Bug#690183: ITP: apt-fast -- shellscript wrapper for apt-get or aptitude
 

Package: wnpp
Severity: wishlist
Owner: Dominique Lasserre <lasserre.d@gmail.com>

* Package name : apt-fast
Version : 0.7.6
Upstream Authors: Matt Parnell, Dominique Lasserre
* URL : https://github.com/ilikenwf/apt-fast.git
* License : GPL-3+
Programming Lang: Bash
Description : shellscript wrapper for apt-get or aptitude

apt-fast is a shellscript wrapper for apt-get that can drastically improve apt
download times by downloading packages in parallel, with multiple connections
per package.
It uses aria2 oder axel as download managers and apt-get --print-uris to get
download URLs.
Bash and Zsh autocompletion included (derived from distributed apt-get).


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: http://lists.debian.org/20121010215331.7523.77516.reportbug@localhost

Bug#690183: ITP: apt-fast -- shellscript wrapper for apt-get or aptitude
 

On Wed, 10 Oct 2012 23:53:31 +0200
Dominique Lasserre <lasserre.d@gmail.com> wrote:
> * Package name : apt-fast

> apt-fast is a shellscript wrapper for apt-get that can drastically improve apt
> download times by downloading packages in parallel, with multiple connections
> per package.

well, isn't it huge load for repository servers?


--
Regards,

Hideki Yamane henrich @ debian.or.jp/org
http://wiki.debian.org/HidekiYamane


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20121011102647.53585a12d8735b8b7cefc48e@debian.or. jp">http://lists.debian.org/20121011102647.53585a12d8735b8b7cefc48e@debian.or. jp

Bug#690183: ITP: apt-fast -- shellscript wrapper for apt-get or aptitude
 

On Oct 11, Hideki Yamane <henrich@debian.or.jp> wrote:

> > apt-fast is a shellscript wrapper for apt-get that can drastically improve apt
> > download times by downloading packages in parallel, with multiple connections
> > per package.
> well, isn't it huge load for repository servers?
As a mirror operator I consider this antisocial.
As a developer, I doubt that this can provide significant speed
improvements other than for pathological cases.

--
ciao,
Marco

Bug#690183: ITP: apt-fast -- shellscript wrapper for apt-get or aptitude
 

On Thu 11 Oct 2012 02:55:24 Marco d'Itri escribió:
> On Oct 11, Hideki Yamane <henrich@debian.or.jp> wrote:
> > > apt-fast is a shellscript wrapper for apt-get that can drastically
> > > improve apt download times by downloading packages in parallel, with
> > > multiple connections per package.
> >
> > well, isn't it huge load for repository servers?
>
> As a mirror operator I consider this antisocial.

True.

> As a developer, I doubt that this can provide significant speed
> improvements other than for pathological cases.

Well, parallel download does **greatly** improves speed when you access
international servers, like we had to do in Argentina until some few weeks
ago.

Of course, being able to download stuff from two different servers at the same
time had a better end result, and as long as is one download at a time per
server, I think it can be considered socially acceptable.

Regards, Lisandro.

--
The generation of random numbers is too important to be left to chance.
http://www.devtopics.com/best-programming-jokes/

Lisandro Damián Nicanor Pérez Meyer
http://perezmeyer.com.ar/
http://perezmeyer.blogspot.com/

Bug#690183: ITP: apt-fast -- shellscript wrapper for apt-get or aptitude
 

On Thu 11 Oct 2012 09:59:35 Lisandro Damián Nicanor Pérez Meyer escribió:
[snip]
> Well, parallel download does **greatly** improves speed when you access
> international servers, like we had to do in Argentina until some few weeks
> ago.

WRT non i386/amd64 archs.

--
Programming today is a race between software engineers striving to build
bigger and better idiot-proof programs, and the Universe trying to produce
bigger and better idiots. So far, the Universe is winning.
Anonymous

Lisandro Damián Nicanor Pérez Meyer
http://perezmeyer.com.ar/
http://perezmeyer.blogspot.com/

Bug#690183: ITP: apt-fast -- shellscript wrapper for apt-get or aptitude
 

On Thu, Oct 11, 2012 at 09:59:35AM -0300, Lisandro Damián Nicanor Pérez Meyer wrote:
> Of course, being able to download stuff from two different servers at the same
> time had a better end result, and as long as is one download at a time per
> server, I think it can be considered socially acceptable.

Yes, which is why there's

Acquire::Queue-Mode "host";

see apt.conf(5) for the full details on this one. You don't need shell
scripts and things like axel to get this.

--
Copyshops should do vouchers. So that next time some bureaucracy requires you
to mail a form in triplicate, you can mail it just once, add a voucher, and
save on postage.


