 00 00 08 00
[2015-01-04 19:05:48] end_request: I/O error, dev sdg, sector 4196440
...

I reran it, and it failed again and I didn't do anything to cause it:

# time ddrescue --force /dev/sda5 /dev/sdg1 logfile
GNU ddrescue 1.16
Press Ctrl-C to interrupt

Initial status (read from logfile)
rescued: 79615 MB, errsize: 109 MB, errors: 47
Current status
rescued: 79615 MB, errsize: 109 MB, current rate: 0 B/s
ipos: 2147 MB, errors: 47, average rate: 0 B/s
opos: 2147 MB, time since last successful read: 25 s
Trimming failed blocks...
ddrescue: write error: Input/output error

real 0m24.515s
user 0m0.000s
sys 0m0.000s


dmesg:
...
[2015-01-04 19:13:44] kjournald starting. Commit interval 5 seconds
[2015-01-04 19:13:44] EXT3-fs (sdg1): warning: mounting fs with errors,
running e2fsck is recommended
[2015-01-04 19:13:44] EXT3-fs (sdg1): using internal journal
[2015-01-04 19:13:44] EXT3-fs (sdg1): mounted filesystem with ordered
data mode
[2015-01-04 19:15:27] sd 11:0:0:0: [sdg] Unhandled sense code
[2015-01-04 19:15:27] sd 11:0:0:0: [sdg] Result: hostbyte=DID_ERROR
driverbyte=DRIVER_SENSE
[2015-01-04 19:15:27] sd 11:0:0:0: [sdg] Sense Key : Hardware Error
[current]
[2015-01-04 19:15:27] sd 11:0:0:0: [sdg] Add. Sense: No additional
sense information
[2015-01-04 19:15:27] sd 11:0:0:0: [sdg] CDB: Read(10): 28 00 00 40 08
58 00 00 08 00
[2015-01-04 19:15:27] end_request: I/O error, dev sdg, sector 4196440

I mounted both drives and did a df:
$ df
Filesystem 1K-blocks
Used Available Use% Mounted on
rootfs 960504
733580 178132 81% /
udev 10240
0 10240 0% /dev
tmpfs 206056
11356 194700 6% /run
/dev/disk/by-uuid/242d81f2-8a0d-48e1-a144-22ab4d2b9407 960504
733580 178132 81% /
tmpfs 5120
8 5112 1% /run/lock
tmpfs 1193200
80 1193120 1% /run/shm
/dev/sdb9 50013916
16658680 30814644 36% /home
/dev/sdb5 960504
36180 875532 4% /tmp
/dev/sdb8 49982172
6441960 41001204 14% /usr
/dev/sdb6 4804736
1738124 2822544 39% /var
/dev/sda5 76637568
31256040 41488692 43% /media/_others
/dev/sdg1 76637568
31256040 41488692 43% /media/_others_

