dd progress info for long operations

On Wed, 30 Sep 2009 02:21:21 +0200, Simon Sibbez <simon.sibbez.TakeThisOut@buerotiger.de> wrote:

>Grant wrote:
>
>> Preparing an old hdd for reuse, first I want to wipe it to zeroes
>> with dd. I'm curious about progress too, so here's how to see the
>> dd status each ten seconds.
>
>Don't do that in a busybox environment. The dd implementation just
>quits when it receives the signal.

Learn something every day

Grant.
--
http://bugsplatter.id.au

On 2009-09-29 at 20:30 ADT, Grant <g_r_a_n_t_ RemoveThis @bugsplatter.id.au> wrote:
> Hi there,
>
> Preparing an old hdd for reuse, first I want to wipe it to zeroes
> with dd. I'm curious about progress too, so here's how to see the
> dd status each ten seconds.
>
> clear hdd:
> ~# dd if=/dev/zero bs=4k of=/dev/hdc
>
> discover pid:
> ~$ ps aux|grep dd\ |grep -v grep
> root 1685 30.7 0.2 1892 708 pts/3 R+ 09:06 5:48 dd if /dev/zero bs 4k of /dev/hdc
> ^^^^
>
> In another root terminal, drive the dd status request signal:
> ~# while :; do kill -usr1 1685; sleep 10; done

killall is your friend...

$ killall -usr1 dd

(assuming you only have the one dd running, of course)


Jim

On Tue, 29 Sep 2009 23:06:34 -0300, Jim Diamond <Jim.Diamond DeleteThis @nospam.AcadiaU.ca> wrote:

>On 2009-09-29 at 20:30 ADT, Grant <g_r_a_n_t_ DeleteThis @bugsplatter.id.au> wrote:
>> Hi there,
>>
>> Preparing an old hdd for reuse, first I want to wipe it to zeroes
>> with dd. I'm curious about progress too, so here's how to see the
>> dd status each ten seconds.
>>
>> clear hdd:
>> ~# dd if=/dev/zero bs=4k of=/dev/hdc
>>
>> discover pid:
>> ~$ ps aux|grep dd\ |grep -v grep
>> root 1685 30.7 0.2 1892 708 pts/3 R+ 09:06 5:48 dd if /dev/zero bs 4k of /dev/hdc
>> ^^^^
>>
>> In another root terminal, drive the dd status request signal:
>> ~# while :; do kill -usr1 1685; sleep 10; done
>
>killall is your friend...
>
>$ killall -usr1 dd
>
>(assuming you only have the one dd running, of course)

Well, if you had several running, each in its own terminal?

Hmm, almost worked...

root@deltree:~# while :; do killall -usr1 dd; sleep 10; done
Cannot find user sr1
^C
root@deltree:~# while :; do killall -USR1 dd; sleep 10; done

And yes, two running instances of dd updated together

Grant.
--
http://bugsplatter.id.au

Grant wrote:
> Hi there,
>
> Preparing an old hdd for reuse, first I want to wipe it to zeroes
> with dd. I'm curious about progress too, so here's how to see the
> dd status each ten seconds.
>
> clear hdd:
> ~# dd if=/dev/zero bs=4k of=/dev/hdc
>
> discover pid:
> ~$ ps aux|grep dd\ |grep -v grep
> root 1685 30.7 0.2 1892 708 pts/3 R+ 09:06 5:48 dd if /dev/zero bs 4k of /dev/hdc
> ^^^^
>
> In another root terminal, drive the dd status request signal:
> ~# while :; do kill -usr1 1685; sleep 10; done
>
> Each dd progress report looks like:
> 8721060+0 records in
> 8721060+0 records out
> 35721461760 bytes (36 GB) copied, 1366.35 seconds, 26.1 MB/s
>
> Grant.

