Hello!
This is my first post to this list, so please be gentle!!
I've got a small home network where I'm running an IMAP mail
server for family mail, fetching mail from multiple POP accounts
at various places on the net.
When using Fetchmail 6.2.2 on Redhat8 trying to fetch mail from
my Earthlink account (NGPopper), On at least a daily basis, I
get an e-mail (always SPAM) that causes fetchmail to fail
fetching the mail and repeated attempts only continue to
retrieve the same e-mail, causing my local mail file to get quite
large and for the Earthlink account to never get cleared unless
I go in manually and delete the mail. This has happened ever since
I started using fetchmail, but recently has gotten so bad that I've
begun looking for either a solution or an alternative.
I'm using a fetchmail/sendmail/procmail/SpamAssassin combination
and I've isolated the problem somewhat, so here goes.....
It appears, when watching the logfile, that there is a long pause between
sending the 'MAIL FROM' to sendmail and geting the 'sender ok' response.
I *do* get the 'sender ok' response, but only after 30-45 seconds.
Ok, so sendmail accepts the mail, but the next operation on the POP
socket ends in failure. So, the mail never gets cleared from POP and the
next time through the same thing will happen all over again.
So, it seems that it takes sendmail some time to process certain e-mail
addresses and the POP server has given up on me!!! I've tried having fetchmail
deliver right to procmail (I'm only fetching for a small group of local mail
boxes), however the amount of time it takes for Procmail/SpamAssassin/Razor
can end up causing the same problem.
As a test, I grabbed the source to fetchmail and forced it to always use
my own address in the smtp 'MAIL FROM' and now all of the trouble e-mails get
fetched zippity quick!!! Interestingly enough, all of the 'From Addresses' in
Outlook Express still have the original 'From', but I guess that comes
from the headers somewhere (I don't pretend to be an expert here).
I'm tempted to leave it this way, but I'm troubled that I may be creating
problems
that I'll have to clean up later. Ideally, I could ping the POP3 socket with
some
kind of NO-OP while waiting on a lengthy SMTP transaction, but I have yet to
attempt that. I wanted to get an expert opinion from this group first.
Is there a better way?
Brian Pence
Author - AbsoluteTelnet Telnet/SSH client
http://www.celestialsoftware.net
http://www.absolutetelnet.com