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20121011162250.GD25977@grep.be">http://lists.debian.org/20121011162250.GD25977@grep.be

Bug#690183: ITP: apt-fast -- shellscript wrapper for apt-get or aptitude
 

On Jo, 11 oct 12, 18:22:50, Wouter Verhelst wrote:
> On Thu, Oct 11, 2012 at 09:59:35AM -0300, Lisandro Damián Nicanor Pérez Meyer wrote:
> > Of course, being able to download stuff from two different servers at the same
> > time had a better end result, and as long as is one download at a time per
> > server, I think it can be considered socially acceptable.
>
> Yes, which is why there's
>
> Acquire::Queue-Mode "host";
>
> see apt.conf(5) for the full details on this one. You don't need shell
> scripts and things like axel to get this.

Sorry, but I don't understand how this is supposed to increase my
download speed.

Kind regards,
Andrei
--
Offtopic discussions among Debian users and developers:
http://lists.alioth.debian.org/mailman/listinfo/d-community-offtopic

Bug#690183: ITP: apt-fast -- shellscript wrapper for apt-get or aptitude
 

On Sat, Oct 13, 2012 at 3:02 PM, Andrei POPESCU
<andreimpopescu@gmail.com> wrote:
> On Jo, 11 oct 12, 18:22:50, Wouter Verhelst wrote:
>> On Thu, Oct 11, 2012 at 09:59:35AM -0300, Lisandro Damián Nicanor Pérez Meyer wrote:
>> > Of course, being able to download stuff from two different servers at the same
>> > time had a better end result, and as long as is one download at a time per
>> > server, I think it can be considered socially acceptable.
>>
>> Yes, which is why there's
>>
>> Acquire::Queue-Mode "host";
>>
>> see apt.conf(5) for the full details on this one. You don't need shell
>> scripts and things like axel to get this.
>
> Sorry, but I don't understand how this is supposed to increase my
> download speed.

It works just like in the other mails you get which promise an "increase":
You have to believe that it works - and it eventually will. ¹

If you don't fall for such placebo effects you are unfortunately out of luck
as "host" is the default, but you can use a service like http.debian.net
(which only works so well because "host" is the default …) to get a
real benefit - assuming that you really have more bandwidth available than
one mirror is able to provide you without being unfair/"anti-social"
to all the other users using the same mirror, which is how apt-fast works.


If you wonder why APT is getting all packages from one source (the first one
mentioned in the sources.list) even if it could choose between two or more
sources in your sources.list: It is assumed your sources.list is ordered from
most to least desired source: e.g. CD-ROM, local mirror, "internet" mirror.
Getting a package from the internet if it could be acquired through CD-ROM
would be a waste of resources (be it time, money, traffic, $yourmetric or a
wild mixture of all of them). Same for "local" vs. "internet" mirror.
So the mentioned setting effects situations in which package A and B are
available from different sources - these are acquired in parallel if possible.


¹ note that Wouter hasn't promised that. It was just pointed out that the
desired behavior is actually already implemented (and I am just making it
clear that this is the default value).


Best regards

David Kalnischkies


--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: http://lists.debian.org/CAAZ6_fC22r=cneyfoS9dBR1CKDVoBMXn74fqFH0g24RcO_zFy A@mail.gmail.com

Bug#690183: ITP: apt-fast -- shellscript wrapper for apt-get or aptitude
 

On Sb, 13 oct 12, 16:55:17, David Kalnischkies wrote:
>
> If you don't fall for such placebo effects you are unfortunately out of luck
> as "host" is the default, but you can use a service like http.debian.net
> (which only works so well because "host" is the default …) to get a
> real benefit - assuming that you really have more bandwidth available than
> one mirror is able to provide you without being unfair/"anti-social"
> to all the other users using the same mirror, which is how apt-fast works.

I have yet to find a reliable mirror that is able to saturate my
bandwidth :(
I'll try http.debian.net

> If you wonder why APT is getting all packages from one source (the first one
> mentioned in the sources.list) even if it could choose between two or more
> sources in your sources.list: It is assumed your sources.list is ordered from
> most to least desired source: e.g. CD-ROM, local mirror, "internet" mirror.
> Getting a package from the internet if it could be acquired through CD-ROM
> would be a waste of resources (be it time, money, traffic, $yourmetric or a
> wild mixture of all of them). Same for "local" vs. "internet" mirror.

This is documented in sources.list(5)

> So the mentioned setting effects situations in which package A and B are
> available from different sources - these are acquired in parallel if possible.

Which does not bring significant improvements in most cases.

> ¹ note that Wouter hasn't promised that. It was just pointed out that the
> desired behavior is actually already implemented (and I am just making it
> clear that this is the default value).

Would be nice to mention this also in the manpage ;)

Kind regards,
Andrei
--
Offtopic discussions among Debian users and developers:
http://lists.alioth.debian.org/mailman/listinfo/d-community-offtopic