Before you _deliberately_ cc: me again, please take the time to note the
simple request in my .sig, and that the Reply-To: is set that way for a reason.
>Uhm, the processes are owned by whom (that might give you a clue)?
One user, out of over a dozen who use procmail to handle their mail.
Well, there's a little clue. That one user is obviously doing something
improperly. Disable their .forward until it is resolved. That resolves
YOUR problem with the process count.
17; but only one has several hundred procmail processes running
There is no indication that the procmailrc contains a loop or anything
peculiar
An experienced admin might choose to add a verbose log setting to the
~/.procmailrc file in question and find out which recipe is doing something
stupid. An admin not being paid to do development work for their users
might just disable the .forward and then send a message to the user
advising them that their account has a problem.
>Is mail not being delivered normally - globally, or just to certain users?
Mail appears to be being delivered normally. But... how would we know? :-)
Contents of mailboxes being updated, perhaps for all users except for the
one who is apparently generating a mail loop is a good indication.
>What's in your mailqueue?
There are 2464 files in /var/spool/mqueue... suggestions???
I dunno, check for LARGE ones that look like replies to replies to replies
to replies, or at least grep for messages created by or destined for the
affected user.
>I could go on and on, but I think _some_ sleuthing on your part would be
>appropriate.
How do you know we didn't?
The same omnipotence that allows us to answer questions when not provided
with even the most basic of details about the problem. I can't begin to
describe how much of a waste of time it is to have to list all manner of
possibilities (including a great number which involve other processes on
your system), when the person asking the question hasn't indicated which
simple things they've already investigated and at least focus the problem
on some particular factor such as a user or script. None of us have root
on your server (well, not that we're going to openly admit anyway -
omnipotence has it's advantages).
We're trying to figure out where to sleuth and how to interpret the data.
[snip]
User "A" (and only user "A") seems to be creeping up in procmail process
count over time.
There you go. If you must, sleuth his procmailrc. Contact the user (say,
by telephone, as they have an account, and perhaps you, as a sysadmin, have
those details) and advise them that their procmailrc is broken. Perhaps
they are relying upon the output of an external program which changed and
now causes their script to do something it wasn't prepared for.
At this point, anything offered here is pure conjecture because we don't
see the script in question. Not that I'm suggesting that you post your
user's script - perhaps it should be *THEIR* job to find the solution -
your solution is to disable their .forward until they resolve it.
---
Sean B. Straw / Professional Software Engineering
Procmail disclaimer: <http://www.professional.org/procmail/disclaimer.html>
Please DO NOT carbon me on list replies. I'll get my copy from the list.
_______________________________________________
procmail mailing list
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail