OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
Subject: Re: more woes...new installation help
From: Scott Smith (scottversity.com)
Date: Fri Mar 31 2000 - 18:08:17 CST

Is there any reason you're not using `timex'?

Just curious,
Scott

> At 11:17 AM -0500 2000/3/23, Wietse Venema wrote:
>
> > /usr/bin/time ./smtp-source -m1000 -s20 -l15360 -c
> >-tnulllocalhost localhost
>
> Running this same command on an ancient Sun Ultra Enterprise
> 1/140 server that I recently built with Solaris 7 and UFS Logging
> (data going to an ancient Segate ST32550W, log mirrored between a
> 64MB partition on this disk and a 64MB partition on the boot/root/OS
> disk, namely an IBM DCAS32160), I get 1:02.1 clock time to complete,
> 1.7 user and 3.8 system -- with stock out-of-the-box Sun syslogd
> (synchronous writes and all that crap).
>
> I would say that this machine would hopefully get the snot beat
> out of it by a much more recent (and much faster) Linux box, but 62.1
> seconds beats the 85 seconds it took under your preferred Linux
> configuration (synchronous directory writes, asynchronous syslog
> writes).
>
>
> Fascinating. Absolutely fascinating. Now I want to see how hard
> I can push this thing from another machine, one that hopefully will
> be able to blow it's doors off. ;-)
>
> Actually, I tried it from a much faster UltraSPARC IIi machine
> (270 Mhz), and it doesn't really go much faster. I guess postfix is
> just too smart for me, and will only accept mail so fast, regardless
> of what machine is trying to send it. Cool.
>
> Out of curiosity, I ran the same test again on that UltraSPARC
> IIi machine (which is the basic configuration we use on our outbound
> mail relays), and I got 1000 messages in 27.2 seconds, user time of
> 0.8, and system time of 1.9. I'm a pretty happy puppy. ;-)
>
>
>
> BTW, this Ultra 1/140 is going to be our new sacrificial ORBS
> lamb -- if customers are on the ORBS and can't get off immediately,
> they can point their mail to this machine and it will act as a relay
> for them.
>
> Of course, before we put it into production, I will have already
> contacted both the ORBS and MAPS RBL folks to make sure that no one
> else in the world accepts mail from it unless they ignore both of
> these lists, but then that's not my problem. ;-)
>
> Once it goes into production, I'll take our current sacrificial
> ORBS lamb out of production (the UltraSPARC IIi system I tested
> above), and do a slight re-work on it so that we now have three
> outbound mail relays, each capable of handling ~35 messages per
> second (under ideal conditions, etc...).
>
> --
> These are my opinions -- not to be taken as official Skynet policy
> ======================================================================
> Brad Knowles, <blkskynet.be> || Belgacom Skynet SA/NV
> Systems Architect, Mail/News/FTP/Proxy Admin || Rue Colonel Bourg, 124
> Phone/Fax: +32-2-706.13.11/12.49 || B-1140 Brussels
> http://www.skynet.be || Belgium
>
>