bug-ddrescue
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Bug-ddrescue] speed and AHCI vs IDE, 64bit vs. 32-bit


From: Eric Stout
Subject: [Bug-ddrescue] speed and AHCI vs IDE, 64bit vs. 32-bit
Date: Fri, 16 Dec 2011 17:13:44 -0800 (PST)

Hello,

I am trying to recover what I can off a 500GB Western Digital 2.5" SATA drive.  Unfortunately I don't have much time.  (HP has sent a replacement drive under the notebook warranty and says they will charge my credit card over $200 if they don't receive the failing HD in a few days.)

I've been running "ddrescue -n -f /dev/sdc /dev/sdb logfile" using a SystemRescueCD for a few hours.  It proceeding at an average rate of less than 2mB/s and falling.  (The current status includes 28GB rescued, 341 errors, 12MB errsize.)  I'm afraid that it won't finish even the first pass in time.  

Is there anything I can do to speed this up?

I've mounted the failing and destination SATA drives internally in a desktop (Z68 chipset, i5-2500k processor).  Does it matter if the BIOS is set to run the SATA drives using RAID (XHD), AHCI, or IDE?  I had it on (software) RAID (XHD), ran it for a while then rebooted using AHCI.  I don't know what the HP G72 notebook that the failing HD came from was using.  Will ddrescue work better with AHCI, IDE, or RAID (XHD)?  Would it be bad to switch between them and restarting ddrescue (using the same destination and logfile).

SystemRescueCD (2.4.1) gives me a choice of running 32 bit or 64 bit kernels.  Will this make a difference?

Would it be faster or better to use a different system?  I could mount the failing HD in a Core2 duo or i3-370m notebook and put the destination drive in a USB 2.0 enclosure.

Is writing to an image file on an ext4 partition faster?  Does formatting the destination device if the destination is a whole device help or make any difference?

I love ddrescue and have used it successfully before.  I really appreciate everything you're doing.  Should I just suck it up and be happy with an average rate over 1MB/s (and only getting to half the drive)?

Thank you,
Eric


Thank you
--- On Wed, 12/7/11, William F. <address@hidden> wrote:

From: William F. <address@hidden>
Subject: Re: [Bug-ddrescue] Bug-ddrescue Digest, Vol 71, Issue 1
To: address@hidden
Date: Wednesday, December 7, 2011, 9:22 AM

Looks like media error. Some kind of noise? Was the HDD detected by BIOS?

Regards.


