Groups > comp.os.linux.misc > #13209 > unrolled thread

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

Started by	Ant <ant@zimage.comANT>
First post	2015-01-04 06:55 -0800
Last post	2015-01-05 14:13 +0100
Articles	20 on this page of 55 — 11 participants

Back to article view | Back to comp.os.linux.misc

Page 1 of 3 [1] 2 3 Next page →

#13209 — How to force copy retries until finally copied all due to a dying old HDD in Linux?

From	Ant <ant@zimage.comANT>
Date	2015-01-04 06:55 -0800
Subject	How to force copy retries until finally copied all due to a dying old HDD in Linux?
Message-ID	<wsednYGSWe_iyTTJnZ2dnUU7-UednZ2d@earthlink.com>

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.

[toc] | [next] | [standalone]

#13211 — Re: How to force copy retries until finally copied all due to a dying old HDD in Linux?

From	Robert Heller <heller@deepsoft.com>
Date	2015-01-04 09:23 -0600
Subject	Re: How to force copy retries until finally copied all due to a dying old HDD in Linux?
Message-ID	<nZidnadeyqNIxzTJnZ2dnUU7-N2dnZ2d@giganews.com>
In reply to	#13209

At Sun, 04 Jan 2015 06:55:59 -0800 Ant <ant@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
heller@deepsoft.com       -- Webhosting Services

[toc] | [prev] | [next] | [standalone]

#13214

From	Ant <ant@zimage.comANT>
Date	2015-01-04 09:19 -0800
Message-ID	<La2dncl-EqmA6zTJnZ2dnUU7-NmdnZ2d@earthlink.com>
In reply to	#13211

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(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.

[toc] | [prev] | [next] | [standalone]

#13225

From	Rich <rich@example.invalid>
Date	2015-01-04 18:27 +0000
Message-ID	<m8c0m6$p8f$2@dont-email.me>
In reply to	#13214

In comp.os.linux.misc Ant <ant@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.

[toc] | [prev] | [next] | [standalone]

#13226

From	Robert Heller <heller@deepsoft.com>
Date	2015-01-04 13:54 -0600
Message-ID	<BtednS378fbxBzTJnZ2dnUU7-bmdnZ2d@giganews.com>
In reply to	#13225

At Sun, 4 Jan 2015 18:27:18 +0000 (UTC) Rich <rich@example.invalid> wrote:

> 
> In comp.os.linux.misc Ant <ant@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.

>            

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

[toc] | [prev] | [next] | [standalone]

#13705

From	ebenZEROONE@verizon.net (Hactar)
Date	2015-02-19 19:00 -0500
Message-ID	<n1airb-lnv.ln1@pc.home>
In reply to	#13226

In article <BtednS378fbxBzTJnZ2dnUU7-bmdnZ2d@giganews.com>,
Robert Heller  <heller@deepsoft.com> wrote:
> At Sun, 4 Jan 2015 18:27:18 +0000 (UTC) Rich <rich@example.invalid> wrote:
> 
> > 
> > In comp.os.linux.misc Ant <ant@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    QebWenE01R@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

[toc] | [prev] | [next] | [standalone]

#13259

From	Marc Haber <mh+usenetspam1118@zugschl.us>
Date	2015-01-05 14:10 +0100
Message-ID	<m8e2gg$kb2$1@news1.tnib.de>
In reply to	#13214

Ant <ant@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

[toc] | [prev] | [next] | [standalone]

#13213

From	Lusotec <nomail@nomail.not>
Date	2015-01-04 15:39 +0000
Message-ID	<m8bmr9$chl$1@dont-email.me>
In reply to	#13209

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

[toc] | [prev] | [next] | [standalone]

#13215

From	Ant <ant@zimage.comANT>
Date	2015-01-04 09:25 -0800
Message-ID	<vOudnV92u-I-6jTJnZ2dnUU7-fWdnZ2d@earthlink.com>
In reply to	#13213

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)
    /\___/\         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.

[toc] | [prev] | [next] | [standalone]

#13217

From	Lusotec <nomail@nomail.not>
Date	2015-01-04 17:49 +0000
Message-ID	<m8bufc$ffs$1@dont-email.me>
In reply to	#13215

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

[toc] | [prev] | [next] | [standalone]

#13227

From	Ant <ant@zimage.comANT>
Date	2015-01-04 13:07 -0800
Message-ID	<8J6dnYTNS9IJNjTJnZ2dnUU7-T2dnZ2d@earthlink.com>
In reply to	#13217

>>> 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.

