Re: [leafnode-list] 'Prioritize' servers ?

["D.J. Bolderman" <list-leafnode@xxxxxxxxxxxxxxxxxxx>]

On vr, 23 mei 2003, Matthias Andree wrote:

> > I have defined 2 servers: first my ISP, and then a pay server. When
> > leafnode downloads messages from the first server, how does it act
> > when that same article is also on the second server ? Does the article
> > still gets downloaded, or does leafnode make some kind of link ?
> 
> Leafnode queries the serves in the order presented in your configuration
> file. It will download any articles in the interesting groups from the
> first server, then it will download only those articles that the first
> hadn't had from the second, then download only those articles that the
> first didn't have from the third and so on. It will not download
> the same messages twice. (If it does, some server must have tampered
> with the Message-ID, which is expressly forbidden by Usenet standards.)
> 
> Leafnode requests the list of Message-IDs in that group from the second
> server and then figures it has already downloaded the article in
> question (by the Message-ID) and will skip it.

Ok, thanks for straightening that out.

> > This brings me to my next question: the retention of my pay server is
> > great, so when a multipart message is partially expired on the first 
> > server, does leafnode automatically fill in the gaps with parts of that
> > message from the pay server, or does it download the whole message
> > from the pay server ?
> 
> What exactly is a "multipart message"?

Sorry, I meant an article which is split into idividual messages, as
you described below.

> If it's a file split into individual messages that must be merged by the
> newsreader software, then yes, because leafnode doesn't care about
> whether a specific article is part of a split or no.  Technically, such
> split articles (seem to be common with Subject: blah.jpg (3/4) for
> example) are self-contained articles, and leafnode will fetch the
> missing ones from the second server.

Thanks. This is the solution I was looking for. When I was using
Windows, I had a binary newsreader which would fetch articles first
from my free ISP server, and only from my pay server when articles
were missing.

At this moment, I'm using Leafnode for text-only groups, so I have
delaybody set to zero. This is not an ideal situtation for binary
groups, because I want to make a selection of articles to fetch,
instead of downloading them all.
To overcome this 'problem' can i run  2 instances of Leafnode ? 
eg. one instance with delaybody 0 and one with delaybody 1 ?

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