* Adam Borowski , 2012-07-22, 23:51:
having an option to disable fsync in dpkg without unreliable LD_PRELOAD
tricks would be great.
http://bugs.debian.org/613428
--
Jakub Wilk
--
To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact
Hi,
On Sat, 21 Jul 2012 19:13:45 -0400
Joey Hess wrote:
> Which is why I asked for actual, real-world benchmarks...
As I said in my presentation, I just tested a few case,
fonts-horai-umefont, poppler-data and openclipart-png.
Install fonts- and poppler-data is almost same time, but opencli
On Mon, 23 Jul 2012, Adam Borowski wrote:
> You tested ext4. On btrfs, dpkg is around an order of magnitude slower,
> making using it without eatmydata a laughable idea.
>
> And that's on a filesystem whose features include:
> * transactions (so all dpkg processing could be done without a single
On Sun, Jul 22, 2012 at 07:58:42PM +0300, Uoti Urpala wrote:
> Simon Paillard wrote:
> > , I understand debian-installer ask dpkg not to fsync:
>
> > - Run dpkg with --force-unsafe-io during installation; syncing is
>
> This only affects one particular instance of syncing (which I think may
On Sun, Jul 22, 2012 at 01:58:59AM +0200, Adam Borowski wrote:
> > BTW, when we switched to building udebx with xz, Philipp Kern benchmarked
> > it using little or no additional CPU to decompress xz produced with
> > -Zxz -z1 -Sextreme http://lists.debian.org/debian-boot/2011/10/msg00247.html
> Per
Simon Paillard wrote:
> On Sat, Jul 21, 2012 at 09:25:50PM +0300, Uoti Urpala wrote:
> > So at least in this case the biggest performance problem by far is the
> > inappropriate use of fsync() or other disk synchronization primitives,
> > and CPU use for unpacking is pretty much irrelevant.
>
> Th
On Sat, Jul 21, 2012 at 11:42:03AM -0400, Joey Hess wrote:
> Hideki Yamane wrote:
> > On Sun, 8 Jul 2012 17:58:16 +0200
> > Adam Borowski wrote:
> > > • xz -6 (the default) is a lot slower when compressing, fast when
> > > decompressing, needs only 10MB memory, 58% size
> > > • xz -9 has v
Mike Hommey wrote:
> Note that slower decompression doesn't necessarily mean longer
> installation time. I/O is still more time consuming than CPU.
Which is why I asked for actual, real-world benchmarks...
--
see shy jo
signature.asc
Description: Digital signature
Hi,
On Sat, Jul 21, 2012 at 09:25:50PM +0300, Uoti Urpala wrote:
> Joey Hess wrote:
> > Hideki Yamane wrote:
> > > I tested as well, and sometimes decompression with xz is so slw,
> > > it takes 6-8 times than default gz.
> >
> > I was just watching your DebConf presentation "Lets shrink De
brian m. carlson wrote:
> On Sat, Jul 21, 2012 at 09:25:50PM +0300, Uoti Urpala wrote:
> > So at least in this case the biggest performance problem by far is the
> > inappropriate use of fsync() or other disk synchronization primitives,
> > and CPU use for unpacking is pretty much irrelevant.
>
>
On Sat, Jul 21, 2012 at 09:25:50PM +0300, Uoti Urpala wrote:
> Most of the time taken by cdebootstrap is wasted by dpkg on doing
> useless file syncs:
>
> cdebootstrap --arch=amd64 unstable debian-tree/
>
> from local package cache on ext4: 138 seconds
>
> on tmpfs where dpkg can't waste time on
Joey Hess wrote:
> Hideki Yamane wrote:
> > I tested as well, and sometimes decompression with xz is so slw,
> > it takes 6-8 times than default gz.
