On Mon, 24 Jun 2013, paul condon wrote:
> I have two 3TB Seagate external HDs. They were purchased from
> different stores at slightly different times earlier this year, here
> in Colorado. I want them to have ext4 file systems on them,
> excepting if someone on this list can give a reason otherwise. I
Ext4 is a good option.
> have googled and gotten a lot of hits, which indicate to me that
> this is a well known problem. Unfortunately, I have difficulty
> following the instructions, and all my efforts have not reached a
> successful conclusion. Now with further trys, it seems to me that
> stuff has been written onto the drives that needs to be wiped off
> because I get messages that from the disk utility in xfce4 that it
> won't overwrite a disk with data on it.
Note that old versions of fdisk, parted, etc can't handle 3TB drives
properly. If you're running the latest release of Debian (Wheezy) you
should be OK though -- if not, be careful.
Some ancient SATA cards and quite a few USB adapters can't handle 3TB
drives at all -- make sure that 'unformatted size' is reported correctly
and you're probably OK.
> So I want to use dd to wipe a complete drive. For this I have found
> the following:
> # dd if=/dev/zero of=/dev/sdg bs=1M
It's very unlikely that you need to wipe the full drive. You may just
have to clear metadata at the front. This would do that:
dd if=/dev/zero of=/dev/sdg bs=1M count=1
> In the above, I have already changed the HD device to sdg (from
> sda), but I wonder about bs=1M. Could the process go faster with a
> larger block size?
Not really. bs=1M is fine for what is being discussed. dd just pushes
'count' data blocks of size 'bs' through to its output (/dev/sdg in your
example). You can hit performance issues if the chosen bs clashes with
an internal cache and/or the native block size of the disk, but that
requires small (kB'ish) block sizes.
> What are the criteria for choosing a value for bs? And, how long
> should a 3T wipe take to complete? The job has been running for about
(3*1024*1024 MB)/(200 MB/sec) / (3600 sec/hour) = 4.3 hours
200 MB/sec may be optimistic. I get 180 MB/sec on my (mediocre) system.
Note that if you're connected with USB2, then you're looking at 10
MB/sec and 86 hours...
> 12 hours. Would it go faster with a different bs? Faster enough to
> make the waste of 12hrs running worthwhile? Is there some way to
> invoke an 'progress indicator' for dd? And, in general, is there a
> better way?
From 'man dd' (towards the end):
Sending a USR1 signal to a running `dd' process makes it print I/O statistics
to standard error and then resume copying.
$ dd if=/dev/zero of=/dev/null& pid=$!
$ kill -USR1 $pid; sleep 1; kill $pid
18335302+0 records in 18335302+0 records out 9387674624 bytes (9.4 GB)
copied, 34.6279 seconds, 271 MB/s
-- Brad
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20130624162518.ga21...@aether.jlab.org
