Bug#85730: Inn2 does not reread entirely.
Russ Allbery
rra at stanford.edu
Tue Feb 13 00:57:37 UTC 2001
Andreas Krüger <andreas.krueger at online.life.de> writes:
> For one thing, that error messages should clearly mention xrefslave. I
> have a strong feeling that other news administrators will continue to
> burn time in similar circumstances, until those error messages are made
> clearer.
> Here's one simple possible patch to that respect:
> bash-2.02$ diff -u innd/art.c.original innd/art.c
> --- innd/art.c.original Thu Jan 11 09:55:23 2001
> +++ innd/art.c Mon Feb 12 19:44:16 2001
> @@ -2459,11 +2459,11 @@
> if (innconf->xrefslave) {
> if (ARTxrefslave() == FALSE) {
> if (HDR(_xref)) {
> - (void)sprintf(buff, "%d Invalid Xref header \"%s\"",
> + (void)sprintf(buff, "%d Xref header \"%s\" invalid in xrefslave mode.",
> NNTP_REJECTIT_VAL,
> MaxLength(HDR(_xref), HDR(_xref)));
> } else {
> - (void)sprintf(buff, "%d No Xref header",
> + (void)sprintf(buff, "%d Xref header required in xrefslave mode.",
> NNTP_REJECTIT_VAL);
> }
> ARTlog(&Data, ART_REJECT, buff);
This looks good; I'll put this into CVS.
> Secondly, inn should set its internal state "innconf->xrefslave" to the
> default "no" whenever it is about to reread inn.conf.
The problem here that you're running into is that ctlinnd reload doesn't
reload inn.conf. This is a known shortcoming in INN at present and is
listed in the TODO file as a project for someone with a lot of time (it's
hard to fix, since inn.conf is pervasive).
The recommended action after changing something in inn.conf is to restart
INN with:
ctlinnd xexec innd
which shuts it down completely and restarts it. (You encountered this
problem because nnrpd re-reads inn.conf each time it starts, so the change
to nnrpdpostinghost was "immediate" while the change to xrefslave wouldn't
be picked up until INN was stopped and restarted.)
--
Russ Allbery (rra at stanford.edu) <http://www.eyrie.org/~eagle/>
More information about the inn-bugs
mailing list