My company recently got a working install of RT and Postfix up and it
was running fine without fetchmail, just that everyone had to log into
the RT portal. We decided this last monday to install fetchmail on the
CentOS box to make it easier for the employees to create tickets and
comment on them too. However, after five days of research and
problems, I am admitting defeat to fetchmail. I could really use some
help with the configuration, any help, because I am helplessly lost. I
just reinstalled fetchmail, so it needs the fetchmailrc file (which is
the root of my problem) and if someone could help me configure that
file, and the rest of it too, it would be extremely hellpful!
It’s actually quite simple. We use a very basic configuration here:
All mail destined for RT goes to an IMAP inbox somewhere (different
server than RT here).
We created a user “rtimap” on our linux box running RT.
poll servername.domain.somewhere proto IMAP user “username” with pass
"password" fetchall idle
The magic is done in /etc/aliases (could be configured by maildrop,
procmail and others, too, I believe.)
All mails are delivered directly to the utility rt-mailgate, which is
part of the rt installation (afair).
Just set the queue they should go to and the right url.
We encountered problems with https here, so you may want to try http, at
rtimap: “|/usr/bin/rt-mailgate --queue ‘Incoming’ --action correspond
(don’t forget to run “newaliases”)
We try to start fetchmail as cronjob every 10 minutes:
*/10 * * * * /usr/bin/fetchmail -s 2>/dev/null >/dev/null
But fetchmail will be kept logged in to the mailbox (keyword “idle” in
.fetchmailrc), and react instantly to new mails, so this is just in case
the connection vanishes to reestablish it.
I highly recommend the >/dev/null 2>/dev/null part, otherwise mails
generated from cron will go to rtimap, which will create a ticket via
aliases for each mail.