Question about whitelist, greylist and dnsrbl
2013-10-13 by Jim Klimov
Hello all, Currently my MTAs call various filtering routines in such an order that DNS RBL lookup is performed by the MTA, and hosts which are not instantly blacklisted, go on to more complicated filters - such as milter-greylist. Sometimes this is a bad thing, i.e. hosts that are configured as explicitly trusted (static whitelist) in the milter are shot down due to DNS RBL before the milter is even invoked. On the other hand, auto-whited sources (past greylist timeout) which might indeed be spammers, should be shot down by DNS RBL even if they wiggled through our filters on their earlier attempts. At what moment are auto-whited hosts accepted - at the start of rule processing, or at the appropriate "greylist" ACL which finds that the timeout reached zero, or at some other moment? Namely, I wonder if I can utilize the "dnsrbl" rules within the milter-greylist in such a way that static whitelisted ip-addresses, domains, RCPTs and FROMs are honoured and always accepted, but the auto-whited hosts might still be rejected due to their newly found presence in DNS RBLs. Is this as simple as defining racl's in the order of "whitelist, dnsrbl, greylist", or are there some more complicated considerations? Thanks, //Jim Klimov