[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: sendmail is slow for mass mail

On Tue, 3 Apr 2001, Craig Sanders wrote:

> On Tue, Apr 03, 2001 at 07:11:06AM -0400, Chris Wagner wrote:
> > So, what happened to sendmail?  How did it earn it's fall from grace?
> > When I got into it, sendmail was it.  I've never looked closely at the
> > mail system since.
> * it's slow.
> * it doesn't scale well.

Short: FUD, spread by one who's not kept up with current developements -
Look at whats being done in 8.12 wrt multi-queues and
multi-runners/queue.  8.12 is faster than 8.9.3 ever was (much faster
than 8.10/11), and has a hell of lot more function

Long: Sendmail was basically in `maintenance' mode for several years,
and a few of its competitors (who didn't have legacy concerns) were able
to leapfrog it in performance and scalability.  Sendmail has had quite
a bit of rework...

> * it suffers from security-flaw-of-the-month-itis.
> sendmail was not designed with security in mind (the net was a much
> nicer place back then), and it shows.

No, but it has come a long way since then, a significant portion has
been rewritten since 8.9.3  The last incident I can recall turned out
to actually be the kernel capability bug...

> * it's configuration language is overly complex for the task at hand
> (the m4 macros helped a lot, but it's still way more complex than it
> needs to be)

Depends upon the task at hand eh?  For a end-node, sure sendmail.cf
hacking isn't needed - but the provided m4 features cover that pretty
well with FEATURE(nullclient, `<smart_host>') -- what could be easier.

Sendmail *is* the kitchen sink of MTAs - and yes, there is a cost to
that, but significant tuning *is* going on...  The other side is that
with the kitchen sink comes everything:

Mixing/matching of DB, LDAP, text, HESIOD, PH (not on Debian), etc.
for *ANY* map: aliases, access, etc.

Why suffer with `sendmail compatible' when you can have the `REAL

Rick Nelson
"And the next time you consider complaining that running Lucid Emacs
19.05 via NFS from a remote Linux machine in Paraguay doesn't seem to
get the background colors right, you'll know who to thank."
(By Matt Welsh)

Reply to: