Xiaoyan Ma <xma(_at_)haas(_dot_)berkeley(_dot_)edu> writes:
We had a user who misconfigured his .forward and procmailrc file, which
generated undeliverable mail to loop and filled our /var/mail spool. I
wonder if there is a way to prevent this type of undetected loop.
If a loop is undetected, then how do you know it's happening?
What you mean to ask is "is it possible to automatically detect loops
of this type?" Since breaking a mail loop means either bouncing or
throwing out the looping message(s), you have to decide which ones
should be considered "looping". For example, qmail and postfix add a
Delivered-To: header to each message they deliver and will bounce any
message that already contains a Delivered-To: header with the specified
address. Sendmail will bounce messages when the number of Received:
headers goes over some specified limit. How do *you* want to detect
looping messages?
The other way to deal with the problem is to places disk quotas on the
mailspool, such that when a user uses too much disk space, for whatever
reason, their mail will simply bounce until they move or delete some of
their mail. That all solves some general denial of service problems,
though it creates others (e.g., michael sends a bunch of large messages
to ann, causing other messages to her to bounce).
Write a policy, think *hard* about its consequences, then implement it.
Philip Guenther
_______________________________________________
procmail mailing list
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail