Re: Multi threads downloader

[Jeffrey Borg <jeffrey@borgs.net>]

> I don't like these fscked downloaders and discourage users of using them.
> >From the user point of view It's ok, but on the other hand my squid goes
> crazy. And let's get am example scenario:
> 
> 1. Client is using such kind of multithreaded downloader.
> 2. Proxy is squid type.
> 3. Server from wich download is, does not support resuming.
> 
> In this case downloader says to squid get for me file.zip at positions 0k,
> 1000k, 2000k, 3000k... because of lack of capabilities at the server side,
> the squid will handle such a request by downloading file from 0k, then
> from 0 to 1000k, then from 0 to 2000k....
> Believe me you won't like to see this happening at your site. Even flowcontrol
> brokes. Virtualy squid download file at same speed as client, but when above
> scenario occurs, squid must to download at full speed between 0 and 1000k...

I do know that netapps cache appliance when trying to do this from a
server wich does not support resuming will send the data to the client
from byte 0

I am pretty sure squid does this too

anyway the other downside to this is that the proxy dosen't cache such
things anyway.

So I think lack of capabilities at the serverside equate to lack of
abilities at the proxy.

But I am not 100% sure for squid

and the idea sucks why not try to find a closer mirror in the first place.

Jeff