Equivalent of DD command conv=noerror, sync
Hi. My Mac OS Drive has started to seriously degrade. Not sure if it's a hardware issue but I'm taking no risks and trying to copy data off to an external drive. I've used CC Cloner but this throws up errors - mainly on copying video files. So I started to use the DD command to copy the erroneous files. However I received loads of input/output errors so I started to use the dd command with conv=noerror, sync parameter which forces the file to be copied whilst forcing through the input/output errors that I was seeing. This works well but there are hundreds of files so will take me months.
So - I found your amazing GUI tool. As a novice, please can you tell me:
1. Can I use it to copy directories of files rather than the whole drive
2. Will it force through the copy of all files that have these annoying input/output errors?
Really hope you can help as I think you may be a life saver!
Thanks
Jon
Question information
- Language:
- English Edit question
- Status:
- Answered
- Assignee:
- Hamish McIntyre-Bhatty Edit question
- Last query:
- Last reply:
Revision history for this message
|
#1 |
Hi,
1. It can't copy only directories, sorry, but you can tell it to copy individual files, though that could be very slow and tedious. Instead, you might want to copy to a partition on a new drive.
2. Yes, it will, and because it uses ddrescue rather than dd, it will probably do it faster as well.
Hope this helps,
Hamish
Revision history for this message
|
#2 |
Thanks Hamish.
Understood.
Do you mind if I ask you a couple of quick extra questions:
1. If it can't copy directories, can it copy packages? eg iPhoto, FinalCutPro package
If not, I presume I will need to copy the full 3TB (single partition) that I have on my Mac? Or is there a way of limiting what I copy across?
2. Does it copy a full replica of the contents or does it create an image file? If an image file - how do I get the contents from it?
3. I can't seem to specify any parameters like the conv=noerror for dd. Do I need to specify parameters or will your tool simply ignore the errors and copy the data over anyway?
Many thanks again and I hope you can help.
Revision history for this message
|
#3 |
Hi,
Ask me 50 questions if you want XD
1. OS X packages? Those are actually directories, but OS X just makes it look like they aren't, so you'll have to copy the full partition, sorry about that, it's just how ddrescue (the command line program that this is a GUI for) was designed.
2. It creates full replicas (assuming you get all the data off). These can be images, which you can mount using the GUI at the end of the recovery process, or you can copy to another partition on a different drive. If you copy to a different partition, make sure it is equal size, or preferably, a little bit bigger than the source.
3. No parameters needed, all the one's you're likely to need are in the "Settings" window :)
Hamish
Revision history for this message
|
#4 |
One last question - sorry!
When I select the input, there seems to be a number of disks detected, but I know I only have one disk and one partition. How do I know which is the right one to choose?
eg on my MacAir I have a 500GB hard drive and it shows up three times:
/dev/disk1 is Type-Device with 499GB
/dev/disk0s2 is Type-Partition with 499GB
The list also shows /dev/disk0 as type-Device and size of 500GB
How do i know which to choose to copy the data from as they are all the same one?
Thanks Hamish.
Revision history for this message
|
#5 |
Hi,
That's okay, it's why I'm the support contact XD
Okay, each name that ends with e.g. "/dev/disk0s1", "/dev/disk2s2", "/dev/disk1s3" is a partition, and all other names e.g. "/dev/disk0", "/dev/disk3" are devices.
So, it seems you have more than one hard drive connected. /dev/disk0 is probably your boot drive, but to double check, could you send me a screenshot of the Disk Info Window please? Normally your boot drive will have an HFS+ filesystem. Do you store your data on mac OS's boot drive?
Hamish
Revision history for this message
|
#6 |
Sure - how do I upload a screenshot to you? I don't see an image upload function on launchpad.
Revision history for this message
|
#7 |
I forgot you can only do that for bugs XD
How about you email me at hamishmb at live.co.uk?
Hamish
Revision history for this message
|
#8 |
I have emailed you the screenshot and an explanation.
Many thanks
Revision history for this message
|
#9 |
Hi Hamish. Did you receive the screenshot via email? Thanks. J.
Revision history for this message
|
#10 |
Hi,
Yep, I did, I'm just slow, busy, and prone to getting sidetracked XD
Thanks for the screenshot, it clears it up for me a lot :)
Okay,
Never heard of a fusion drive... sounds complicated.
1) Whichever disk has your data is the right one, the size looks smaller because my GUI rounds the sizes to make them look nicer XD
2) disk1s2 looks like a good candidate - it is bigger than the source. You need to make sure you don't have any data that you want to keep on this partition before you copy to it, because anything on it will be overwritten. If you're sure of all that, this partition looks like a good candidate.
