Learn the DD command

djbelieny · 08-24-2010, 08:43 AM

Great post, the most comprehensive dd tutorial ever.
Thank you

exvor · 08-24-2010, 10:54 PM

I had a quick question about this particular modifier, conv=notrunc. What exactly does this do as the man page is a bit sketchy about it? In the past I have created disk images of dvd's and even hard drives and have never had issues with them.

Anyone shed some light on why it would be needed and possible ramifications if it is not used?

vsurlan · 08-25-2010, 08:03 AM

Is it just me or does it really look like a link to the manual page for dd would fit nicely in the original post as well?

Something like: Manual page for dd would be perfect.

mariapeter12 · 08-27-2010, 12:52 AM

Good tutorial about DD command in linux but we should not forget that the incorrect use of the DD command may Cause Data loss in Linux So proper care must be taken while using this Powerful Command.

tiekookeit · 09-01-2010, 01:52 PM

just AWESOME

AwesomeMachine · 09-05-2010, 11:56 PM

Quote:

Originally Posted by archtoad6

I seem to only be posting corrections, sorry. ...

From the bash man page (search for "metacharacter"):

This means that the '>' separates words on the CLI the same as a space does, & therefore putting spaces around a '>' is NOT necessary for the command to work.

Furthermore, that is also my personal experience.

Anyone who wants to can test this by comparing the results of the following 2 commands:

Code:

>space_test0
> space_test1

Each will create an empty file in the current directory.

I DO generally put a space before '>' for clarity, especially if it's in a code snippet that I'm going to post. (See my post #361 above.)

OK, archtoad,

Do it the way you want. But if you don't even try what I said then don't post that you don't know how to do it. The > is a redirect operator. Why don't you just try what I said?

AwesomeMachine · 09-06-2010, 12:23 AM

Quote:

Originally Posted by exvor

I had a quick question about this particular modifier, conv=notrunc. What exactly does this do as the man page is a bit sketchy about it? In the past I have created disk images of dvd's and even hard drives and have never had issues with them.

Anyone shed some light on why it would be needed and possible ramifications if it is not used?

exvor,

This is a popular subject. The conversion parameter, notrunc, tells dd not to truncate the output file. This applies if the output file currently exists.

If you want to write a sector of 0xffh bytes to the beginning of an existing file:

Code:

dd if=/dev/zero bs=512 count=1 | sed 's,\x00,\xff,g' | of=/home/sam/existing.file

will replace the entire file with 512 0xffh bytes, leaving the file 512 bytes long and eliminating the preexisting contents.

Code:

dd if=/dev/zero bs=512 count=1 | sed 's,\x00,\xff,g' | of=/home/sam/existing.file conv=notrunc

tells dd to just replace the specified bytes and leave the rest of the output file alone. Without this option, dd would create an output file 512 bytes long.

exvor · 09-07-2010, 07:53 PM

Ahh ok that makes sense now.

AwesomeMachine · 09-26-2010, 01:29 AM

Quote:

Originally Posted by vsurlan

Is it just me or does it really look like a link to the manual page for dd would fit nicely in the original post as well?

Something like: Manual page for dd would be perfect.

Hi vsurlan,

I actually put a link to the htmlized man page, but when I saved the whole post it was too large and the system refused to save, so I had to cut out some new content and that got cut. If you look at the second post the man page is written out pretty much verbatim.

JZL240I-U · 10-15-2010, 01:15 AM

Hi AwesomeMachine and other knowledgeable persons

.

How is /dev/zero working? I mean what mechanism tells dd to use an endless stream of binary zeros e.g. for an input file in dd if=/dev/zero ... ? I can't imagine that the zeros truly get freshly created for more than one blocksize (bs) -- or do they?

Reason for my inquiry (apart from the joy in generalized knowledge) is this: last night I wiped an old 10 GB disk with "dd if=/dev/urandom of=/dev/sdb". That took more than 1 1/2 hours on my old 700 Mhz Athlon.

After this experience my idea was to first create a file of, say, 1 MB of random data and then using this block of data repeatedly as input for dd. Would that be possible and how would one go about that?

rn_ · 10-15-2010, 08:35 AM

Quote:

Originally Posted by JZL240I-U

How is /dev/zero working?

My understanding is that the device driver behind the device responds to the read request by writing zeros.

Quote:

Originally Posted by JZL240I-U

Would that be possible and how would one go about that?

you could use something like:

Code:

while true
do
  cat filewithrandomdata
done | dd of=/dev/sdb

HTH.
-RN.

JZL240I-U · 10-18-2010, 01:44 AM

Quote:

Originally Posted by rn_

My understanding is that the device driver behind the device responds to the read request by writing zeros....

How wasteful

. I mean, creating once the necessary block of zeros and then never increasing the address pointer would do admirably without pestering the CPU for zeros. Ah well...

Quote:

Originally Posted by rn_

...you could use something like:...

Thanks, I copied that and will try it next time.

<edit> Okay, I tried it. It still took close to an hour. While it is an improvement it still is nothing like fast. CPU-load stayed at 100%. </edit>

divyashree · 10-18-2010, 02:11 AM

Nice article,very much helpful

rn_ · 10-18-2010, 08:13 AM

Quote:

Originally Posted by JZL240I-U

How wasteful

. I mean, creating once the necessary block of zeros and then never increasing the address pointer would do admirably without pestering the CPU for zeros. Ah well...

perhaps that's exactly what it does; i wouldn't know because i haven't seen the code, and if i did, i doubt if i'd be able to make heads or tails of it.

onebuck · 10-18-2010, 08:50 AM

Hi,

You can always 'man command';

Quote:

excerpt 'man null';
NAME

null, zero - data sink DESCRIPTION

Data written on a null or zero special file is discarded. Reads from the null special file always return end of file, whereas reads from zero always return \0 characters.
null and zero are typically created by:

mknod -m 666 /dev/null c 1 3
mknod -m 666 /dev/zero c 1 5
chown root:mem /dev/null /dev/zero

NOTES

If these devices are not writable and readable for all users, many programs will act strange. FILES

/dev/null
/dev/zero SEE ALSO

chown(1), mknod(1)