[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[leafnode-list] Re: leafnode-2: fetchnews always says "0 articles posted"
> Matthias Andree:
> Ah gee. Why did it go into failed.postings? I guess we need a counter
> for failed postings as red flags...
that's what i'd like to kow myself. unfortunately, i do not post that
often. for this to diagnose correctly, i'd have to wait until the cron-
scheduled fetchnews runs and immediately afterwards check the log of the
NNTP dialog as it shows on the wire. this is possible, because i keep
a tcpdump instance on the network interface, but i have to hurry,
because this log is obviously growing rapidly and therefore rotated out
of existence often.
i'll try nonetheless.
> > i noticed that the line "logstderr = 1" in etc/leafnode/config
> > applies only to fetchnews, but not leafnode. currently, i have this
> > knob turned off.
> Yup, that's because some so-called "superservers" of the inetd kind
> were reported to spit out stderr stuff over the wire, which isn't
> intended, so leafnode suppresses this.
well, i'd like to get the meaningless disconnects from leafnode clients
out of the log somehow. they are logged as errors, but they really are
more like informational message. and i don't want to see them
syslog(3)ed, so i have a leafnode as a supervised process like this:
# $Header: /service/leafnode/run,v 1.6 2007/05/14 11:49:40 root Exp root $
exec /command/chpst \
-i /etc/ipsvd/leafnode \
-u news:news \
127.0.0.1 119 \
see <URL:http://smarden.org/runit/> and <URL:http://smarden.org/ipsvd/>.
now i have the leafnode "cruft" in a log rotated by size.
leafnode-list mailing list