Is it done since they look the same? I briefly looked at the copied
data, and they seem OK? Also, its logfile:
$ cat logfile
# Rescue Logfile. Created by GNU ddrescue version 1.16
# Command line: ddrescue --force /dev/sda5 /dev/sdg1 logfile
# current_pos current_status
0x8000C000 *
# pos size status
0x00000000 0x00E18000 +
0x00E18000 0x00008000 *
0x00E20000 0x00000200 -
0x00E20200 0x00010000 *
0x00E30200 0x00000200 -
0x00E30400 0x00020000 *
0x00E50400 0x00002C00 +
0x00E53000 0x00000200 -
0x00E53200 0x0003D400 *
0x00E90600 0x00000A00 +
0x00E91000 0x00000200 -
0x00E91200 0x00079E00 *
0x00F0B000 0x00004000 +
0x00F0F000 0x00000200 -
0x00F0F200 0x000EFC00 *
0x00FFEE00 0x00000200 -
0x00FFF000 0x001DF800 *
0x011DE800 0x00009800 +
0x011E8000 0x00000200 -
0x011E8200 0x003B5800 *
0x0159DA00 0x0001D600 +
0x015BB000 0x00000200 -
0x015BB200 0x0074DA00 *
0x01D08C00 0x00000400 +
0x01D09000 0x00000200 -
0x01D09200 0x00E9B000 *
0x02BA4200 0x002B0E00 +
0x02E55000 0x00000200 -
0x02E55200 0x01A85200 *
0x048DA400 0x005B9C00 +
0x04E94000 0x00000200 -
0x04E94200 0x02F50800 *
0x07DE4A00 0x0BD9B600 +
0x13B80000 0x00010000 *
0x13B90000 0x0B9B1000 +
0x1F541000 0x0000F000 *
0x1F550000 0x04986000 +
0x23ED6000 0x0000A000 *
0x23EE0000 0x03F58000 +
0x27E38000 0x00008000 *
0x27E40000 0x03B95000 +
0x2B9D5000 0x0000B000 *
0x2B9E0000 0x05DF5000 +
0x317D5000 0x0000B000 *
0x317E0000 0x04318000 +
0x35AF8000 0x00008000 *
0x35B00000 0x0A380000 +
0x3FE80000 0x00010000 *
0x3FE90000 0x01D3A000 +
0x41BCA000 0x00006000 *
0x41BD0000 0x0A030000 +
0x4BC00000 0x00010000 *
0x4BC10000 0x000DC000 +
0x4BCEC000 0x00004000 *
0x4BCF0000 0x04310000 +
0x50000000 0x00010000 *
0x50010000 0x07A75000 +
0x57A85000 0x0000B000 *
0x57A90000 0x0852A000 +
0x5FFBA000 0x00006000 *
0x5FFC0000 0x01BE8000 +
0x61BA8000 0x00008000 *
0x61BB0000 0x008D0000 +
0x62480000 0x00010000 *
0x62490000 0x01DC4000 +
0x64254000 0x0000C000 *
0x64260000 0x003B6000 +
0x64616000 0x0000A000 *
0x64620000 0x0145F000 +
0x65A7F000 0x00001000 *
0x65A80000 0x00784000 +
0x66204000 0x0000C000 *
0x66210000 0x0032D000 +
0x6653D000 0x00003000 *
0x66540000 0x01AC0000 +
0x68000000 0x00010000 *
0x68010000 0x027F0000 +
0x6A800000 0x00010000 *
0x6A810000 0x01270000 +
0x6BA80000 0x00010000 *
0x6BA90000 0x08F70000 +
0x74A00000 0x00010000 *
0x74A10000 0x0816C000 +
0x7CB7C000 0x00004000 *
0x7CB80000 0x03482000 +
0x80002000 0x0000A000 *
0x8000C000 0x07D37000 +
0x87D43000 0x00000E00 /
0x87D43E00 0x00000200 -
0x87D44000 0x0043C000 +
0x88180000 0x00000E00 /
0x88180E00 0x00000200 -
0x88181000 0x06954000 +
0x8EAD5000 0x00000E00 /
0x8EAD5E00 0x00000200 -
0x8EAD6000 0x0832A000 +
0x96E00000 0x00000E00 /
0x96E00E00 0x00000200 -
0x96E01000 0x011FF000 +
0x98000000 0x00000E00 /
0x98000E00 0x00000200 -
0x98001000 0x06383000 +
0x9E384000 0x00000E00 /
0x9E384E00 0x00000200 -
0x9E385000 0x066FB000 +
0xA4A80000 0x00000E00 /
0xA4A80E00 0x00000200 -
0xA4A81000 0x0DDF5000 +
0xB2876000 0x00000E00 /
0xB2876E00 0x00000200 -
0xB2877000 0x03DC3000 +
0xB663A000 0x00000E00 /
0xB663AE00 0x00000200 -
0xB663B000 0x006FB000 +
0xB6D36000 0x00000E00 /
0xB6D36E00 0x00000200 -
0xB6D37000 0x021D6000 +
0xB8F0D000 0x00000E00 /
0xB8F0DE00 0x00000200 -
0xB8F0E000 0x11D70F2000 +

What do you think?
--
"What, like I had to live with all those ants? Do you know what I did to
those ants? HoooHooo! No more ants!" --unknown

/\___/\ Ant(Dude) @ http://antfarm.ma.cx (Personal Web Site)
/ /\ /\ \ Ant's Quality Foraged Links: http://aqfl.net
| |o o| |
\ _ / If crediting, then use Ant nickname and AQFL URL/link.
( ) If e-mailing, then axe ANT from its address if needed.

A song (i/wa)s playing on this computer: Tears For Fears - Mad World
(World Mix)

[Marc Haber]

Ant <a...@zimage.comANT> wrote:
>This is more complex than I expected as a newbie. So, I can't copy files
>directly like cp command?

If that fails, the file system is likely damaged in a way that would
make it hard to recover the file in the first place. Your chances are
a bit higher with an image since an image tool can try to read a bad
sector repeatedly which _might_ eventually succeed.

> I do NOT need everything from the old dying
>drive. I thought there was a way to do that to an existing HDD (don't
>want to mess up my destination drive).

You don't need to mess up the destination drive if you copy to an
image file and mount it loopback.

>Anyways, I tried that one of the unreadable file as a quick test:
>$ cp Mac\ OS\ X.dmg ~/Downloads/test
>cp: reading `Mac OS X.dmg': Input/output error
>cp: failed to extend `/home/ant/Downloads/test/Mac OS X.dmg':
>Input/output error

.dmg is a disk image? Those files are notoriously huge and therefore
extremely likely to be affected by the drive damage. With your level
of knowlegde, your best chance is having a backup.

