Re: Wheezy release: CDs are not big enough any more...
[Steve McIntyre]
> (http://www.debian.org/releases/stable/amd64/ch04s03.html.en)
While it is refreshing to see "cat debian.iso > /dev/sdX" instead of
the usual dd nonsense (it seems there's an extremely widespread myth
that you need to use dd any time you're reading or writing block
devices), I think "cp" is even more straightforward. Bonus: you can
easily run it with sudo. ("sudo cat debian.iso > /dev/sdX" does not
do what a novice might think.)
Though I suppose it might be annoying to those who feel the need for
alias cp='cp -i' in .bashrc. But hey, it's their choice to be annoyed
by things like this. (:
--
To UNSUBSCRIBE, email to debian-cd-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120515174055.gd2...@p12n.org
Re: Wheezy release: CDs are not big enough any more...
[Samuel Thibault] > > I think "cp" is even more straightforward. > > Does cp accept that way since a long time? I'm not sure, but I've been using things like "cp boot.img /dev/fd0" for probably 10 or 15 years on various Linux and Unix systems. (The fact that I referred to a floppy drive may give some idea of how long) I am not sure where the idea came from that reading or writing block devices always requires 'dd', but if I were to guess, I'd say we can blame tape drives (which aren't even block devices, but char devices). As I recall, you can choose the block size when you format or write a tape, and maybe there are ancient systems out there in which userspace must be explicit when reading and writing them. (Normally, though, I _think_ you just tell the kernel tape driver the right parameters using, e.g. 'mt', then let it handle writing full blocks.) Peter -- To UNSUBSCRIBE, email to debian-cd-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120515230005.ge2...@p12n.org
Re: Wheezy release: CDs are not big enough any more...
[Steve McIntyre]
> The major win with dd onto a raw device is that you can specify the
> block size. For most USB sticks, using a block size of 4MB or so is
> going to be *much* faster than using the default for dd (512 bytes)
> or cp (10 KB IIRC).
That seemed a little fishy to me, since none of the above commands do
any fsync by default, so I just benched it locally.
Writing a 50 MB businesscard image to a USB flash drive on my system
(numbers are MB/s):
dd bs=512 1.771.781.77
dd bs=1024 1.791.761.77
dd bs=2048 1.771.781.78
dd bs=4096 2.542.532.51
dd bs=8192 2.482.502.55
dd bs=4194304 2.502.502.54
cp 2.492.472.48
So it appears that if you aren't going to specify a bs= parameter here,
there's no point in using dd, unless you just happen to think its
command line syntax is particularly charming. And even if you do
specify bs=, you'll only barely beat cp.
For completeness, the same test writing a small file (1 MB),
unsurprisingly, is quite inconclusive:
dd bs=512 1.441.040.98
dd bs=1024 1.001.061.04
dd bs=2048 0.821.041.05
dd bs=4096 1.301.311.35
dd bs=8192 1.061.521.56
dd bs=4194304 1.191.281.27
cp 1.141.291.27
--
Peter
#!/bin/sh
infile=/tmp/debian-6.0.2.1-amd64-businesscard.iso
MB=$(stat -c'scale=2; %s/1048576' $infile | bc)
outfile=/dev/sdd
test_start () {
label=$1
sudo sysctl vm.drop_caches=1 > /dev/null # probably not really needed
t0=$(date +%s.%N)
}
test_end () {
sync
echo "$label : $(echo "scale=2; $MB / ($(date +'%s.%N') - $t0)" | bc) MB/s"
}
for n in $(seq 1 3); do
for sz in 512 1024 2048 4096 8192 $((1024*4096)); do
test_start bs=$sz
dd bs=$sz if=$infile of=$outfile 2> /dev/null
test_end
done
test_start cp
cp $infile $outfile
test_end
done
--
To UNSUBSCRIBE, email to debian-cd-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120516020029.gf2...@p12n.org
Re: Wheezy release: CDs are not big enough any more...
[Steve McIntyre] > You're not measuring the time taken to sync to the flash drive > either, so all you're going to be seeing is the speed of writing to > cache. Huh, I figured the 'sync' call at the end of each test run covered that. > I've done lots of work with USB flash and MMC/SD cards over the last > few years, and the best results are typically achieved using "dd > bs=4M oflag=sync". That way, you'll normally get nicely-aligned date > writes big enough to cover the internal flash page size and remove > the horrendous effects of read-modify-write cycles. Not noticeable in my test runs, so maybe I have an abnormal flash disk. (The fact that it has a USB interface, rather than something closer to the flash controller, probably makes a difference.) Anyway, I've never been against people recommending things like "dd bs=4M oflag=sync" when writing to disk media. My pet peeve is when people recommend "dd" but without any options other than if= and of=. It is clear that many such people don't have a clue _why_ they use dd, except an irrational, dare I say cargo-cult, aversion to cp with block devices. -- To UNSUBSCRIBE, email to debian-cd-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120516220409.gg2...@p12n.org
Re: More tasks option in Tasksel: what tasks do you want there? (reloaded)
> On Mon, Sep 8, 2014 at 11:15 PM, Thomas Goirand wrote: > > - database-server: commonly one would expect MySQL, and postgress gets > > installed [Paul Wise] > Isn't tasksel for people with no expectations? People who know > something about the technology they are looking for will install the > relevant packages instead of following tasksel recommendations. Yeah but in what possible world would anybody want a "database server" but not care which DBMS it is? I mean, on the basic level of MySQL vs. Postgres, not "which fork of mysql is cool this week." I just can't fathom the use case for this particular task. Yes, there are cases where you need a DBMS but you don't have an opinion, but I suspect in that case what you need the DBMS for is some other package, which then Depends: or Recommends: on a suitable DB engine. Peter -- To UNSUBSCRIBE, email to debian-cd-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140908230757.ga3...@p12n.org
