Hi,
The dspam setup we have running at the moment is fairly basic. pgsql
storage driver, integrated with postfix through check_client_access
and a pipe transport, and alises/transports for spam@domain and
notspam@domain. Everyone is part of a global classification group
that feeds off a well-trained globaluser, and the training mode is
train-on-error. But the strange thing we're seeing is that when a
message comes into, say, spam@domain.com for retraining as spam, by
the time it gets to the dspam-retrain script the sender address has
turned from what it should be, into the value of ClientIdent in the
dspam.conf file, be that value secret@Relay1, or
gobbledygook@wtf.com. Has anyone come across this problem before?
I can provide whatever configuration information you might want, but
I'm stumped, honestly. I can't figure out where it's going wrong.
The secret@Relay1 is telling me that it's going in through the dspam
filter before it's being transported out to the dspam-retrain script.
I'm baffled, though. Any ideas?
Regards,
Adrian Overbury
Received on Thu Nov 8 03:52:48 2007
This archive was generated by hypermail 2.1.8 : Fri Nov 09 2007 - 00:00:09 CET