How to delete all partitions for a complete reinstall

[Tony Arnold]

Alex,

On Tue, 2006-08-15 at 11:55 +0200, Alexander Skwar wrote:

> With Erics command, a "huge" number of write requests is sent (512,
> to be exact). And every time, just 1 byte is written. That's tremendously
> slow!

And possible 512 reads too, as for each byte written, dd would have to
read the block first before writing it back again. It could cache the
block, of course.

> But test the following, which will create a FILE of 1MiB in the
> current directory:
> 
>         time dd if=/dev/zero of=zero.bs1    count=$((1*1024*1024)) bs=1
>         time dd if=/dev/zero of=zero.bs1MiB count=1 bs=$((1*1024*1024))

Interesting! On my machine the real time for the first command was about
8 seconds. See:

zzalsaca at beeblebrox:~$ time dd if=/dev/zero of=zero.bs1    count=
$((1*1024*1024)) bs=1
1048576+0 records in
1048576+0 records out
Segmentation fault

real    0m8.070s
user    0m0.556s
sys     0m6.968s
zzalsaca at beeblebrox:~$ time dd if=/dev/zero of=zero.bs2    bs=
$((1*1024*1024)) count=1
1+0 records in
1+0 records out
Segmentation fault

real    0m0.035s
user    0m0.000s
sys     0m0.020s
zzalsaca at beeblebrox:~$ ls -l zero.bs*
-rw-r--r-- 1 zzalsaca zzalsaca 1048576 2006-08-15 13:39 zero.bs1
-rw-r--r-- 1 zzalsaca zzalsaca 1048576 2006-08-15 13:40 zero.bs2

But I also get a segmentation fault in the middle of dd's output. Is
this due to running the -26 kernel? or something else?

Regards,
Tony.
-- 
Tony Arnold <tony.arnold at manchester.ac.uk>
University of Manchester