On Thu, Mar 28, 2002 at 05:55:53PM +0100, Steffen Evers ([EMAIL PROTECTED]) wrote: > On Wed, Mar 27, 2002 at 23:33, Barry Mathieu wrote: > > I removed the /etc/fetchmailrc file, and issued > > > > /etc/init.d/fetchmail stop > > > > and then I logged in as root and issued > > > > dpkg-reconfigure fetchmail-common > > > > and I made sure a non system wide fetchmail setup was completed. > > > > Then I issued a, "pon" and after a few minutes my system connects > > (initiates ppp to my ISP) without me issuing any command; like a daemon. > > So my problem is still not fixed. > > > > I'm beginning to wonder about exim. Before the system attempts to > > connect, the following occurs in /var/log/syslog: > > > > Mar 27 22:38:01 debian /USR/SBIN/CRON[4209]: (mail) CMD ( if [ -x > > /usr/sbin/exim -a -f /etc/exim/exim.conf ]; then /usr/sbin/exim -q ; fi) > > > > > > I thought exim simply interacts with /var/spool/mail and did not attempt > > to create a connection. Maybe I'm wrong? > > Is your problem about an unintended dial on demand (DoD) or about > unintended retrievel of your emails? > > If it is about DoD than it is for sure this exim bug: > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=96633&repeatmerged=yes > > The current Debian woody is simply unusable for DoD with exim. I have > spend days to find out more about this bug. > > Thr problem is that EVERY time exim is started, it initiates a DNS > lookup of your local machine name. No matter how you configure it. This > is at least true when zou do not have a full qualified domain name for > your machine. > > If someone likes to look after it, I can provide all information I have > gatherd about this problem so far. The exim upstream developer gave me > some hints as well where the problem could be ... > > By, Steffen
Not very good news to hear. My problem is unintended DoD. I checked the backup of my old Potato configuration and there was no /etc/ppp/ip-up.d/fetchmail. So, I renamed it to something else in my new Woody setup (fetchmail_daemon). I also issued a "/etc/init.d/fetchmail stop". Now there is no way /etc/init.d/fetchmail should be able to run as a daemon when I start DoD by issuing pon. I have ppp configured for DoD; I don't use diald. Sure enough, when the following appeared in my log, the modem began a connection: Mar 29 08:38:02 debian /USR/SBIN/CRON[7292]: (mail) CMD ( if [ -x /usr/sbin/exim -a -f /etc/exim/exim.conf ]; then /usr/sbin/exim -q ; fi) Exim does this every 15 minutes. I definitely don't wan't my computer to obtain a connection every 15 minutes and stay connected for 5 minutes. As a newbie I really don't understand the full ramification of a DNS lookup. After reading the bug report, it appears that exim doesn't think it knows the name of the machine it is currently running on, and to search for, and assign, the name, it resorts to something that begins a modem connection. I simply can't have exim doing this every 15 minutes. I typically only collect mail twice per day, is there a way I could set exim to run using a timed cron job? That is, not run exim as a daemon. I'll start reading. My goal is to have DoD running so using a browser, or sending mail, can be a single step process. I can't believe such a terrible bug is not being more discussed in the mail group. Thanks for your help, Barry Mathieu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