Em 07-12-2011 15:00, address@hidden escreveu:
> Send Bug-ddrescue mailing list submissions to
>     address@hidden
>
> To subscribe or unsubscribe via the World Wide Web, visit
>     https://lists.gnu.org/mailman/listinfo/bug-ddrescue
> or, via email, send a message with subject or body 'help' to
>     address@hidden
>
> You can reach the person managing the list at
>     address@hidden
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Bug-ddrescue digest..."
>
>
> Today's Topics:
>
>     1. Question about reported errorsize (Tom Williams)
>     2. Re: Version 1.15-rc1 of GNU ddrescue released (Tom Williams)
>     3. Re: Question about reported errorsize (andrew zajac)
>     4. Re: Question about reported errorsize (Tom Williams)
>     5. Request for example of rescuing a dvd data disk (Bob)
>     6. Re: Version 1.15-rc1 of GNU ddrescue released (Antonio Diaz Diaz)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Tue, 06 Dec 2011 09:26:03 -0800
> From: Tom Williams<address@hidden>
> To: address@hidden
> Subject: [Bug-ddrescue] Question about reported errorsize
> Message-ID:<address@hidden>
> Content-Type: text/plain; charset=ISO-8859-1
>
> I'm running ddrescue from a Trinity Rescue CD (3.4-bld-372) on a 500GB
> hard drive with two NTFS partitions on it.  The hard drive is dying and
> I'm hoping to move the data to a new hard drive so it can be saved.
>
> I've issued his command:
>
> ddrescue /dev/sda /dev/sdb
>
> and the process is underway now.
>
> ddrescue has reported only ONE error and the errorsize is 500GB.  It has
> 51MB of data and is in the "splitting" phase now. It looks as if it's
> going to try to split the 500GB of "errors" reported in "errorsize".
>
> My question:  does the fact that 500GB was reported as the errorsize
> mean this process is most likely going to fail?  Is it worth spending
> the time to go through the "splitting" phase?
>
> I believe I'm running ddrescue 1.14, but I'm not sure.
>
> Thanks!
>
> Peace...
>
> Tom
>
>
>
> ------------------------------
>
> Message: 2
> Date: Tue, 06 Dec 2011 09:56:10 -0800
> From: Tom Williams<address@hidden>
> To: address@hidden
> Subject: Re: [Bug-ddrescue] Version 1.15-rc1 of GNU ddrescue released
> Message-ID:<address@hidden>
> Content-Type: text/plain; charset=ISO-8859-1
>
> I have worked up a small patch to add the ddrescue version number to the
> "show help" (-h) output.  This way, when someone runs "ddrescue -h", the
> version of ddrescue being run is included in the output.  :)
>
> Here is the patch:
>
> -------------------- START ----------------------------
>
> address@hidden:~/build/ddrescue-1.15-rc1$ diff -up main-orig.cc main.cc
> --- main-orig.cc        2011-12-06 09:45:04.000000000 -0800
> +++ main.cc     2011-12-06 09:47:48.000000000 -0800
> @@ -60,7 +60,7 @@ const int o_binary = 0;
>
>   void show_help( const int cluster, const int hardbs ) throw()
>     {
> -  std::printf( "%s - Data recovery tool.\n", Program_name );
> +  std::printf( "%s %s - Data recovery tool.\n", Program_name,
> PROGVERSION );
>     std::printf( "Copies data from one file or block device to another,\n"
>                  "trying hard to rescue data in case of read errors.\n"
>                  "\nUsage: %s [options] infile outfile [logfile]\n",
> invocation_name );
> [1]+  Done                    gedit arg_parser.cc
> address@hidden:~/build/ddrescue-1.15-rc1$
>
> ---------------------- END -----------------------------
>
> Peace...
>
> Tom
>
> On 11/23/2011 07:28 AM, Antonio Diaz Diaz wrote:
>> Version 1.15-rc1 of GNU ddrescue is ready for testing here
>> http://download.savannah.gnu.org/releases/ddrescue/ddrescue-1.15-rc1.tar.gz
>>
>> http://download.savannah.gnu.org/releases/ddrescue/ddrescue-1.15-rc1.tar.lz
>>
>>
>> The md5sums are:
>> 0d4182237e5bae450530f919c60ebed7  ddrescue-1.15-rc1.tar.gz
>> 9fbbf10cf008aed6765d1e24925a8169  ddrescue-1.15-rc1.tar.lz
>>
>> Please, test it and report any bugs you find.
>>
>> GNU ddrescue is a data recovery tool. It copies data from one file or
>> block device (hard disc, cdrom, etc) to another, trying hard to rescue
>> data in case of read errors.
>>
>> GNU Ddrescuelog is a tool that manipulates ddrescue logfiles, shows
>> logfile contents, converts logfiles to/from other formats, compares
>> logfiles, tests rescue status, and can delete a logfile if the rescue
>> is done.
>>
>> The homepage is at http://www.gnu.org/software/ddrescue/ddrescue.html.
>>
>> This version is also available in lzip format. If your distro doesn't
>> yet distribute the lzip program, you can download it from
