Rinne,
I use Russell Coker's postal program for SMTP performance testing. Are you
just going to be relaying?
I am finding that using JDBC for a spool is higher performance than the file
system, but please test it in your environment. Also, if you are going to
do a backup, one benefit of using JDBC (at some expense of resources) is
that you can configure MySQL to do real-time replication. I use that with
our system such that in the event of a failure of the live server, I have a
real-time backup.
--- Noel
-----Original Message-----
From: Rinne Ville [mailto:[EMAIL PROTECTED]]
Sent: Monday, January 13, 2003 3:39
To: 'James Users List'
Subject: James performance tips
Hi,
We are in the testing phase of a project where we are using James as an smtp
proxy. Once we enter production we are expecting to proxy around 6 million
message per month initially. We will be doing performance testing fairly
soon and I was just wondering if any of you have some basic performance tips
relating to things such as how many delivery threads should we use etc. ? We
did some performance testing before the actual project in a short proof of
concept phase with pretty much the default configuration of James and based
on the results we got from those tests I'm not expecting any trouble but I
thought I'd ask anyways to avoid making any obvious mistakes. We'll be
deploying on a single cpu p4 Linux-box with an identical machine used as a
backup.
cheers,
Ville Rinne
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