3) If the OS is on the same partition (disk1s2) then you need to copy somewhere else. However, if your OS is on disk1, but a different partition (e.g. disk1s3) then you're okay. If your OS is there, DDRescue-GUI will try to overwrite it, so you either need to put the OS somewhere else, or copy to a different drive.
4) Store the log file somewhere that is neither on the source or destination drive. Something like a USB stick, or the desktop on your recovery OS would be a good idea :)
Side note: Mavericks doesn't get security updates any more, so I feel obliged to say that if you still use it for your day to day OS, you should upgrade. Should be fine for data recovery though, the GUI runs on Snow Leopard XD
Hope this helps,
Hamish
Revision history for this message
|
#11 |
Thanks Hamish - I know the busy feeling. ;-D
And finally...
'If the OS is on the same partition (disk1s2) then you need to copy somewhere else.'
How do I know which partition the OS is on? Is there an easy way to find out?
Revision history for this message
|
#12 |
Hi,
It seems that if you go to disk utility, you can see the mount point of the selected drive near the bottom of the window. All you should need to do is find the drive whose mount point is '/' and that will be your boot drive. You can find the name, or identifier for that disk by clicking on info at the top of the window.
Hamish
Revision history for this message
|
#13 |
OK. So far so good. Thank you Hamish. Understood everything and I'm ready to Start.
Now, my newbie questions on the settings:
No. of times to retry bad sectors: Now set to Forever (as I know there are bad sectors on there)
Max numbers of errors before exiting: Infinite
Number of clusters to copy at a time: Default (128)
1. So, should I choose the above settings (to force the bad sectors) even though it may take a while
2. Also - Should I choose best recovery or fastest recovery?
3. Should I use any of the other settings? eg read the input file/disk backwards, preallocate space on disc for output etc - are any of these worthwhile?
Hope you can help?
Thanks for your patience!
Revision history for this message
|
#14 |
Hi,
1) You might want to retry 3 times. This is because retrying (ddrescue does this at the end) can be very, very slow, and by the time a read attempt has failed 3 times it's probably not going to succeed. Your choice XD
2) Probably fastest given you know you have bad data corruption or bad sectors.
3) These are worth trying if you're not getting anywhere fast. The beauty of ddrescue is that if you use a log file, you can stop and start multiple times with different settings, and resume where you left off :)
I blame me for not writing documentation ;)
Hamish
Revision history for this message
|
#15 |
Aaaah - help!
I pressed a button on my keyboard by mistake and the DDrescue aborted. It's now asking if I want to reset, mount Image/Disk or Quit - but actually I want to continue where I left off. How do I continue without starting again??
I have a the log file which I created when we started. How can I continue pls?
Thanks
Revision history for this message
|
#16 |
Hi Hamish - not sure if this came through as a notification to you? Sorry if I'm chasing you! Just got about 15 days worth of backup to go and I'm keen to progress before my hard drive fails completely.
As I accidentally aborted - I want to continue from where I left off. Three questions:
1. which option should I choose?
Reset, Mount Image or Quit?
2. And then how do I continue to backup from where I left off (I’ve already copied over 800GB so don’t want to lose that!)?
I presume from previous posts that I select the same input, output and log file and try to remember the settings that I used I presume?
3. Also - should I choose to over-write the output file? I don’t want to lose what I’ve done so far as I've already successfully copied over 800GB.
Many thanks
J.
Revision history for this message
|
#17 |
Hi,
Nope, the other never came through as a notification XD
1) You can reset, or quit, and then resume your backup. Mounting (to access your recovered data) probably won't work until your backup is complete.
2) Pretty much. If you use slightly different settings it won't matter though, as long as the input, output, and log settings are *identical*.
3) You only need to do that if you're recovering to another device. If recovering to a file, it shouldn't be needed. The GUI will autodetermine this for you, so if you leave the setting alone, it will be fine :)
Well done, sounds like it's working!
Hamish
Revision history for this message
|
#18 |
Thanks
Was going very smoothly until I accidentally hit a key on my keyboard!
I tried to select the previous log file, but it then the only button option is to Save it. Then it says the Log file already exists and asks if I want to replace it.
Shall I choose replace? (I've taken a copy just in case).
Also - not sure what you mean about Overwrite output file. I previously had checked this option as I am recovering my entire hard drive to an external hard drive.
If I keep this checked will it overwrite my previous half-completed backup? I want it to carry on where I left off.
Thanks.