>> http://www.nongnu.org/lzip/lzip.html
>>
>>
>> Changes in this version:
>>
>>    * The new option "-I, --verify-input-size" has been added.
>>
>>    * The new option "-x, --extend-outfile" has been added.
>>
>>    * Ddrescue now verifies that infile, outfile and logfile are all
>> different.
>>
>>    * Non-tried blocks are now read aligned to cluster-size sectors.
>>
>>    * The "split or skip" algorithm of the split pass has been improved.
>>
>>    * The default block size for ddrescuelog has been changed to 512,
>> the value used by ddrescue.
>>
>>
>> Regards,
>> Antonio Diaz, GNU ddrescue author and maintainer.
>>
>>
>> _______________________________________________
>> Bug-ddrescue mailing list
>> address@hidden
>> https://lists.gnu.org/mailman/listinfo/bug-ddrescue
>>
>
>
>
> ------------------------------
>
> Message: 3
> Date: Tue, 6 Dec 2011 11:28:43 -0800 (PST)
> From: andrew zajac<address@hidden>
> To: address@hidden, Tom Williams<address@hidden>
> Subject: Re: [Bug-ddrescue] Question about reported errorsize
> Message-ID:
>     <address@hidden>
> Content-Type: text/plain; charset=iso-8859-1
>
> Hi Tom.
>
> That's a good question.
>
> First off, I like to see what's going on while I read a faulty drive.  In another console, I usually run something like:
> tail -f /var/log/kernel
>
> and that shows me what the kernel sees - I pay attention to devices dropping offline.  In this case, I would wonder if the drive went offline but the kernel didn't notice.  This happens sometimes when the kernel can see the drive controller even if there is no drive attached, as with some USB-SATA connections.
>
> That being said, I like those kinds of interfaces because you can hotplug the drive and potentially bring it back online without power cycling it.
>
> I would run Gnu ddrescue in the same way, but I would use a log.  That way, you can stop and resume the recover.  I regularly have to physically unplug and plug back in the USB connector to a drive that has dropped out like that.  I find some drives tend to faint easily at high speeds, which makes the slower USB speeds an asset in these cases.  Using -D or -d switches can also slow things down and regulate speed in a useful way.
>
> Good luck.
>
> Andrew Zajac
> AndrewZajac.ca
> Ubuntu-Rescue-Remix.org
>
>
>
>
> --- On Tue, 12/6/11, Tom Williams<address@hidden>  wrote:
>
>> From: Tom Williams<address@hidden>
>> Subject: [Bug-ddrescue] Question about reported errorsize
>> To: address@hidden
>> Received: Tuesday, December 6, 2011, 12:26 PM
>> I'm running ddrescue from a Trinity
>> Rescue CD (3.4-bld-372) on a 500GB
>> hard drive with two NTFS partitions on it.? The hard
>> drive is dying and
>> I'm hoping to move the data to a new hard drive so it can
>> be saved.
>>
>> I've issued his command:
>>
>> ddrescue /dev/sda /dev/sdb
>>
>> and the process is underway now.
>>
>> ddrescue has reported only ONE error and the errorsize is
>> 500GB.? It has
>> 51MB of data and is in the "splitting" phase now. It looks
>> as if it's
>> going to try to split the 500GB of "errors" reported in
>> "errorsize".
>>
>> My question:? does the fact that 500GB was reported as
>> the errorsize
>> mean this process is most likely going to fail?? Is it
>> worth spending
>> the time to go through the "splitting" phase?
>>
>> I believe I'm running ddrescue 1.14, but I'm not sure.
>>
>> Thanks!
>>
>> Peace...
>>
>> Tom
>>
>> _______________________________________________
>> Bug-ddrescue mailing list
>> address@hidden
>> https://lists.gnu.org/mailman/listinfo/bug-ddrescue
>>
>
>
> ------------------------------
>
> Message: 4
> Date: Tue, 06 Dec 2011 13:16:18 -0800
> From: Tom Williams<address@hidden>
> Cc: address@hidden
> Subject: Re: [Bug-ddrescue] Question about reported errorsize
> Message-ID:<address@hidden>
> Content-Type: text/plain; charset=ISO-8859-1
>
> Thanks for the tip.  :)
>
> I used Alt-F2 to open a new console and was able to look at the kernel
> log.  As expected, there are a TON of errors reading the "problem"
> drive.  I'm seeing errors with "STATUS {DRDY ERR}" and failed command
> "READ DMA" in the log.  The end of the log has messages that read
> "Buffer I/O error on device sda, logical block {nnnnn}"
>
> The "problem" drive is connected to my computer internally and the drive
> I'm trying to rescue to is an external USB drive.
>
> I'm planning on letting it run to see what happens. :)
>
> Thanks!
>
> Peace...
>
> Tom
>
> On 12/06/2011 11:28 AM, andrew zajac wrote:
>> Hi Tom.
>>
>> That's a good question.