>
> I was just watching your DebConf presentation "Lets shrink Debian
> package archive" and I think there you said decompression with xz was
>
On Sat, Jul 21, 2012 at 11:42:03AM -0400, Joey Hess wrote:
> Hideki Yamane wrote:
> > On Sun, 8 Jul 2012 17:58:16 +0200
> > Adam Borowski wrote:
> > > • xz -6 (the default) is a lot slower when compressing, fast when
> > > decompressing, needs only 10MB memory, 58% size
> > > • xz -9 has v
Hideki Yamane wrote:
> On Sun, 8 Jul 2012 17:58:16 +0200
> Adam Borowski wrote:
> > • xz -6 (the default) is a lot slower when compressing, fast when
> > decompressing, needs only 10MB memory, 58% size
> > • xz -9 has very slow compression, takes gobs of memory, 56% size
> > (Obviously
Hi,
On Sun, 8 Jul 2012 17:58:16 +0200
Adam Borowski wrote:
> • xz -6 (the default) is a lot slower when compressing, fast when
> decompressing, needs only 10MB memory, 58% size
> • xz -9 has very slow compression, takes gobs of memory, 56% size
> (Obviously, the "size" numbers are dra
On Sat, Jul 07, 2012 at 04:22:58PM -0600, Stefano Zacchiroli wrote:
> Ansgar has been experimenting with .deb sizes to make the packages
> needed for a minimal desktop installation fit in the first CD. It looks
> like that's doable by switching to xz compression for the involved
> binaries. Would
Le Sun, Jul 08, 2012 at 02:12:44PM -0600, Joey Hess a écrit :
>
> grub-legacy is still used for multipath and sataraid.
> Something was going to be done to make grub2 support those, but
> I don't know the status.
Hi,
Grub-legacy is also useful for booting virtual machine images with pv-grub
(suc
> Ben Hutchings writes:
[...]
> - twm: no-one should have to suffer this
And, exactly, why not? Before I've switched to Openbox, it was
one of the two WM's I've used, along with FVWM. And they say
[1] that it still can be handy at times.
The “obscure” lab
On Sun, Jul 08, 2012 at 02:12:44PM -0600, Joey Hess wrote:
> Cyril Brulebois wrote:
> > It looks to me like a current debian-installer build installs grub2,
> > with no option for grub-legacy, even in expert mode.
>
> grub-legacy is still used for multipath and sataraid.
> Something was going to b
Cyril Brulebois wrote:
> It looks to me like a current debian-installer build installs grub2,
> with no option for grub-legacy, even in expert mode.
grub-legacy is still used for multipath and sataraid.
Something was going to be done to make grub2 support those, but
I don't know the status.
--
s
Wouter Verhelst (08/07/2012):
> On Sun, Jul 08, 2012 at 01:17:32AM +0200, Bastian Blank wrote:
> > Even grub-legacy?
>
> Yes; d-i in expert mode still has the ability to explicitly choose for
> grub legacy, if you really want to.
It looks to me like a current debian-installer build installs grub
Hi
Stefano Zacchiroli writes:
> Ansgar has been experimenting with .deb sizes to make the packages
> needed for a minimal desktop installation fit in the first CD. It looks
> like that's doable by switching to xz compression for the involved
> binaries. Would you grant freeze exceptions for pac
On Sun, Jul 08, 2012 at 01:17:32AM +0200, Bastian Blank wrote:
> On Sat, Jul 07, 2012 at 04:36:34PM -0600, Joey Hess wrote:
[...]
> > > - grub-legacy
> > These are all installed by d-i in various situations.
>
> Even grub-legacy?
Yes; d-i in expert mode still has the ability to explicitly choose
Hi,
On Sat, 7 Jul 2012 04:47:35 +0100
Steve McIntyre wrote:
> So, yes - looks like xz will make a difference here for the wheezy
> release, for amd64 at least. It's enough that we'd probably even have
> space for the installation manual and release notes to fit \o/.
BTW, I'll talk about using x
On Sat, Jul 07, 2012 at 08:42:07PM +0100, Ben Hutchings wrote:
> On Sat, 2012-07-07 at 16:14 +0200, Bastian Blank wrote:
> - ftp, telnet: mostly redundant with wget and nc, unless you really like
> cleartext authentication
I don't get why anyone would talk about "authentication" in the context of
On Sat, Jul 07, 2012 at 04:22:58PM -0600, Stefano Zacchiroli wrote:
> On Sun, Jul 08, 2012 at 12:03:44AM +0200, Cyril Brulebois wrote:
> > Ansgar Burchardt (07/07/2012):
> > > Another question is if the release team would consider unblocking
> > > updated packages (with just the switch to xz for b
Quoting Ben Hutchings (b...@decadent.org.uk):
> > partman-crypto still installs this.
>
> Seems like a bug...? The package is orphaned and dm-crypt has support
> for a compatible encryption mode.
Given the level of attention which partman-crypto got during the
squeeze->wheezy release, I'd bet
On Sat, 2012-07-07 at 16:36 -0600, Joey Hess wrote:
> > - ftp, telnet: mostly redundant with wget and nc, unless you really like
> > cleartext authentication
> > - procmail: server
>
> These are priority standard.
Which is not the same as being important to include in a desktop
installation.
