copying part of a spool from one reader to another

Katsuhiro Kondou kondou at nec.co.jp
Tue May 1 15:29:21 UTC 2001


In article <Pine.LNX.3.96.1010430204431.27736A-100000 at potato.cts.ucla.edu>,
	Chris Stromsoe <cbs at cts.ucla.edu> wrote;

}  -Should this work, or am I missing something obvious and/or stupid?

That should work.  I don't think you need to care whether
dropped disks.  Alive articles are rejected when fed.
What is needed are;

- recreate history and overview on replaced side
- artcutoff is set to 0 on replaced side
- build token list from history on good side and
  use it for innxmit

That worked on my production servers.

}  -The man page for inn.conf mentions that "usually" xreflave's only have
} one feeder.  Any gotchas?  Danger above and beyond normal operation?

It's ok in your case.

}  -I want to keep the Xref header the same so that examination of the
} articles on either machine won't show any differences.  Does innxmit pass
} the article completely "as is" to the system it's feeding?  Does innd in
} xreflave mode try to muck with Xref to modify it to match the PTR of the
} sending host?  (I don't want cases where one machine shows an article
} originating from the feeder, but the other shows it originating from a
} reader.)

Innxmit never modifies Xref: and innd neither does
if in xrefslave mode.

}  -It looks like I'm going to have to feed around 700,000 articles over. 
} Is innxmit the right way to go for this? 

Yes.
-- 
Katsuhiro Kondou


More information about the inn-workers mailing list