OK. I need to make free disk space before I do that.


>> 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

Is that safe to use in Debian stable (not Ubuntu)? Debian's official 
servers don't have it according to its 
https://packages.debian.org/search?keywords=DDRescue-GUI+&searchon=names&suite=all&section=all 
...
-- 
"The ants are a people not strong, yet they prepare their meat in the 
summer." --Proverbs 30:25 (Bible)
    /\___/\         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.

[toc] | [prev] | [next] | [standalone]

#13230

From	William Unruh <unruh@invalid.ca>
Date	2015-01-04 21:57 +0000
Message-ID	<m8cd0d$bq4$1@dont-email.me>
In reply to	#13227

On 2015-01-04, Ant <ant@zimage.comANT> 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.
>
> OK. I need to make free disk space before I do that.

If you are that short of disk space, go out and buy a new disk! You need
backups anyway. 

>
>
>>> 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
>
> Is that safe to use in Debian stable (not Ubuntu)? Debian's official 
> servers don't have it according to its 
> https://packages.debian.org/search?keywords=DDRescue-GUI+&searchon=names&suite=all&section=all 
> ...

[toc] | [prev] | [next] | [standalone]

#13391

From	Ant <ant@zimage.comANT>
Date	2015-01-09 18:32 -0800
Message-ID	<D-adnXYAA_GjEi3JnZ2dnUU7-e-dnZ2d@earthlink.com>
In reply to	#13217

On 1/4/2015 1:07 PM, Ant 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.
>
> OK. I need to make free disk space before I do that.

I resumed 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. :(
-- 
"To conquer the world, we must be as meticulous and calculating as a 
colony of ants on the march." --Julius Caesar
    /\___/\         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.

[toc] | [prev] | [next] | [standalone]

#13393

From	Gernot Fink <g.fink@gmx.net>
Date	2015-01-10 08:42 +0100
Message-ID	<1p17ob-594.ln1@garv.home>
In reply to	#13391

In article <D-adnXYAA_GjEi3JnZ2dnUU7-e-dnZ2d@earthlink.com>,
	Ant <ant@zimage.comANT> writes:
> 
> # time ddrescue -r -1 /dev/sda5 /media/storage2/CopiedDyingSDA5 ImageFileLog
> Isn't 70 GB free enough for this? How much disk space do I really need 
> for this? This is sure frustrating. :(

ddrescue create a mountable 1/1 image of the bad partition.
You need exact the the same space for source and destination.

Gernot

[toc] | [prev] | [next] | [standalone]

#13396

From	Ant <ant@zimage.comANT>
Date	2015-01-10 07:51 -0800
Message-ID	<KN6dnarjy8MC1yzJnZ2dnUU7-cOdnZ2d@earthlink.com>
In reply to	#13393

>> # time ddrescue -r -1 /dev/sda5 /media/storage2/CopiedDyingSDA5 ImageFileLog
>> Isn't 70 GB free enough for this? How much disk space do I really need
>> for this? This is sure frustrating. :(
>
> ddrescue create a mountable 1/1 image of the bad partition.
> You need exact the the same space for source and destination.

The 80 GB HDDs (dying sda and external USB sdg) have almost the same sizes:
$ df
Filesystem                                             1K-blocks 
Used Available Use% Mounted on
rootfs                                                    960504 
733648    178064  81% /
udev                                                       10240 
0     10240   0% /dev
tmpfs                                                     206056 
7180    198876   4% /run
/dev/disk/by-uuid/242d81f2-8a0d-48e1-a144-22ab4d2b9407    960504 
733648    178064  81% /
tmpfs                                                       5120 
4      5116   1% /run/lock
tmpfs                                                    1193200 
80   1193120   1% /run/shm
/dev/sdb9                                               50013916 
16657464  30815860  36% /home
/dev/sdb5                                                 960504 
23984    887728   3% /tmp
/dev/sdb8                                               49982172 
6442208  41000956  14% /usr
/dev/sdb6                                                4804736 
1810652   2750016  40% /var
/dev/sdg1                                               76921888 
184068  72830340   1% /media/storage2
/dev/sda5                                               76637568 
31256040  41488692  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 is in FAT32 and NTFS. Can 
ddrescue use it to make an image file? I don't want to lose any data on 
it though.
-- 
"Only two great groups of animals, men and ants, indulge in highly 
organized mass warfare." --Charles H. Maskins
    /\___/\         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.

