On Thu, 10 Jul 1997 proff@suburbia.net wrote: > The different article id's is unfortunately a problem that can't > be resolved. Infact nntpcache goes to quite some lengths to avoid Do the article numbers in the groups have to be exactly sequential? I mean, how exactly is the reader using those to know that an article's already been read? Is it just, "I've read up to article 77594?" Or, does it actually store a list of read articles. If they don't need to be sequential, couldn't you hash up a *new* article number based on the message-id or epoch timestamp of posting? Just perusing RFC977, it seems that a reader can select an article based on number, or message-id. Is the message-id just a fall back, or a worthwhile approch? > to the same group from different servers. Theoretically you could > run two news servers on the same news-spool via mirroring or nfs. Code Right, but eventually, you expand beyond the limits of a single NFS server, not to mention that you have a *big* single point of failure there. > You could also [without modification] just use a round-robin-dns entry > for the two servers. Right, but again, they would have to be completely in sync. Maybe using INN slaves is the right way to go. Maybe I ask too much. :)
