[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: milter errors.
Re: milter errors.
Mon, 09 May 2005 16:57:47 -0500
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.6) Gecko/20050328 Fedora/1.7.6-1.2.5
I have found that this error usually results from spamd failing. I run
spamd on a different system than the servers which connect via the
milter. We have multiple mail servers all querying the same spamd server.
Quick numbers on one of our inbox servers indicate 2,300 failures
("Milter (spamassassin): to error state") over the course of 84,000
I have a check that runs every minute on each mail server, and on the
spamd server to be sure the milter is available and that spamd server is
I don't know if resources are your problem. But, I would suggest
checking and logging failures of spamd as a first step.
To check if the mitler is running, just test for the existence of the
socket you use. Here's the script that I use for spamd. It requires
netcat (nc). Lastly, the biggest bottleneck seems to be DNS tests for
Sorry for the digression from the milter aspect of the question, but I
think that the problem could be in spamd's response. This is a decent
way to test if that is the case.
if [ -f $LOCKFILE ]
echo "Exiting. Lock file $LOCKFILE exists."
net_test="`echo quit | nc -w 45 localhost 783 | head -1|cut -f1
| cut -f1 -d"/"`"
if [ "$net_test" != "SPAMD" ]
echo `date` spamd-check failed >> /var/log/spamd.log
echo `date` spamd-check ok >> /var/log/spamd.log
rm -f $LOCKFILE
Miles Mawyer wrote:
> I keep getting numoreous errors related to the milter and I'm trying to
> figure out why, and/or what they mean. Below are some excerpts of the log
> files. Message seem to be skipped a lot as well, I'll see that they passed
> through the milter server however they have no spamassassion headers?
> What do I need to do to figure out whats going on.
> It's a dual xeon 2.4g with 3gigs of ram. I'm running 10 spamd children and
> ending up with about 300ms left of memory on average. Seems kinda steep. Any
- milter errors., Miles Mawyer, 2005/05/09
- Re: milter errors.,
Chris Crowley <=