Theo Van Dinter wrote:
On Sat, Jan 27, 2007 at 06:52:29PM -0500, Thomas Bolioli wrote:
/etc/procmail and it is fired off with a user .forward file "|IFS=' ' &&
exec /usr/bin/procmail || exit 75 #tpblists". Still looking into Net::DNS.
A few ideas. First, do DROPPRIVS=yes if you haven't already. Second, why are
you using a .forward file? Just set procmail as the MDA.
DROPPRIVS is already set to yes. In answer to the second question,
legacy. This machine has an upgrade legacy of 6 years. I set it up this
way because I was not having SA do checks for every account and I was
experimenting when I first setup spam filtering. Changing that may
become my sunday morning task...
I am still at a complete loss to explain why some users (when running SA
from the cmdline) can do rbl checks and others can't. I have set the
user_prefs files to be exactly the same, eliminating any config deltas
from potentially causing this. I have confirmed though that the problem
is that the DNS queries are definitely timing out and upping the timeout
to 60 secs does nothing but delay the inevitable. I was mistaken that it
was the SPF* tests zeroed out that was causing the issue. But it looked
that way for a while.
Now, the only thing clustering the groups (ie; those that work, and
those that do not) is the two accounts (there may be more but I will not
be digging into my clients email accounts) that do not successfully
check RBLs get by far the most amount of spam compared to the others
that work.
Anyone with ideas, they would be greatly appreciated but right now I
need to determine if it is SA that is having issues with the lookups or
are the accounts screwed up in some way. bind does not seem to be
throttled either so the volume of queries should not be the issue either.