Another probably unhelpful nitpick, your drive is not dying of SMART
errors, the SMART errors are the result of drive damage that probably
happened due to the drive's old age. SMART is only the reporter of the
bad news, and makes it possible to pull off the remains of your data
before the drive goes south for good.

Greetings
Marc
--
-------------------------------------- !! No courtesy copies, please !! -----
Marc Haber | " Questions are the | Mailadresse im Header
Mannheim, Germany | Beginning of Wisdom " | http://www.zugschlus.de/
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 621 72739834

[Marc Haber]

Ant <a...@zimage.comANT> wrote:
>On 1/4/2015 7:39 AM, Lusotec wrote:
>> https://www.gnu.org/software/ddrescue/manual/ddrescue_manual.html
>
>"If you use a device or a partition as destination, any data stored
>there will be overwritten." I don't want to erase anything on my
>destinated drive (/dev/sdb*) though.

Then don't use a device or a partition as destination.

>I just wanted to copy a few unreadable files.

If they're unreadable, how do you want to read them?

[Marc Haber]

Ant <a...@zimage.comANT> wrote:
>On 1/4/2015 2:12 PM, Rich wrote:
>
>>> Even with ddrescue? I thought it retries until it finally copies.
>>
>> If the disk sector is corrupted, no amount of retries will make it
>> "finally copy". It is simply gone. Retries are only useful when there
>> is an intermittent issue that occurs sometimes, but not other times.
>
>Ah, so http://www.kalysto.org/utilities/dd_rhelp/index.en.html is
>useless in this scenario. Interesting. Thanks.

It might help you to get the undestroyed data off the drive before it
will eventually fail completely. Destroyed data is, well, destroyed.
You might get back parts of them, but don't expect them to be remotely
intact.

[Robert Nichols]

That's bad! It says that your destination drive (sdg1) is also having
problems.

--
Bob Nichols AT comcast.net I am "RNichols42"

[Chris Davies]

In comp.os.linux.misc William Unruh <un...@invalid.ca> wrote:
> ddrescue just allows you to conitue if there are errors (it essentially

> retries once) . It does NOT keep retrying until it reads it [...]

You must have a different version of ddrescue to the one that I
use. http://www.gnu.org/software/ddrescue/ddrescue.html

Synopsis: ddrescue /path/to/broken/file /path/to/saved/copy /tmp/log.tmp

Don't omit the third parameter as ddrescue uses the "log file" to record
the current state of play with the copy.

Chris

[Ant]

I resumed this issue since I finally found free time. I decided to do an
image from scratch since people told me to do that instead, but it looks
like I run out of disk space on the destination external USB2 HDD. This
probably why it failed from partition to partition too last weekend. The
original data used is about 30 GB. Why would it need to fill up the
whole drive?

# time ddrescue -r -1 /dev/sda5 /media/storage2/CopiedDyingSDA5 ImageFileLog

GNU ddrescue 1.16
Press Ctrl-C to interrupt

rescued: 78579 MB, errsize: 119 MB, current rate: 14811 kB/s
ipos: 78698 MB, errors: 46, average rate: 12289 kB/s
opos: 78698 MB, time since last successful read: 0 s
Copying non-tried blocks...
ddrescue: write error: No space left on device

real 106m34.617s
user 0m2.536s
sys 23m53.486s


Isn't 70 GB free enough for this? How much disk space do I really need
for this? This is sure frustrating. :(
--
"Hey dad, look. I put honey on my back, and now the ants are carrying me
home." --Chris from Family Guy

[Robert Nichols]

On 01/09/2015 08:33 PM, Ant wrote:
> I resumed this issue since I finally found free time. I decided to do an image
> from scratch since people told me to do that instead, but it looks like I run
> out of disk space on the destination external USB2 HDD. This probably why it
> failed from partition to partition too last weekend. The original data used is
> about 30 GB. Why would it need to fill up the whole drive?
>
> # time ddrescue -r -1 /dev/sda5 /media/storage2/CopiedDyingSDA5 ImageFileLog
> GNU ddrescue 1.16
> Press Ctrl-C to interrupt
> rescued: 78579 MB, errsize: 119 MB, current rate: 14811 kB/s
> ipos: 78698 MB, errors: 46, average rate: 12289 kB/s
> opos: 78698 MB, time since last successful read: 0 s
> Copying non-tried blocks...
> ddrescue: write error: No space left on device
>
> real 106m34.617s
> user 0m2.536s
> sys 23m53.486s
>
>
> Isn't 70 GB free enough for this? How much disk space do I really need for this?
> This is sure frustrating. :(

How large is partition /dev/sda5? That's the whole partition, regardless of how
much space is or is not "used" in the filesystem. That is the amount of space
you need at the destination. /ddrescue/ neither knows nor cares about filesystem
structure. It copies the whole partition.