> >
On Sat, Jul 07, 2012 at 04:36:34PM -0600, Joey Hess wrote:
> > - ftp, telnet: mostly redundant with wget and nc, unless you really like
> > cleartext authentication
> > - procmail: server
> These are priority standard.
This is fixeable.
> > - pcmciautils: PCMCIA is dead
> > - jfsutils, reiserfspr
> - ftp, telnet: mostly redundant with wget and nc, unless you really like
> cleartext authentication
> - procmail: server
These are priority standard.
> - pcmciautils: PCMCIA is dead
> - jfsutils, reiserfsprogs, ufsutils: obscure
> - discover
> - openssh-server: server, not desktop
> - grub-lega
On Sun, Jul 08, 2012 at 12:03:44AM +0200, Cyril Brulebois wrote:
> Ansgar Burchardt (07/07/2012):
> > Another question is if the release team would consider unblocking
> > updated packages (with just the switch to xz for binaries). I think
> > it would be nice to be able to get a useful desktop s
Ansgar Burchardt (07/07/2012):
> Another question is if the release team would consider unblocking
> updated packages (with just the switch to xz for binaries). I think
> it would be nice to be able to get a useful desktop system using just
> the first CD, but I'm not sure if they would make an e
On Sat, 2012-07-07 at 16:14 +0200, Bastian Blank wrote:
> On Sat, Jul 07, 2012 at 02:34:44PM +0100, Steve McIntyre wrote:
> > The list posted there is the full sorted list of *all* packages, as
> > applied to the full set of CDs. The last one on CD#1 is
> > gnome-packagekit-data, as I said, and I d
On Sat, Jul 07, 2012 at 02:34:44PM +0100, Steve McIntyre wrote:
> The list posted there is the full sorted list of *all* packages, as
> applied to the full set of CDs. The last one on CD#1 is
> gnome-packagekit-data, as I said, and I don't see any debug packages
> above that in the list.
Ups, I di
On Sat, Jul 07, 2012 at 03:10:15PM +0200, Bastian Blank wrote:
>On Fri, Jul 06, 2012 at 10:10:04PM +0100, Steve McIntyre wrote:
>> http://cdimage.debian.org/cdimage/tmp/new-tasks/gnome-cd.list.gz
>
>Why does this contain debug packages?
The list posted there is the full sorted list of *all* pack
On Fri, Jul 06, 2012 at 10:10:04PM +0100, Steve McIntyre wrote:
> http://cdimage.debian.org/cdimage/tmp/new-tasks/gnome-cd.list.gz
Why does this contain debug packages?
Bastian
--
I've already got a female to worry about. Her name is the Enterprise.
-- Kirk, "The Corbomite Ma
Steve McIntyre writes:
> On Fri, Jul 06, 2012 at 09:00:32PM -0600, Ansgar Burchardt wrote:
>>I tried recompressing all packages in wheezy with xz. The total size
>>for all amd64+all packages was reduced from 42GB to 33 GB (about 20%).
>>A per-package listing is available from [1]
>>
>> [1]
On Fri, Jul 06, 2012 at 09:00:32PM -0600, Ansgar Burchardt wrote:
>Hi,
>
>Steve McIntyre writes:
>> Back in May I warned about CD sizes[1] for the Wheezy release,
>> pointing out that CD#1 isn't big enough any more to provide usable
>> Gnome or KDE installations. There was some discussion about wh
Hi,
Steve McIntyre writes:
> Back in May I warned about CD sizes[1] for the Wheezy release,
> pointing out that CD#1 isn't big enough any more to provide usable
> Gnome or KDE installations. There was some discussion about what to do
> about that (change compression to xz, switch to the lighter d
As suggested by Ansgar IRL, here's a summary of what compression types
are showing up on each CD (by looking at data.tar.$EXT for all the
.debs and .udebs):
>Gnome
>=
>
>The last package on amd64 CD#1 is gnome-packagekit-data. task-desktop
>fits on CD#1, but task-gnome-desktop is ~110 packages
Steve McIntyre writes:
> Back in May I warned about CD sizes[1] for the Wheezy release,
> pointing out that CD#1 isn't big enough any more to provide usable
> Gnome or KDE installations.
Indeed. CD1 was really problematic in squeeze too:
http://lists.debian.org/debian-boot/2011/08/msg00172.html
Hey people,
Following up on this again...
Back in May I warned about CD sizes[1] for the Wheezy release,
pointing out that CD#1 isn't big enough any more to provide usable
Gnome or KDE installations. There was some discussion about what to do
about that (change compression to xz, switch to the li
42 matches
Mail list logo