Revision history for this message
|
#19 |
Hi,
Yep - Saving and Replacing the log file is fine, that's the nature of the OS.
If you had it checked before, leave it checked, it's fine. If the settings in SettingsWindow are different it doesn't matter, as long as the input, output and log files are the same.
Hamish
On 17 Feb 2017, 11:37 +0000, jon Summ <email address hidden>, wrote:
Question #452008 on DDRescue-GUI changed:
https:/
Status: Answered => Open
jon Summ is still having a problem:
Thanks
Was going very smoothly until I accidentally hit a key on my keyboard!
I tried to select the previous log file, but it then the only button option is to Save it. Then it says the Log file already exists and asks if I want to replace it.
Shall I choose replace? (I've taken a copy just in case).
Also - not sure what you mean about Overwrite output file. I previously had checked this option as I am recovering my entire hard drive to an external hard drive.
If I keep this checked will it overwrite my previous half-completed backup? I want it to carry on where I left off.
Thanks.
--
You received this question notification because you are subscribed to
the question.
Revision history for this message
|
#20 |
Thanks - that worked - Phew!!
Resumed from 28%
Once it's finished, I presume I click on Mount Image? What will that do? Will it mount on the new output disk so I can view my files through finder?
Still got 15-20 days left on the recovery....
Ho hum...
Revision history for this message
|
#21 |
Yep, that's what it will do, although you can mount the image yourself
as well by using Disk Utility if you want to :)
Hamish
On 17/02/17 12:18, jon Summ wrote:
> Question #452008 on DDRescue-GUI changed:
> https:/
>
> Status: Answered => Open
>
> jon Summ is still having a problem:
> Thanks - that worked - Phew!!
>
> Resumed from 28%
>
> Once it's finished, I presume I click on Mount Image? What will that
> do? Will it mount on the new output disk so I can view my files through
> finder?
>
> Still got 15-20 days left on the recovery....
>
> Ho hum...
>
Revision history for this message
|
#22 |
Hi Hamish
The DD rescue seems to have halted - not sure why. It doesn't seem to have frozen but there hasn't been any data copied for the last few hours.
Should I abort and restart from where it left off using the log file?
I've sent you a screenshot to your email but it doesn't seem to show anything anyway but see if you can find any pointers.
What do you recommend?
Hope you can help
Many thanks
J.
Revision history for this message
|
#23 |
Ok. Quick update. I aborted and as i did so, the DDrescue gui window (the left bottom white panel) showed that I had actually copied more than it was showing when it looked like it has stalled (it jumped from 930 GB to 1030GB). So it seemed like it was still copying but the GUI hadn't updated. Anyhow, I've now continued from where I left off and it seems ok.
Perhaps you need a refresh button on the GUI?
Or perhaps it had actually stalled? Either way - all seems ok now.
Revision history for this message
|
#24 |
Hi,
Sorry for slow reply. Hmm, sounds like a bug in the GUI -- I'll look into that. The screenshot seems to be saying the recovery was still going, but as long as it's working I guess it doesn't matter :)
Revision history for this message
|
#25 |
Wow - really going slow now.
Any tips to improve the speed?
I've selected the following settings:
Use Direct Disk Access - checked
Overwrite output file/disk - checked
Other options unchecked
No. of times to retry bad sectors: 5
Max no. of errors before exiting: Default (infinite)
No. clusters to copy at one time: 256
remaining time: 531 days!!! Not sure I can wait that long!
Any ideas how to improve speed?
Thanks
J.
Revision history for this message
|
#26 |
Hi,
The time remaining estimates are usually very pessimistic and kinda inaccurate, particularly the GUI one XD If there's an estimate available in the terminal output, it's probably better.
Try setting the number of clusters to copy at a time to the lowest. That might make it skip through bad areas faster, but will make the good areas slower. You also also try the reverse option, which reads from the end of your source disk first, where they may be lots of good areas on the drive. Also, try to make sure your drive isn't overheating, as this can sometimes happen with data recovery and make it slower.
There aren't really any hard and fast rules here, maybe just experiment a bit and try to find something that works a little better. If you're still on the "Copying non-tried blocks" phase, it should skip over the bad areas fairly quickly.
Hope this helps,
Hamish
Revision history for this message
|
#27 |
Quickly meaning maybe in a few hours XD
Revision history for this message
|
#28 |
^ For example there are some non-trimmed blocks in your screenshot -- these are areas that ddrescue knows have bad sectors inside, but that it hasn't yet tried to read and has instead skipped over them to try and get the easily readable data first.
Revision history for this message
|
#29 |
Thanks again Hamish - really appreciate your help.