[Ant]

On 1/10/2015 6:21 AM, Robert Nichols wrote:

> How large is partition /dev/sda5? That's the whole partition, regardless
> of how
> much space is or is not "used" in the filesystem. That is the amount of
> space
> you need at the destination. /ddrescue/ neither knows nor cares about
> filesystem
> structure. It copies the whole partition.

Oh, that is why! Dang it! So, my destinated drive/partition isn't big
enough due to imaging it (not just used data) even though both drives
are the same size. :( As you requested:

$ df -h
Filesystem Size Used Avail
Use% Mounted on
rootfs 938M 717M 174M
81% /
udev 10M 0 10M
0% /dev
tmpfs 202M 6.9M 195M
4% /run
/dev/disk/by-uuid/242d81f2-8a0d-48e1-a144-22ab4d2b9407 938M 717M 174M
81% /
tmpfs 5.0M 4.0K 5.0M
1% /run/lock
tmpfs 1.2G 80K 1.2G
1% /run/shm
/dev/sdb9 48G 16G 30G
36% /home
/dev/sdb5 938M 24M 867M
3% /tmp
/dev/sdb8 48G 6.2G 40G
14% /usr
/dev/sdb6 4.6G 1.8G 2.7G
40% /var
/dev/sdg1 74G 180M 70G
1% /media/storage2
/dev/sda5 74G 30G 40G
43% /media/_others

FYI. sda5 (an ext3 partition on the dying drive I care for) is not
mounted, but /media/storage2 (ext4) is mounted in case that matters.

I do have another external USB2 drive, but it was formatted in FAT32 and
has data. Can ddrescue use it to make an image file? I don't want to
lose any data on it though.
--
"The ants sought personal revenge for my having sprayed them the day
before." --Oliver Smith

[Robert Nichols]

FAT32 can't hold a 74GB file. The maximum file size for FAT32 is 4GB
minus 1 byte. /dev/sdg1 *might* be large enough for the image if you
unmounted it and went straight to the raw partition, but that's going to
destroy the current contents of /dev/sdg1. I can't tell from the above
info. Compare the two sizes when you run "grep -E 'sda5|sdg1'
/proc/partitions" or the sizes you see when you run "fdisk -lu /dev/sda
/dev/sdg".

[Ant]

On 1/11/2015 6:04 AM, Robert Nichols wrote:
...

> FAT32 can't hold a 74GB file. The maximum file size for FAT32 is 4GB
> minus 1 byte. /dev/sdg1 *might* be large enough for the image if you
> unmounted it and went straight to the raw partition, but that's going to
> destroy the current contents of /dev/sdg1. I can't tell from the above
> info. Compare the two sizes when you run "grep -E 'sda5|sdg1'
> /proc/partitions" or the sizes you see when you run "fdisk -lu /dev/sda
> /dev/sdg".

$ df | grep -E 'sda5|sdg1'
/dev/sdg1 76921888
184068 72830340 1% /media/storage2

/dev/sda5 76637568
31256040 41488692 43% /media/_others

# fdisk -lu /dev/sda /dev/sdg

Disk /dev/sda: 80.0 GB, 80026361856 bytes
255 heads, 63 sectors/track, 9729 cylinders, total 156301488 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x6a3a6846

Device Boot Start End Blocks Id System
/dev/sda2 578340 156296384 77859022+ 5 Extended
/dev/sda5 581632 156295167 77856768 83 Linux

Disk /dev/sdg: 80.0 GB, 80026361856 bytes
255 heads, 63 sectors/track, 9729 cylinders, total 156301488 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x1c7391d2

Device Boot Start End Blocks Id System
/dev/sdg1 2048 156301311 78149632 83 Linux


As for ummounted sdg1, I think I did that when I tried to copy sda5
partition into it (was an ext4, but turned into ext3), but not as an
image. It failed as shown in
http://pastie.org/pastes/9813433/text?key=y5rqiou5fsukcgvngnamw ... I'll
try again as an image file this time and follow-up. :)
--
/\___/\ +1 Year Ant(Dude) @ http://antfarm.ma.cx (Personal Web Site)

[Ant]

It failed quickly when I tried it:

