Mark Pilgrim: My attempts at compartmentalization have failed. There is only one inbox.
Nearly three decades ago, I had an opportunity to witness the effect of an email sent to a select group being widely forwarded. I no longer remember the details, but I do remember the decision that I made at that time. Ironically, it is not a decision that I widely publicize as I do not condone breaking of netiquette. But it was that decision that leads me to open source, open standards, and blogging.
But this post isn’t about that. It is about a failure that has begun to affect me:
From xxxx@gmail.com Sun Mar 28 15:25:52 2010
Return-Path: <xxxx@gmail.com>
X-Spam-Flag: YES
X-Spam-Checker-Version: SpamAssassin 3.2.5 (2008-06-10) on vanadium.sabren.com
X-Spam-Level: *****
X-Spam-Status: Yes, score=5.8 required=5.0 tests=DNS_FROM_OPENWHOIS,
FH_DATE_PAST_20XX autolearn=disabled version=3.2.5
X-Spam-Report:
* 2.4 DNS_FROM_OPENWHOIS RBL: Envelope sender listed in bl.open-whois.org.
* 3.4 FH_DATE_PAST_20XX The date is grossly in the future.
I don’t understand DNS_FROM_OPENWHOIS RBL or why that would affect everything sent from gmail, but I do recognize FH_DATE_PAST_20XX as bug 6269. Apparently, when combined this means that I didn’t see anything in the past week or so from a number of sources, including not only gmail but yahoo and my publisher.
Update: I need to clarify that this does not affect the apache secretary email address.
Sam, the DNS_FROM_OPENWHOIS is https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6157
And I actually check (and clean) the spam folder in GMail several times a day, just in case.
Sam, the DNS_FROM_OPENWHOIS is https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6157
Thanks!
And I actually check (and clean) the spam folder in GMail several times a day, just in case.
I actually have a two layer system: SpamAssassin is run on the server to eliminate obvious spam, and then ThunderBird’s Bayesian filter is run on the client to process the rest. I do periodically check my Junk folder, but I had pretty much forgotten about the server checks as until recently they “just worked” and I had never found any non-SPAM in the emails that were filtered out by that process.
“I’ve also added a few whiltelist_from lines as failsafes.”
Are you using procmail or some other sorting/forwarding step before spamassassin? If so, I think it would be better to put whitelisting, blacklisting, etc. there, to avoid a call to the expensive spamassassin process. (And this, of course, applied after the filtering done by your smtp demon, which avoids calls to the expensive procmail process — a burst of a few thousand spam emails hitting your server more or less all at once can really slow things down if each one gets processed by spamassassin.)
I can’t imagine optimizing for the case where I get a few thousand emails more at less all at once originating from sites I whitelisted, but sure, it does make sense to skip running spamassassin entirely for sites I’ve whitelisted. For future reference, here’s the syntax: