On Sat, 21 Nov 2009, Charles Gregory wrote: > > Did you look at the logs you posted? > NONE of the DNS tests are being launched on msg 26661....
Yes, that is the problem. They run with `spamassassin`, but they do not run from `spamd`. Do other people see this running `spamd --debug`, although it works with `spamassassin`? > Also, for that message, there are a suspicious set of > entries that do not appear in the second msg but appear in > the first: > > Fri Nov 20 11:49:31 2009 [26661] dbg: config: fixed relative path: > /var/lib/spamassassin/3.002005/updates_spamassassin_org/20_dnsbl_tests.cf > > Fri Nov 20 11:49:31 2009 [26661] dbg: config: using > "/var/lib/spamassassin/3.002005/updates_spamassassin_org/20_dnsbl_tests.cf" > for included file > > Fri Nov 20 11:49:31 2009 [26661] dbg: config: read file > /var/lib/spamassassin/3.002005/updates_spamassassin_org/20_dnsbl_tests.cf > > ..... so spamassassin definitely thinks something is > different about the enviornment, and tries to fix things. > Why is it a 'relative path', and does this interfere with > any other path-based functions? `sa-update` creates its cf with relative paths, which is correct. When I ran the test Benny recommended using `spamassassin` but with full '-D' debugging instead of just '-D metadata', I see all the same messages about fixing relative paths, and I still get the same results: `spamassassin` runs RBL tests, but running through spamd does not. Both logs show that all the cf files were read and parsed correctly. I don't think that's the problem. $ head /var/lib/spamassassin/3.002005/updates_spamassassin_org.cf # UPDATE version 795855 include updates_spamassassin_org/10_default_prefs.cf include updates_spamassassin_org/20_advance_fee.cf include updates_spamassassin_org/20_body_tests.cf include updates_spamassassin_org/20_compensate.cf include updates_spamassassin_org/20_dnsbl_tests.cf include updates_spamassassin_org/20_drugs.cf include updates_spamassassin_org/20_dynrdns.cf include updates_spamassassin_org/20_fake_helo_tests.cf include updates_spamassassin_org/20_head_tests.cf ... Thanks for the ideas though... I will try comparing the full debugging output of running `spamassassin -D` versus the startup of `spamd --debug` and see where things are different. Then I guess I'll have to start hacking debug messages into the confounded source. --mark--