# time ddrescue -r -1 /dev/sda5 /dev/sdg1/CopiedDyingSDA5 ImageFileLog
ddrescue: Can't open output file: Not a directory

real 0m0.002s
user 0m0.000s
sys 0m0.000s


# ls -all /dev/sdg1
brw-rw---T 1 root floppy 8, 97 Jan 11 10:10 /dev/sdg1

$ df
Filesystem 1K-blocks
Used Available Use% Mounted on
rootfs 960504

733700 178012 81% /

udev 10240
0 10240 0% /dev
tmpfs 206056

7484 198572 4% /run
/dev/disk/by-uuid/242d81f2-8a0d-48e1-a144-22ab4d2b9407 960504
733700 178012 81% /
tmpfs 5120
4 5116 1% /run/lock

tmpfs 1193200
80 1193120 1% /run/shm
/dev/sdb9 50013916

16660140 30813184 36% /home
/dev/sdb5 960504
24068 887644 3% /tmp
/dev/sdb8 49982172
6445300 40997864 14% /usr
/dev/sdb6 4804736
1821408 2739260 40% /var


Both source internal drive and external USB2 destination drives were
unmounted. Did I do this wrong? :(

[Robert Nichols]

On 01/11/2015 12:20 PM, Ant wrote:
> On 1/11/2015 10:14 AM, Ant wrote:
>> Device Boot Start End Blocks Id System
>> /dev/sda2 578340 156296384 77859022+ 5 Extended
>> /dev/sda5 581632 156295167 77856768 83 Linux
>>
>> Disk /dev/sdg: 80.0 GB, 80026361856 bytes
>> 255 heads, 63 sectors/track, 9729 cylinders, total 156301488 sectors
>> Units = sectors of 1 * 512 = 512 bytes
>> Sector size (logical/physical): 512 bytes / 512 bytes
>> I/O size (minimum/optimal): 512 bytes / 512 bytes
>> Disk identifier: 0x1c7391d2
>>
>> Device Boot Start End Blocks Id System
>> /dev/sdg1 2048 156301311 78149632 83 Linux
>>
>>
>> As for ummounted sdg1, I think I did that when I tried to copy sda5
>> partition into it (was an ext4, but turned into ext3), but not as an
>> image. It failed as shown in
>> http://pastie.org/pastes/9813433/text?key=y5rqiou5fsukcgvngnamw ... I'll
>> try again as an image file this time and follow-up. :)
>
> It failed quickly when I tried it:
>
> # time ddrescue -r -1 /dev/sda5 /dev/sdg1/CopiedDyingSDA5 ImageFileLog
> ddrescue: Can't open output file: Not a directory

You want just

time ddrescue -r -1 /dev/sda5 /dev/sdg1 ImageFileLog

You don't get to add a name. When it's done, sdg1 will be a mountable
filesystem identical to however much of sda5 could be recovered. Even the
filesystem label (if any) and the UUID will be the same, so you might
want to use /tune2fs/ to change those afterward. There should be enough
space, as sdg1 is slightly larger than sda5.

[Ant]

On 1/11/2015 5:23 PM, Robert Nichols wrote:
...

>> It failed quickly when I tried it:
>>
>> # time ddrescue -r -1 /dev/sda5 /dev/sdg1/CopiedDyingSDA5 ImageFileLog
>> ddrescue: Can't open output file: Not a directory
>
> You want just
>
> time ddrescue -r -1 /dev/sda5 /dev/sdg1 ImageFileLog
>
> You don't get to add a name. When it's done, sdg1 will be a mountable
> filesystem identical to however much of sda5 could be recovered. Even the
> filesystem label (if any) and the UUID will be the same, so you might
> want to use /tune2fs/ to change those afterward. There should be enough
> space, as sdg1 is slightly larger than sda5.

Ah. Well, I ended up with a brand new 2 TB WD HDD (partitioned in half
with EXT4 FS) and my hardware friend replaced the dying HDD in my PC.
Dying HDD is my sdg(5 is the one I want to copy) and the new HDD
partition I am making an image is in /media/Storage2b/CopiedDyingStorage
and 1 TB better be enough. I hope I am doing this correctly for over a
day so far:

# time ddrescue /dev/sdg5 /media/Storage2b/CopiedDyingStorage ImageFileLog

GNU ddrescue 1.16
Press Ctrl-C to interrupt

Initial status (read from logfile)

rescued: 8641 MB, errsize: 29099 MB, errors: 42
Current status
rescued: 8641 MB, errsize: 29099 MB, current rate: 0 B/s
rescued: 8718 MB, errsize: 71006 MB, current rate: 16384 B/s
ipos: 18350 kB, errors: 56, average rate: 1041 B/s
opos: 18350 kB, time since last successful read: 0 s
Splitting failed blocks...

I saw "-rwxrwxrwx 1 root root 8.2G Jan 12 22:34 CopiedDyingStorage". It
is still going.