Yep - I am still on the "Copying non-tried blocks" phase but a fair bit to go... (2TB still!)
Silly question now: When you say 'Try setting the number of clusters to copy at a time to the lowest.' - do mean greater than 256 or as small a number as possible?
Just checking...
:-0
Also - can I continue with the same process using the same log file by aborting and trying to read in reverse?
Will this still keep the previous part of the backup?
Thanks again
J.
Revision history for this message
|
#30 |
Hi,
Yep, smallest number :P
Yep that'll be fine. As long as the three main settings (input, output, log) are the same, you can change the other ones as much as you like :)
Hamish
On 21 Feb 2017, 00:42 +0000, jon Summ <email address hidden>, wrote:
Question #452008 on DDRescue-GUI changed:
https:/
Status: Answered => Open
jon Summ is still having a problem:
Thanks again Hamish - really appreciate your help.
Yep - I am still on the "Copying non-tried blocks" phase but a fair bit
to go... (2TB still!)
Silly question now: When you say 'Try setting the number of clusters to
copy at a time to the lowest.' - do mean greater than 256 or as small a
number as possible?
Just checking...
:-0
Also - can I continue with the same process using the same log file by aborting and trying to read in reverse?
Will this still keep the previous part of the backup?
Thanks again
J.
--
You received this question notification because you are subscribed to
the question.
Revision history for this message
|
#31 |
Hi Hamish
Copying is now going really dog-slow.
What does it I do a soft-run do? If I run that, what happens to anything that gets left behind?
Can I go back after and try to rescue the bad sectors after or are they gone for good?
Also - could it be that the hard drive has overheated and has gone slow? Should I leave it for a few hours before trying again? It seems that any settings I try don't fix the copying speed back to the levels I was acheiving a few days ago.
Getting a bit worried now.
Any thoughts?
J.
Revision history for this message
|
#32 |
And could a pre-allocation of space help in the settings? (Not tried that yet!)
Revision history for this message
|
#33 |
Hi,
A soft-run grabs the good parts of the disk only. After that, you can restart the GUI with other settings to try and get your bad sectors.
Yeah, could have overheated, or could just be struggling with very damaged parts of the disk. Do you have a fan you could point at it? Might be safer than turning it off.
Preallocating space shouldn't make any difference unless your destination drive is slow.
Hamish
Revision history for this message
|
#35 |
Hi Hamish
Even the soft run is going very slow - average 6306 b/s. This is using blocks of 128kb
Any thoughts on speeding up?
I've tried backwards too.
This is killing me!
Revision history for this message
|
#36 |
Hi,
Sounds like a total pain. I think it must be overheating. If you have it installed internally, maybe you should let it cool for a few hours, if external, try to get a cooling fan or just turn it off for a while. Does it feel hot?
Also, the the number of errors count or Unreadable data size going up, because if so it might just be a tricky part of the drive.
Hamish
Revision history for this message
|
#37 |
Errors Slightly going up but not much.
I've actually put the drive to sleep overnight so it shouldn't be overheating.
Sometimes even on soft run the current rate of transfer drops to zero which I didn't think should happen
Note - this slowness only occurred once I accidentally aborted the first time.
During the first run, the transfer rate was a lot better.
I'm loathe to start again but do you think it might be an option (annoyingly as I've already transferred 1129GB (38%) which took 10 days.
Revision history for this message
|
#38 |
Actually - speeding up now in bursts, slowing down occasionally but slightly better now - fingers crossed!
Revision history for this message
|
#39 |
Hi,
Nope, restarting won't make it faster. If errors are going up then this is to be expected. ddrescue will skip over the bad parts of the drive increasingly quickly to try and find the good parts. If you look under "Input Position" or "ipos" you should be able to see where on the disk it's trying to read the data from at the moment.
That's actually really good progress, so to be honest I'd just let it do its thing.
Hamish
Revision history for this message
|
#40 |
^ To be expected because this must be a badly damaged area of the disk that is very difficult to read -- very slow to read.
Revision history for this message
|
#41 |
OK - I will let it run for now -
It's been stuck on IPOS 1130 GB for ages now. (non tried is 1869, rescued is 1129GB).
Is this the longest running support thread you've had to deal with?? Really sorry!
Revision history for this message
|
#42 |
Hi,
Glad it's speeding up :) Ah, it not giving you the position in MB makes that metric fairly unhelpful :P
Umm, maybe, but I don't mind, it fine :)
Hamish
Can you help with this problem?
Provide an answer of your own, or ask jon Summ for more information if necessary.