On Jan 20, 2004, at 11:56 AM, Jesse Norell wrote:
That's probably the local_recipient_maps setting. You can just
disable it
I don't currently have it enabled and I disabled the errors on unknown
users. How can I set postfix to use the database when checking for
valid users?
This is what we have, with postfix+postgres:
# grep local_recipient_maps /etc/postfix/main.cf
local_recipient_maps = proxy:pgsql:/etc/postfix/aliases.cf
# cat /etc/postfix/aliases.cf
user = postfix
password = <removed>
dbname = dbmail
hosts = forever.kci.net
table = aliases
select_field = deliver_to
where_field = alias
Note, this is the exact same lookup file we use for postfix
aliases lookups in the database.
Thanks. I added that to the configuration but I notice this in the
output logs:
Jan 20 12:07:40 uptech postfix/proxymap[7677]: fatal: unsupported
dictionary type: pgsql
Jan 20 12:07:41 uptech postfix/smtpd[7666]: warning: premature
end-of-input on private/proxymap socket while reading input attribute
name
Jan 20 12:07:41 uptech postfix/smtpd[7666]: warning: private/proxymap
socket: service dict_proxy_open: Success
Jan 20 12:07:41 uptech postfix/master[7658]: warning: process
/usr/lib/postfix/proxymap pid 7677 exit status 1
Jan 20 12:07:41 uptech postfix/master[7658]: warning:
/usr/lib/postfix/proxymap: bad command startup -- throttling
Offhand, I don't remember anything in your config that called
spamassasin .. your content_filter pointed to a virus scanner, didn't
it? That's how our setup works - we have amavis run from the
content_filter line, and it calls spamassasin itsself... your's is
a different setup if you were using a master.cf entry to call it.
Perhaps you used to have a transport map entry pointing to spamassasin,
and now have it pointing to dbmail?
Looking at it deeper it is amavis that triggers spamassasin.
Spamassasin is working and outputing a spam level header but it isn't
going into the database for some reason. It shows up while the header
breakdown cycle occurs in the code though (in the log at least it
recognizes it).
