After wiping and reloading everything (rebuilding SA, wiping user_prefs and local.cf), the issue with TRUSTED_NETWORKS seems to be resolved.
I've had a couple of interesting problems. I'd like in some cases to see the values tagged to the tests in the full report. I had some success with this by running "spamc -R" but it was inconsistant. Has anyone run SA with full reporting? Had any problems?
Likewise, I added "--username" to spamd and had some difficulties. It worked at first until I started playing with switches ;)
Can I get away with any old account? what kind of access will
it need? eg can it have /bin/false as its shell?)
While playing with the above two variables, I was running two iterations
of spamd; one (2.64) on the default port and the newer SA on somes
other random port. The specific results were that my mail wasn't going thru at all. At one point, pine was complaining about
a corrupt mailbox format (It appears that spam was only adding the
full report to the mbx without headers and the body). And syslog was getting a stream of connections refused ie
connect(AF_INET) to spamd at 127.0.0.1 failed, retrying (#1 of 3): Connection refused
I know I'm not giving much information here; it's likely some silly thing I was doing (or doing too many things at once).
A couple of interesting points/questions; to resolve "bayes db version" error, I ran sa-learn --sync. Will all the users have to do this?
Likewise, spamassassin --lint showed various errrors in my SA2.64 user_prefs ie rules that no longer work in SA3.0X until I removed them. Will all users have to do this also?
This may be another philosophical issue in that we previously gave users too much control over their SA environment and perhaps the way to go is to indeed have switches, bayes_db_path etc defined only globally.
=-=-=-=-=-=-=-=-=-=- generated by /dev/dave -=-=-=-=-=-=-=-=-=-=-=-= David Stern University of Maryland Institute for Advanced Computer Studies
