> | i have
> | smtp_accept_max_per_host = 4
> That setting works fine for us. Can't imagine what's wrong. You've HUP'd
> the daemon of course (sorry for suggesting the obvious).
no need to hup. it's been in the config since ice cream was invented.
> Were it working, you'd be getting messages like this logged in the mainlog:
> "Connection from a.b.c.d refused: too many connections from that IP address"
# grep "too many connections from that IP address" /var/spool/exim/log/main
#
and log is 14:48 long, rolls midnight gmt
> Not helping with the problem at hand, - but generally related to smtp
> flooding - we recently found changing DNSBL tests from "deny" to "drop"
> lowered the connection count significantly, due to a buggy botnet that
> failed to close connections after being told to "5xx off".
hmmm. may try that. thanks.
randy
