Thomas Hochstein napisał(a): >> Intermittently I get errors logged like in subject while mail is locally >> spam-scanned (traditional method of external scanning and the mail >> filter re-injecting the message with exim -oMr spam-scanned -bS). >> > > Why don't you do it during SMTP-time? >
Because that's the method I inherited (it's default in our hosting management package) and it works fine with custom script. For all the talk about scanning in ACLs I haven't found any example on the net how to use custom script in ACL during SMTP time, as opposed to simply connecting SpamAssassin, which I don't like and use statistical filter (DSPAM) before it plus other tricks in my custom mail filter under development (like storing authenticated senders and their recipients in SQL DB and using them to let spam filters know mail from those addresses is clean, obviously the goal is reducing false positives). I admit I would prefer this solution - I could do reject or fakereject (and quarantine), I just don't know ACLs yet well enough to implement this. I can't just throw not very well tested solution into production hosts. > >> Does anybody know the cause behind it? >> > > I bet the scanning process takes too long. > This is really unlikely; the mail in question (that was lost) had a few KB size, and mail delivery worked just fine before and after this happened, including the domain of that mail. I have: smtp_receive_timeout = 5m The loadavg (which I record in cron job) showed the machine was quiet at the time. Since this is "Exim to Exim" injection, mail filtering on the way was not a problem, filtering was already done and the timeout happened when transport with: command = /usr/sbin/exim -oMr spam-scanned -bS ..tried to run Exim locally. -- Marcin Król -- ## List details at http://lists.exim.org/mailman/listinfo/exim-users ## Exim details at http://www.exim.org/ ## Please use the Wiki with this list - http://wiki.exim.org/