[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [leafnode-list] Bandwidth limiter for fetch downloads?

 > sss@xxxxxxxxxxx wrote:

> > Thats pretty odd. you must be with a wimpy ISP or you are grabbing from a
> > news server across the net' somewhere.
> No. That's normal when you fetch news from non-binary newsgroups because
> fetch must do quite some work inbetween, like parsing headers etc. At
> the moment fetch is still written in a way that it sends a command, waits
> for the answer from the server and sends the next command. This makes
> it unnecessarily slow. The next major update (not 1.9.3) will hopefully
> fix that.
> I think that sss@xxxxxxxxxxx fetches a lot of binary groups or groups with
> very long message bodies. In this case, fetch will spend most of its time
> writing bodies to disk. This is not very CPU-intensive, so the suggestions
> made in this thread (like renicing fetch, etc.) won't help.
> No, I can't offer anything better.
> --Cornelius.

OK, that makes sense now.  I actually am running leafnode on my server,   
chez.dyn.tj, and I am about 50/50 balance between binary and text content.
I do notice the bandwidth being lower when it is processing smaller
articles.  As you say, the methods discussed here will not work, yes.  
I also understand that what is involved is way beyond the scope of
The only way I can see to solve this is to setup a dedicated linux server   
on a LAN and utilize the Linux Bandwidth management features becoming
available in the 2.2 kernels.  Ill start to look into that. :) Or get
access to a heavily loaded news server....
- Stefan

leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list