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

Re: [leafnode-list] fetchnews in parallel?

* Michael O'Quinn wrote on Thu, Jul 11, 2002 at 07:47:19AM CEST:
> > On Thu, 11 Jul 2002, Matthias Andree wrote:
> > > Michael O'Quinn schrieb am Mittwoch, den 10. Juli 2002:
> > > 
> > > > Now when I am retrieving this cached news across my internal 10 Base TX
> > > > network, server limitations do become an issue.  But surprisingly, it's
> > > > still not disk bandwidth, but CPU on the server.
> > > 
> > > Leafnode or NewsPlex? If it's leafnode, I might want to peek at a
> O.K., this is where I screwed up.  Leafnode DOES sometimes get bogged down
> when articles are being fetched from it.  I had it fixed firmly in my mind
> that we were talking about fetchnews fetching across the Internet, but
> after actually READING (Duh!) what you snipped, it becomes fairly obvious
> that we are talking about fetching news FROM the leafnode server.  And
> it's actually a 100 Base TX, not 10 Base TX, so server limitations tend to
> stand out more.

Ok, I'm interested in this.  The startup grouinfo issue is known, but
the fetching issue is not (to me).  Over here, I can get around 80
articles/second from a leafnode-2-current server (across loop device)
within one group and around 20 art./sec with usual changing of groups
and _no_ fetchnews/applyfilter/texpire running.  That is with a client
fetchnews-2-current btw, a version 1.9.x fetchnews does worse.  Leafnode
gets hogged whenever someone writes in the spool, because it will try
rebuilding the .overview.  Some numbers could help identifying a problem


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