Re: Delay_Pool

> Hi David and all the others...
>
> We've got the delay pools running based on the client-ip-address in the
> HTTP-Header.
> Now we're facing another problem:
>
> If somebody downloads an object with the maximum allowed bandwidth he
> will never ever be able to make any other requests during this download.
> Did we miss something in the configuration options on allowing
> concurrent downloads/requests from the same machine?

Sounds like the scanning order problem, which could also be considered to be
a problem that delay pools allocations should be split up to the number of
active clients on the pool. Don't have the time to explain what's happening
right now - read the archives. I don't expect to fix it in the near future,
although there are a few possible solutions, most of them have a performance
or memory cost.

David.
Received on Wed Jun 09 1999 - 21:28:21 MDT