$ df
Filesystem 1K-blocks
Used Available Use% Mounted on
rootfs 960504

733704 178008 81% /

udev 10240
0 10240 0% /dev
tmpfs 206056

5936 200120 3% /run
/dev/disk/by-uuid/242d81f2-8a0d-48e1-a144-22ab4d2b9407 960504
733704 178008 81% /
tmpfs 5120
8 5112 1% /run/lock

tmpfs 1193200
80 1193120 1% /run/shm

/dev/sda9 50013916
2123204 45350120 5% /home
/dev/sda5 960504
24568 887144 3% /tmp
/dev/sda8 49982172
6442520 41000644 14% /usr
/dev/sda6 4804736
1801296 2759372 40% /var
/dev/sdb2 961428808
23414852 889176120 3% /media/Storage2b
/dev/sdb1 961428808
351230000 561360972 39% /media/Storage2a

I am surprised the dying HDD still works ever since it went crazy. :P
--
"Is it for pleasure you were made? Not for doing, and for action? Look
at the plants, the sparrows, the ants, spiders, bees, all doing their
business, helping to weld the order of the world. And will you refuse
man's part? And not run the way of nature's ordering?" --Marcus Aurelius
/\___/\ Ant(Dude) @ http://antfarm.ma.cx (Personal Web Site)

[Robert Nichols]

I certainly hope it gets past the bad section and speeds up. At 16384 B/s
it's going to take about 55 days to finish.

Yes, 1TB is more than enough. You should need about 80GB.

[Ant]

Yeah, I am glad it is only a 80 GB HDD. I can't imagine people trying to
recover from very big drives!! Too bad we can't recover some areas and
not the whole drive which I wanted to do. :(

# time ddrescue /dev/sdg5 /media/Storage2b/CopiedDyingStorage ImageFileLog
GNU ddrescue 1.16
Press Ctrl-C to interrupt
Initial status (read from logfile)
rescued: 8641 MB, errsize: 29099 MB, errors: 42
Current status
rescued: 8641 MB, errsize: 29099 MB, current rate: 0 B/s
rescued: 8718 MB, errsize: 71006 MB, current rate: 16384 B/s

rescued: 8719 MB, errsize: 71006 MB, current rate: 0 B/s
rescued: 8719 MB, errsize: 71006 MB, current rate: 0 B/s
rescued: 8720 MB, errsize: 71004 MB, current rate: 0 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
ipos: 32157 MB, errors: 159, average rate: 120 kB/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
ipos: 32159 MB, errors: 159, average rate: 118 kB/s
opos: 32159 MB, time since last successful read: 3.4 h
Splitting failed blocks...


pydemsg:
...
[2015-01-14 08:24:41] sd 14:0:0:0: [sdg] Sense Key : Not Ready [current]
[2015-01-14 08:24:41] sd 14:0:0:0: [sdg] Add. Sense: Medium not present
[2015-01-14 08:24:41] sd 14:0:0:0: [sdg] CDB: Read(10): 28 00 03 c7 4e
c8 00 00 08 00
[2015-01-14 08:24:41] end_request: I/O error, dev sdg, sector 63393480
[2015-01-14 08:24:41] Buffer I/O error on device sdg5, logical block 7851481
[2015-01-14 08:25:12] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:25:43] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:26:13] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:26:44] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:27:15] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:27:47] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:27:47] sd 14:0:0:0: [sdg] Device not ready
[2015-01-14 08:27:47] sd 14:0:0:0: [sdg] Result: hostbyte=DID_ABORT
driverbyte=DRIVER_SENSE
[2015-01-14 08:27:47] sd 14:0:0:0: [sdg] Sense Key : Not Ready [current]
[2015-01-14 08:27:47] sd 14:0:0:0: [sdg] Add. Sense: Medium not present
[2015-01-14 08:27:47] sd 14:0:0:0: [sdg] CDB: Read(10): 28 00 03 c7 4e
c8 00 00 08 00
[2015-01-14 08:27:47] end_request: I/O error, dev sdg, sector 63393480
[2015-01-14 08:27:47] Buffer I/O error on device sdg5, logical block 7851481
[2015-01-14 08:28:18] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:28:48] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:29:19] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:29:50] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:30:21] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:30:53] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
[2015-01-14 08:30:53] sd 14:0:0:0: [sdg] Device not ready
[2015-01-14 08:30:53] sd 14:0:0:0: [sdg] Result: hostbyte=DID_ABORT
driverbyte=DRIVER_SENSE
[2015-01-14 08:30:53] sd 14:0:0:0: [sdg] Sense Key : Not Ready [current]
[2015-01-14 08:30:53] sd 14:0:0:0: [sdg] Add. Sense: Medium not present
[2015-01-14 08:30:53] sd 14:0:0:0: [sdg] CDB: Read(10): 28 00 03 c7 4e
c8 00 00 08 00
[2015-01-14 08:30:53] end_request: I/O error, dev sdg, sector 63393480
[2015-01-14 08:30:53] Buffer I/O error on device sdg5, logical block 7851481
[2015-01-14 08:31:23] usb 2-2: reset high-speed USB device number 6
using ehci_hcd
...
--
"Ants can attack with a grain of rice." --a Malagasy Proverb

