Yeah, repeatedly running process can utilize the great strength of
adaptive URI selector. Actually it is particularly designed for that
purpose :)


On 3月3日, 午前5:09, Peter Poeml <[email protected]> wrote:
> Hi Tatsuhiro,
>
> On Fri, Feb 27, 2009 at 02:28:20AM -0800, Tatsuhiro wrote:
>
> > Congratulations!  Many Great improvements.
> > I am also grad to hear that aria2 is in the backend of client.
> > Are you using adaptive URI selector of aria2 or more conventional one?
>
> So far, the conventional URI selector was in use. But I just noticed the
> extreme advantage that the adaptive one brings about. I did some
> experiments today. I hacked download.opensuse.org to return "broken
> mirrors" when the client sends an "X-Broken-Mirrors: true" header. The
> first mirrors at the top of metalinks will be bogus and broken URLs that
> will lead to various failures. (And everybody is welcome to use that
> server's metalinks for testing.)
>
> And I realized very quickly that trying a series of unreachable mirrors,
> one after the other, will make the client eventually succeed, but it
> will take such a long time for each file to be downloaded that this is
> leading nowhere, when dozens or hundreds of files are to be downloaded.
>
> I had played with aria2c's server-stat-of and server-stat-if directive
> before, but today was the first time I also added uri-selector=adaptive
> which makes it actually use this saved data.
>
> Wow, this is amazing. A difference like day and night.
>
> What a fantastic ground that you, Aurelien Lefebvre and Pascal Rigaux
> have layed here (and whoever else - I was just looking at the changelog).
>
> Thanks!
> Peter
> --
> Contact: [email protected] (a.k.a. [email protected])
>          #opensuse-mirrors on freenode.net
> Info:http://en.opensuse.org/Mirror_Infrastructure
>
> SUSE LINUX Products GmbH
> Research & Development
>
>  application_pgp-signature_part
> < 1K表示ダウンロード
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Metalink Discussion" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/metalink-discussion?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to