Another method is to use the ddrescue command like the following.


ddrescue -v -b4k /dev/zero /dev/hdc


Harold

On 2009-09-29 at 23:23 ADT, Grant <g_r_a_n_t_.DeleteThis@bugsplatter.id.au> wrote:
> On Tue, 29 Sep 2009 23:06:34 -0300, Jim Diamond <Jim.Diamond.DeleteThis@nospam.AcadiaU.ca> wrote:
>
>>On 2009-09-29 at 20:30 ADT, Grant <g_r_a_n_t_.DeleteThis@bugsplatter.id.au> wrote:
>>> Hi there,
>>>
>>> Preparing an old hdd for reuse, first I want to wipe it to zeroes
>>> with dd. I'm curious about progress too, so here's how to see the
>>> dd status each ten seconds.
>>>
>>> clear hdd:
>>> ~# dd if=/dev/zero bs=4k of=/dev/hdc
>>>
>>> discover pid:
>>> ~$ ps aux|grep dd\ |grep -v grep
>>> root 1685 30.7 0.2 1892 708 pts/3 R+ 09:06 5:48 dd if /dev/zero bs 4k of /dev/hdc
>>> ^^^^
>>>
>>> In another root terminal, drive the dd status request signal:
>>> ~# while :; do kill -usr1 1685; sleep 10; done
>>
>>killall is your friend...
>>
>>$ killall -usr1 dd
>>
>>(assuming you only have the one dd running, of course)
>
> Well, if you had several running, each in its own terminal?
>
> Hmm, almost worked...
>
> root@deltree:~# while :; do killall -usr1 dd; sleep 10; done
> Cannot find user sr1
> ^C
> root@deltree:~# while :; do killall -USR1 dd; sleep 10; done
mea culpa!

> And yes, two running instances of dd updated together
What the heck, why not go for 3 or 4 ?

I'm sort of curious about the filling it with zeroes... was there top
secret stuff on it?

Jim

Harold <harold.RemoveThis@everywhere.com> wrote:
> Another method is to use the ddrescue command like the following.
> ddrescue -v -b4k /dev/zero /dev/hdc

Yet another method is to use some buffer program which displays progress
to stderr. There are different programs to choose from, one of my
favorites is bfr, http://www.glines.org/software/bfr

dd if=/dev/zero | bfr -p | dd of=/dev/hdc

regards Henrik
--
The address in the header is only to prevent spam. My real address is:
hc3(at)poolhem.se Examples of addresses which go to spammers:
root@localhost postmaster@localhost

On Wed, 30 Sep 2009 10:04:09 -0300, Jim Diamond <Jim.Diamond DeleteThis @nospam.AcadiaU.ca> wrote:

>On 2009-09-29 at 23:23 ADT, Grant <g_r_a_n_t_ DeleteThis @bugsplatter.id.au> wrote:
>> On Tue, 29 Sep 2009 23:06:34 -0300, Jim Diamond <Jim.Diamond DeleteThis @nospam.AcadiaU.ca> wrote:
>>
>>>On 2009-09-29 at 20:30 ADT, Grant <g_r_a_n_t_ DeleteThis @bugsplatter.id.au> wrote:
>>>> Hi there,
>>>>
>>>> Preparing an old hdd for reuse, first I want to wipe it to zeroes
>>>> with dd. I'm curious about progress too, so here's how to see the
>>>> dd status each ten seconds.
>>>>
>>>> clear hdd:
>>>> ~# dd if=/dev/zero bs=4k of=/dev/hdc
>>>>
>>>> discover pid:
>>>> ~$ ps aux|grep dd\ |grep -v grep
>>>> root 1685 30.7 0.2 1892 708 pts/3 R+ 09:06 5:48 dd if /dev/zero bs 4k of /dev/hdc
>>>> ^^^^
>>>>
>>>> In another root terminal, drive the dd status request signal:
>>>> ~# while :; do kill -usr1 1685; sleep 10; done
>>>
>>>killall is your friend...
>>>
>>>$ killall -usr1 dd
>>>
>>>(assuming you only have the one dd running, of course)
>>
>> Well, if you had several running, each in its own terminal?
>>
>> Hmm, almost worked...
>>
>> root@deltree:~# while :; do killall -usr1 dd; sleep 10; done
>> Cannot find user sr1
>> ^C
>> root@deltree:~# while :; do killall -USR1 dd; sleep 10; done
>mea culpa!
>
>> And yes, two running instances of dd updated together
>What the heck, why not go for 3 or 4 ?