[William Unruh]

On 2015-01-14, Ant <a...@zimage.comANT> wrote:
> On 1/13/2015 6:39 AM, Robert Nichols wrote:
>
>>> Ah. Well, I ended up with a brand new 2 TB WD HDD (partitioned in half
>>> with EXT4 FS) and my hardware friend replaced the dying HDD in my PC.
>>> Dying HDD is my
>>> sdg(5 is the one I want to copy) and the new HDD partition I am making
>>> an image is in /media/Storage2b/CopiedDyingStorage and 1 TB better be
>>> enough. I hope
>>> I am doing this correctly for over a day so far:
>>>
>>> # time ddrescue /dev/sdg5 /media/Storage2b/CopiedDyingStorage
>>> ImageFileLog
>>>
>>> GNU ddrescue 1.16
>>> Press Ctrl-C to interrupt
>>> Initial status (read from logfile)
>>> rescued: 8641 MB, errsize: 29099 MB, errors: 42
>>> Current status
>>> rescued: 8641 MB, errsize: 29099 MB, current rate: 0 B/s
>>> rescued: 8718 MB, errsize: 71006 MB, current rate: 16384 B/s
>>> ipos: 18350 kB, errors: 56, average rate: 1041 B/s
>>> opos: 18350 kB, time since last successful read: 0 s
>>> Splitting failed blocks...

I think that the reason it is taking so long is taht your disk is really
really borked-- ie loads of it is unreadable. It keeps trying a few
times to reread each sector, and cannot. I think you will need to wave
goodbye to all of the data on that disk.
If the disk were good, it would be reading it at many MB/sec.

[Robert Nichols]

On 01/14/2015 10:33 AM, Ant wrote:
> Yeah, I am glad it is only a 80 GB HDD. I can't imagine people trying to recover
> from very big drives!! Too bad we can't recover some areas and not the whole
> drive which I wanted to do. :(

You have some control over it, but you have to have some idea of what's
where. <http://www.gnu.org/software/ddrescue/manual/ddrescue_manual.html#Algorithm>

[Ant]

On 1/14/2015 9:05 PM, Robert Nichols wrote:

>> Yeah, I am glad it is only a 80 GB HDD. I can't imagine people trying
>> to recover
>> from very big drives!! Too bad we can't recover some areas and not the
>> whole
>> drive which I wanted to do. :(
>
> You have some control over it, but you have to have some idea of what's
> where.
> <http://www.gnu.org/software/ddrescue/manual/ddrescue_manual.html#Algorithm>

Ah, interesting. Oh well, I'll let it keep going until the disk is
completely dead or ddrescue is finally done. So far, it
(finish/complete)d 30 GB out of 80.
--
"Above ground I shall be food for kites; below I shall be food for
mole-crickets and ants. Why rob one to feed the other?" --Juang-zu (4th
Century B.C.)

[Ant]

Oh well, I'll let it keep going until the disk is completely dead or
ddrescue is finally done. So far, it (finish/complete)d 30 GB out of 80

GB. This would be a good practice for me anyways. Either way, I need to
destroy this old HDD. ;)
--
"Ants can lift up to 50 times their own weight. And your monitor is
missing. Time to bring out the bugspray." --BBspot's Geek Horoscopes
(2/28/2003)

[Ant]

> Oh well, I'll let it keep going until the disk is completely dead or
> ddrescue is finally done. So far, it (finish/complete)d 30 GB out of 80
> GB. This would be a good practice for me anyways. Either way, I need to
> destroy this old HDD. ;)

OK, I think it is officially dead:

# time ddrescue /dev/sdg5 /media/Storage2b/CopiedDyingStorage ImageFileLog
GNU ddrescue 1.16
Press Ctrl-C to interrupt
Initial status (read from logfile)
rescued: 8641 MB, errsize: 29099 MB, errors: 42
Current status
rescued: 8641 MB, errsize: 29099 MB, current rate: 0 B/s
rescued: 8718 MB, errsize: 71006 MB, current rate: 16384 B/s

rescued: 8719 MB, errsize: 71006 MB, current rate: 0 B/s
rescued: 8719 MB, errsize: 71006 MB, current rate: 0 B/s
rescued: 8720 MB, errsize: 71004 MB, current rate: 0 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
ipos: 32157 MB, errors: 159, average rate: 120 kB/s

rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
ipos: 34550 MB, errors: 159, average rate: 90752 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
ipos: 35348 MB, errors: 159, average rate: 82342 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
ipos: 35348 MB, errors: 159, average rate: 82288 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
ipos: 38008 MB, errors: 159, average rate: 47700 B/s
rescued: 31944 MB, errsize: 47780 MB, current rate: 0 B/s
ipos: 38008 MB, errors: 159, average rate: 47682 B/s
opos: 38008 MB, time since last successful read: 3.5 d
Splitting failed blocks...
Interrupted by user

real 8145m19.830s
user 0m35.534s
sys 4m24.917s

3.5 days it last read. I aborted it. It is not end of the world since I
do have backups from other computers and drives. At least I learned how
to use ddrescue in case I run into this again the future (use ddrescue
to make an image!)! ;)

Actually, since I have an image. Does that mean I can still access the
data in it even though it is only 30 GB out of 80 GB? I wonder how much
can be accessed from it.
--
"Your parents were killed by ants?" --Idle Hands movie

[Robert Nichols]

On 01/17/2015 07:59 PM, Ant wrote:
> 3.5 days it last read. I aborted it. It is not end of the world since I do have
> backups from other computers and drives. At least I learned how to use ddrescue
> in case I run into this again the future (use ddrescue to make an image!)! ;)
>
> Actually, since I have an image. Does that mean I can still access the data in
> it even though it is only 30 GB out of 80 GB? I wonder how much can be accessed
> from it.

Hard to say, but sounds doubtful. You'd need to make another copy of the
image, run "fsck -y" on that, and then see what can be recovered. You
could _try_ running /photorec/ on the current image, but between the loss
of all the file names and the mass of corrupted files and fragments it
would find, it strikes me a strictly an act of desperation.

[Ant]

Alright and thanks. Just wnated to see what it is like to a pretty bad
dead drive that was recovered. Learning experiences. :)
--
This worker ant will be unemployed after Friday, the 23rd, as of
1/15/2015 @ 11 AM PST. :(

[Hactar]

In article <BtednS378fbxBzTJ...@giganews.com>,
Robert Heller <hel...@deepsoft.com> wrote:

> At Sun, 4 Jan 2015 18:27:18 +0000 (UTC) Rich <ri...@example.invalid> wrote:
>
> >
> > In comp.os.linux.misc Ant <a...@zimage.comant> wrote:

> > > On 1/4/2015 7:23 AM, Robert Heller wrote:
> >
> > > > dd can do that:
> > > >
> > > > dd ... conv=noerror
> > > >
> > > > man dd for full details.
> > > >
> > > > Basically, you do a raw DD of the 'bad' disk's file system (eg
> create a copy
> > > > of the file system either onto a new disk partition (of the proper
> size) or to
> > > > an image file. Then you can mount the partition (or image via -o
> loop) and
> > > > then copy the files off. Note: files with bad blocks will likely
> be 'broken'
> > > > and may not be usable, but you should be able to retrieve other files.

> >
> > > This is more complex than I expected as a newbie. So, I can't copy files

> > > directly like cp command? I do NOT need everything from the old dying

> > > drive. I thought there was a way to do that to an existing HDD (don't
> > > want to mess up my destination drive).
> >

> > dd in="/old/drive/Mac OS X.dmg" out="/new/drive/that/is/not/broken/Mac
> OS X.dmg" conv=noerror
> >
> > should work to "copy" a single file. Change "in=" and "out=" files to
> > point to the ones you are trying to copy.
> >
> > Note, that given the "errors" when reading the files, and the fact that
> > the example you gave is a mac dmg file, the result you get from the
> > "copy" process will likely be useless anyway.
>
> Yes indeed. The 'best' option is to copy the whole file system (using dd
> conv=noerror) to a scratch disk or to a file image. Then go though the copy to
> pick out the files that are still usable. The OP seemed to want to copy
> everything still usable. Doing dd on each file is very tedious, since dd does
> not work with wildcards or multiple files.

No, but you can do something like this:

find "$srcdir" -type f | while read filename ; do
dd if="$filename" of="$dest/$filename" conv=noerror
done

But I think you'd be better off using dd_rescue.

--
-eben QebWe...@vTerYizUonI.nOetP ebmanda.redirectme.net:81
LIBRA: A big promotion is just around the corner for someone
much more talented than you. Laughter is the very best medicine,
remember that when your appendix bursts next week. -- Weird Al

[Hactar]

In article <m968gb$ng5$1...@dont-email.me>,

I recovered files from a "dead" HD for my friend. I forget how big it
was but I used essentially this method. It was slow reading the source
and there was lots of clicking, but mounting the image and copying from
it ran at normal speeds.

[Ant]

I somewhat gave up on it. It was too slow and not moving:

22499 Jan 17 17:44 ImageFileLog
31987306496 Jan 14 05:03 CopiedDyingStorageImage

3 days of nothing! I still have the "dead" drive so I could retry if I
wanted to.
--
"The ants are my friends, they're blowin' in the wind. The ant, sir, is
blowin' in the wind." --the misheard lyrics to Bob Dylan's "Blowin' in
the Wind"