Jeff Mayzurk wrote:
>
> >On Tue, Oct 31, 2000 at 12:02:55AM +0100, Peter van Dijk wrote:
> >>DESCRIPTION
> >> qmail-queue reads a mail message from descriptor 0. It
> >> then reads envelope information from descriptor 1. It
> >> places the message into the outgoing queue for future
> >> delivery by qmail-send.
> >
> >Yeah, I read through that and while it implies that you terminate the
> >session by doing a close, it doesn't say that. It could, for example,
> >terminate it by a line consisting only of '.'...
> >
> >I'd change it to read:
> >
> >qmail-queue reads a message from descriptor 0, terminated by a close().
> >[...]
> >
>
> It also doesn't state the requirement that these operations must be
> serialized. Yes, the implication is there, but it's hardly what I'd call
> clear.
>
> Beter yet:
>
> "qmail-queue reads a message from descriptor 0. After EOF is received
> on descriptor 0, it reads envelope information from descriptor 1."
>
> By the way, does anyone have any interest in comparing notes on really high
> volume qmail configs? I'm looking for performance in the range of 200-250k
> remote deliveries per hour. We're halfway there with relatively few
> modifications on a modest dual-processor Sun. Before we really start digging
> into the code (or throwing more hardware at it), I'd like to hear from others
> who are dealing with similar volumes.
I've a similar issue (high performance required).
We deceided to go for multiple smallish intel boxes due to cost for
emails per hour ratios work in favour of intel on Linux (or FreeBSD but
thats a different flame ;-).
Lirking on this list I've seen people achieve between 30k to 150k and
hour. The average appears to be 50k an hour.
If bandwidth is not an issue then next bottle neck is the queue and disk
speed. As far as a I can see once you've moved from a simple DISK, you
can either:
a) Avoide the queue and talk direct to qmail-remote (and take care of
bounces / process mgt yourself).
This appears to be where people are getting above 100k and hour.
b) Buy lots of fast disks - RAID 0+1 (due to write performance RAID 5
appears a bad idea).
On SUN this gets expensive !
c) Some suggest RAM disks / Flash DISKS.
Is it Russ or Ken Jones has used Flash disks for this ?!? can't
rememeber.
If anyone has any hard info / experience here. I would be interested.
If you've any nice patches I would be very interested - but my spin on
the issue is KISS - i.e we will scale by adding small nodes - not
expensive boxen, and also going with simple set-ups for the sanity of
the people whom will run this.
I would be very instered in anyone thoughts on theses issues.
Regards
Greg Cope
>
> Thanks,
>
> -Jeff