Because it's a single drive and all one does is lose time in head
seeking, so writing two adjacent partitions ran at about half speed.

But if I was lucky enough to be wiping 5 drives in parallel? It'd
be fun. But I haven't had that many drives in one spot since once
I bought 5 x 270MB SCSI drives to play with RAID about a decade ago,
discovering RAID on narrow scsi was a silly idea
>
>I'm sort of curious about the filling it with zeroes... was there top
>secret stuff on it?

No. Filling a drive with zeroes gives the hdd controller a chance to
remap any iffy sectors. Modern hdd lie to OS about sector size, so
an iffy sector can only be replaced when OS asks for entire 'real'
(much larger) sector is written. Thus writing all zeroes to a drive
is somewhat like the old 'low level format'.

In fact this is all that the Seagate disk recovery software (for
warranty testing) does, write zeroes to entire drive then test it.

Grant.
--
http://bugsplatter.id.au

On Wed, 30 Sep 2009 09:19:14 -0400, Harold <harold.TakeThisOut@everywhere.com> wrote:

....
>Another method is to use the ddrescue command like the following.
>
>
>ddrescue -v -b4k /dev/zero /dev/hdc

~# ddrescue --version
-bash: ddrescue: command not found

Not in slack-11

Didn't ddrescue features get folded into dd? Or was that a special
dd I saw somewhen?

Grant.
--
http://bugsplatter.id.au

Hallo, Grant,

Du meintest am 01.10.09:

>> Another method is to use the ddrescue command like the following.
>>
>>
>> ddrescue -v -b4k /dev/zero /dev/hdc

> ~# ddrescue --version
> -bash: ddrescue: command not found

> Not in slack-11

Try "linuxpackages.net", search for "any" version. You'll find
"ddrescue-1.10-i486-1McD.tgz".

> Didn't ddrescue features get folded into dd?

No - sorry.

Viele Gruesse
Helmut

"Ubuntu" - an African word, meaning "Slackware is too hard for me".

Grant wrote:

>>killall is your friend...
>>$ killall -usr1 dd

> root@deltree:~# while :; do killall -USR1 dd; sleep 10; done
>
> And yes, two running instances of dd updated together

Even more elegant?

watch -n 10 killall -USR1 dd

--
----------------------------------------------------------------------
Sylvain Robitaille syl.TakeThisOut@encs.concordia.ca

Systems analyst / AITS Concordia University
Faculty of Engineering and Computer Science Montreal, Quebec, Canada
----------------------------------------------------------------------

On Wed, 30 Sep 2009 21:23:26 +0000 (UTC), Sylvain Robitaille <syl.TakeThisOut@alcor.concordia.ca> wrote:

>Grant wrote:
>
>>>killall is your friend...
>>>$ killall -usr1 dd
>
>> root@deltree:~# while :; do killall -USR1 dd; sleep 10; done
>>
>> And yes, two running instances of dd updated together
>
>Even more elegant?
>
> watch -n 10 killall -USR1 dd

I never used watch -- it's hidden under a few high use-polished tools
over in the far corner

Grant.
--
http://bugsplatter.id.au

On Wed, 30 Sep 2009 02:21:21 +0200, Simon Sibbez <simon.sibbez.DeleteThis@buerotiger.de> wrote:

>Grant wrote:
>
>> Preparing an old hdd for reuse, first I want to wipe it to zeroes
>> with dd. I'm curious about progress too, so here's how to see the
>> dd status each ten seconds.
>
>Don't do that in a busybox environment. The dd implementation just
>quits when it receives the signal.

Slack-13 (32bit) installer's dd works fine handling this signal today.

Busybox dd didn't show bytes transferred, so I used 'bs=1M'

Grant.
--
http://bugsplatter.id.au