'dd' help needed

clifford227 · 02-08-2011, 07:38 AM

hi,

would the following command wipe an entire drive clean (partition table included)?

Code:

dd if=/dev/urandom of=/dev/sda bs=512

I was wondering what the 'bs' part means? Would it be better to omit that?

fordeck · 02-08-2011, 07:46 AM

Yes it would over write /dev/sda with random data. The 'bs=512' just tells the command to write 512 bytes at a time and is not needed.

Quote:

Originally Posted by clifford227

hi,

would the following command wipe an entire drive clean (partition table included)?

Code:

dd if=/dev/urandom of=/dev/sda bs=512

I was wondering what the 'bs' part means? Would it be better to omit that?

You can get more information by reading the man page for 'dd'

Regards,

Fordeck

lugoteehalt · 02-08-2011, 07:54 AM

Quote:

Originally Posted by clifford227

I was wondering what the 'bs' part means? Would it be better to omit that?

bs=512 is the default:

Code:

info coreutils 'dd invocation'

So if omitted it would not make any difference. It simply instructs dd how big a chunk of data to remove or place each go. Incidentally the size of the MBR is 512.

syg00 · 02-08-2011, 08:01 AM

And will be unbelievably slow - use a (much) larger bs. The manpage gives a way of finding out the progress.

vinodvb · 02-08-2011, 08:02 AM

Quote:

Originally Posted by clifford227

hi,

would the following command wipe an entire drive clean (partition table included)?

Code:

dd if=/dev/urandom of=/dev/sda bs=512

I was wondering what the 'bs' part means? Would it be better to omit that?

The dd command syntax will be:

Code:

dd if=inputfile of=outputfile bs=byte-size count=blocks

The 'bs' and 'count' options together will give us the output file size of the dd command.

And regarding the 'bs' part, as mentioned earlier it is used along with 'count' to specify the output file size.

Code:

Example : 
bs=512 count=1 --> This gives us a 512 byte size file.
bs=1M count=1 --> Results in a 1MB file (1,048,576 bytes)
bs=1MB count=1 --> Resutls in a 1MB file (1,000,000 bytes)
bs=1G count=1 --> Results in a 1GB file (1,073,741,824 bytes)
bs=1GB count=1 --> results in a 1GB file (1,000,000,000 bytes)

In short, 'bs' will be multiplied 'count' number of times to decide the output file size.

NOTE: dd command to backup the MBR

Code:

dd if=/dev/sda of=/tmp/mbr.out bs=512 count=1

Regards

clifford227 · 02-08-2011, 08:18 AM

Quote:

Originally Posted by vinodvb

The dd command syntax will be:

Code:

dd if=inputfile of=outputfile bs=byte-size count=blocks

Quote:

So the above command should include the 'count' option, and when included will actually wipe out the MBR (master boot record) and not the entire drive.

Sorry just make this clear for me, you mean that by including the count option, the above command would wipe only the MBR, but not the rest of the device?

vinodvb · 02-08-2011, 08:25 AM

Quote:

Originally Posted by clifford227

Code:

dd if=inputfile of=outputfile bs=byte-size count=blocks

Sorry just make this clear for me, you mean that by including the count option, the above command would wipe the MBR, but not the rest of the device?

Sorry for the confusion. If the bs and count options are mentioned, then the output size will be limited to as specified by bs and count. If not the entire drive will be written with random data.

In short, the first 512 bytes contain the MBR and if I limit the output file size to 512 bytes, then only the MBR will be overwritten. If I don't limit the output file size, then the entire drive will be overwritten including the MBR.

i92guboj · 02-08-2011, 08:30 AM

Quote:

Originally Posted by clifford227

Code:

dd if=inputfile of=outputfile bs=byte-size count=blocks

Sorry just make this clear for me, you mean that by including the count option, the above command would wipe only the MBR, but not the rest of the device?

More correctly speaking, bs=512 count=1 will overwrite the first 512 bytes of the file specified by of='...'. If that file is a device node tied to a block device then it will write the first 512 bytes in that device with random bytes. If the device is partitioned using a MBR scheme then that matches the MBR. But the device could also hold meaningless stuff or be partitioned using EFI or some other thing, not necessarily an MBR scheme.

clifford227 · 02-08-2011, 08:38 AM

So the count option limits the portion of the device that will be overwritten?

If I only include the 'bs' option, then the entire drive will be overwritten? (please excuse me, I've got a very tiny brain).

vinodvb · 02-08-2011, 08:41 AM

Quote:

Originally Posted by clifford227

So the count option limits the portion of the device that will be overwritten?

If I only include the 'bs' option, then the entire drive will be overwritten? (please excuse me, I've got a very tiny brain).

Yup, you got it right

clifford227 · 02-08-2011, 08:48 AM

Fantastic, thanks for your help and patience guys

Stephen Morgan · 02-08-2011, 12:06 PM

While we're on the subject, wouldn't /dev/zero work just as well as /dev/urandom, and probably be quicker?

i92guboj · 02-08-2011, 04:42 PM

Quote:

Originally Posted by Stephen Morgan

While we're on the subject, wouldn't /dev/zero work just as well as /dev/urandom, and probably be quicker?

Yes. It's just that people who are afraid that the NSA will waste their time to rescue porn from their hard drives prefer to use /dev/urandom to make it harder to recover all the porn from the drive. If you don't feel that kind of paranoia, just use /dev/zero.

onebuck · 02-08-2011, 09:24 PM

Hi,

Look at 'Learn The DD Command Revised' if you want to learn by examples.