Re: optimization of fetchmail
On Wed, Apr 10, 2002 at 09:05:14AM -0700, Craig Dickson wrote:
> I doubt it makes any difference; I don't think the pop3 protocol
> provides for switching from one user to another without establishing a
> new connection. I would tend to prefer the second form (two users under
> a single poll entry in .fetchmailrc) just because it's less typing and
> easier to read (it's more clear that these two users are on the same
> server).
I think I'll choose the second form, 150 users under 3-4 differerent
pop servers may give me headhache douring the server maintaining if I'll use
the first mehod :-)
> The obvious test would be to install tcpspy and see what actually happens
> when fetchmail runs. What I've observed is that a new connection is made
> for each user entry.
Oh a good way to refresh my tcp cognitions!
> The main thing that would make a difference if you have fetchmailrc
> supporting many users is to run a single system-wide fetchmail daemon
> rather than running it from cron jobs.
I use fetchmail as start-stop-daemon because there is an intensive traffic
and I feared that cron would have taken away too much resources and also
because the fetchmail daemon script has the beautiful option "awaken" very
useful in phase of test.
Talkin' about daemons, the idea of bringing fetchmail under djb daemontools
sounds interesting, does anyone have heard something about it?
--
To UNSUBSCRIBE, email to debian-user-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: