koffiejunkie wrote:
Nate Duehr wrote:
Qmail is fast, and can handle an incredible amount of mail thrown at it,
I have heard and read that claim so many times but, after years of
having to admin qmail servers, have yet to seen it handle huge amounts
of mail with even half the grace that Postfix does.
I have no experience with postfix on massive amounts of mail.
I do believe you, however... since HP is using postfix internally quite
heavily, from what I've heard. (I don't work there.)
The postfix machines I've run are much smaller than the qmail box(es)
were and less important.
I'm lazy and my home server is still running exim, because I've been
running it since exim3 days on Debian... and I haven't felt like
rebuilding it to postfix. But if the hardware ever finally keels over
dead, it'll be rebuilt as postfix/courier.
I regularly
encounter servers that had been compromised (usually via php or a weak
smtp password) and used for sending out massess of spam. 100,000
undeliverable mails in the queue and qmail just about stops functioning.
That's odd. Someone who would go through the time/effort to set up
qmail didn't secure their box? Weird.
Never seen a queue quite that high, but I would assume the box would get
both CPU and I/O bound for most values of "box". (GRIN)
Add in that even if it's "Public Domain", the author never wnated to
work with the community to make it better... he just washed his hands
of it
You're letting him off lightly. He still maintains it is perfect,
doesn't need any of the new features, and is 100% secure. Forgive me
for thinking doesn't have to deal with any real busy production servers.
Yeah, I was being diplomatic. Qmail's author is out of touch with large
mail server admin reality so far that I figured it went without
saying... anyone who really looks into it will find the same things
you've just added above.
Anyone too lazy to look into it, gets what they deserve when they have
to deal with constantly recompiling qmail for their various production
machines.
I gave enough of a hint that they could go do their own
research/homework. :-)
Nate
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