[toc] | [prev] | [next] | [standalone]

#13397

From	Gernot Fink <g.fink@gmx.net>
Date	2015-01-10 20:38 +0100
Message-ID	<bnb8ob-rp6.ln1@garv.home>
In reply to	#13396

In article <KN6dnarjy8MC1yzJnZ2dnUU7-cOdnZ2d@earthlink.com>,
	Ant <ant@zimage.comANT> writes:
>>
>> ddrescue create a mountable 1/1 image of the bad partition.
>> You need exact the the same space for source and destination.
> 
> The 80 GB HDDs (dying sda and external USB sdg) have almost the same sizes:
> $ df

The target filesystem/device must have room for all source blocks.

FAT dont work because it has a 2G file size limit.
NTFS can work but the linux implementation is slow.

Gernot

[toc] | [prev] | [next] | [standalone]

#13399

From	Ant <ant@zimage.comANT>
Date	2015-01-10 17:24 -0800
Message-ID	<hfOdnYiyfpJSTSzJnZ2dnUU7-e-dnZ2d@earthlink.com>
In reply to	#13397

On 1/10/2015 11:38 AM, Gernot Fink wrote:

>>> ddrescue create a mountable 1/1 image of the bad partition.
>>> You need exact the the same space for source and destination.
>>
>> The 80 GB HDDs (dying sda and external USB sdg) have almost the same sizes:
>> $ df
>
> The target filesystem/device must have room for all source blocks.

I guess 80 GB isn't big enough even though it is almost the same size 
(destination drive is slighty bigger). Argh.


> FAT dont work because it has a 2G file size limit.
> NTFS can work but the linux implementation is slow.

Darn and thanks. :(
-- 
"For example, the tiny ant, a creature of great industry, drags with its 
mouth whatever it can, and adds it to the heap which she is piling up, 
not unaware nor careless of the future." --Horace, Satires, Book I, I, 33.
    /\___/\         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.

[toc] | [prev] | [next] | [standalone]

#13401

From	Rich <rich@example.invalid>
Date	2015-01-10 03:35 +0000
Message-ID	<m8q6mt$j9t$1@dont-email.me>
In reply to	#13391

Ant <ant@zimage.comant> wrote:
> On 1/4/2015 1:07 PM, Ant wrote:

> > OK. I need to make free disk space before I do that.

> I resumed 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. 
> ...
> Isn't 70 GB free enough for this? How much disk space do I really need 
> for this?

For an image, you need enough space to store the raw size of the
disk or partition you are reading from.  An image is an exact copy of
each sector of the disk, without regard to whether it is allocated to a
file or free space in the filesystem on the disk/partition.

So, the answer to "how must space [is needed]" is exactly the size of
the disk or the partition.

[toc] | [prev] | [next] | [standalone]

#13402

From	William Unruh <unruh@invalid.ca>
Date	2015-01-10 03:54 +0000
Message-ID	<m8q7qj$ldn$1@dont-email.me>
In reply to	#13391

On 2015-01-10, Ant <ant@zimage.comANT> wrote:
> On 1/4/2015 1:07 PM, Ant 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.
>>
>> OK. I need to make free disk space before I do that.
>
> I resumed 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?

Because dd copies every byte from the drive to the other one, whether
those bytes are parts of files or not. Thus, you need a target that is
at least as large as the original full drive. If you have a 10TB drive,
even if it only has 2MB on it, you need at least a 10TB target 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. :(

[toc] | [prev] | [next] | [standalone]

#13221

From	William Unruh <unruh@invalid.ca>
Date	2015-01-04 18:12 +0000
Message-ID	<m8bvqi$h4v$4@dont-email.me>
In reply to	#13215

On 2015-01-04, Ant <ant@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.

[toc] | [prev] | [next] | [standalone]

Page 1 of 3 [1] 2 3 Next page →

csiph-web

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

Contents

#13209 — How to force copy retries until finally copied all due to a dying old HDD in Linux?

#13211 — Re: How to force copy retries until finally copied all due to a dying old HDD in Linux?

#13214

#13225

#13226

#13705

#13259

#13213

#13215

#13217

#13227

#13230

#13391

#13393

#13396

#13397

#13399

#13401

#13402

#13221