>>
>> First off, I like to see what's going on while I read a faulty drive.  In another console, I usually run something like:
>> tail -f /var/log/kernel
>>
>> and that shows me what the kernel sees - I pay attention to devices dropping offline.  In this case, I would wonder if the drive went offline but the kernel didn't notice.  This happens sometimes when the kernel can see the drive controller even if there is no drive attached, as with some USB-SATA connections.
>>
>> That being said, I like those kinds of interfaces because you can hotplug the drive and potentially bring it back online without power cycling it.
>>
>> I would run Gnu ddrescue in the same way, but I would use a log.  That way, you can stop and resume the recover.  I regularly have to physically unplug and plug back in the USB connector to a drive that has dropped out like that.  I find some drives tend to faint easily at high speeds, which makes the slower USB speeds an asset in these cases.  Using -D or -d switches can also slow things down and regulate speed in a useful way.
>>
>> Good luck.
>>
>> Andrew Zajac
>> AndrewZajac.ca
>> Ubuntu-Rescue-Remix.org
>>
>>
>>
>>
>> --- On Tue, 12/6/11, Tom Williams<address@hidden>  wrote:
>>
>>> From: Tom Williams<address@hidden>
>>> Subject: [Bug-ddrescue] Question about reported errorsize
>>> To: address@hidden
>>> Received: Tuesday, December 6, 2011, 12:26 PM
>>> I'm running ddrescue from a Trinity
>>> Rescue CD (3.4-bld-372) on a 500GB
>>> hard drive with two NTFS partitions on it.  The hard
>>> drive is dying and
>>> I'm hoping to move the data to a new hard drive so it can
>>> be saved.
>>>
>>> I've issued his command:
>>>
>>> ddrescue /dev/sda /dev/sdb
>>>
>>> and the process is underway now.
>>>
>>> ddrescue has reported only ONE error and the errorsize is
>>> 500GB.  It has
>>> 51MB of data and is in the "splitting" phase now. It looks
>>> as if it's
>>> going to try to split the 500GB of "errors" reported in
>>> "errorsize".
>>>
>>> My question:  does the fact that 500GB was reported as
>>> the errorsize
>>> mean this process is most likely going to fail?  Is it
>>> worth spending
>>> the time to go through the "splitting" phase?
>>>
>>> I believe I'm running ddrescue 1.14, but I'm not sure.
>>>
>>> Thanks!
>>>
>>> Peace...
>>>
>>> Tom
>>>
>>> _______________________________________________
>>> Bug-ddrescue mailing list
>>> address@hidden
>>> https://lists.gnu.org/mailman/listinfo/bug-ddrescue
>>>
>
>
>
> ------------------------------
>
> Message: 5
> Date: Wed, 07 Dec 2011 14:10:32 +0000
> From: Bob<address@hidden>
> To: address@hidden
> Subject: [Bug-ddrescue] Request for example of rescuing a dvd data
>     disk
> Message-ID:<address@hidden>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> Hi
>
> I am trying to recover a dvd data disk. It reports the following
> errors:- wrong fs type; bad option; bad super block; missing code page.
>
> I have read the manual but can't find an example recovering a 4.7GB dvd.
> I have only one drive.
>
> Could you please furnish an example to run gddrescue on a data dvd,
> saving to a file on the hard drive and then writing to a fresh dvd with
> only one drive available.
>
> Ubuntu 11.10 AMD64
>
> Thank you
> Bob
>
>
>
>
> ------------------------------
>
> Message: 6
> Date: Wed, 07 Dec 2011 16:47:44 +0100
> From: Antonio Diaz Diaz<address@hidden>
> To: Tom Williams<address@hidden>
> Cc: address@hidden
> Subject: Re: [Bug-ddrescue] Version 1.15-rc1 of GNU ddrescue released
> Message-ID:<address@hidden>
> Content-Type: text/plain; charset=us-ascii; format=flowed
>
> Hello Tom,
>
> Tom Williams wrote:
>> I have worked up a small patch to add the ddrescue version number to the
>> "show help" (-h) output.  This way, when someone runs "ddrescue -h", the
>> version of ddrescue being run is included in the output.  :)
> Any reason you can't run "ddrescue -V"?
>
> The change you propose would include ddrescue's version number in the
> description of the help2man-generated man page, and I am not sure I like
> that.
>
>
> Best regards,
> Antonio.
>
>
>
> ------------------------------
>
> _______________________________________________
> Bug-ddrescue mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/bug-ddrescue
>
>
> End of Bug-ddrescue Digest, Vol 71, Issue 1
> *******************************************
>


_______________________________________________
Bug-ddrescue mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/bug-ddrescue

reply via email to

[Prev in Thread] Current Thread [Next in Thread]