rebooting and crashes on dell server

[BSD Guy]

Theres alot to this problem, but I'll try to be
concise so I don't loose people.  I started out with a
cheap dell server running scsi raid and 5.x  It
crashed or locked up with a kernel panic in a
different process (everything, syslog, you name it)
ever few days.  I swapped ram around figuring it was
at fault since it was bought used on ebay.  Still no
luck.

Tired of messing with it, I bought a new dell
poweredge 2650, with scsi raid, a couple gigs of ram,
duals, and just a new system, configured it, copied
user data over and switched to it.  Sure enough still
the worst stability I've ever seen.  It has dual power
supplies, each in a different UPS. I don't believe
power is the problem or I'd see similar problems on
the other server or router I have plugged in there.  I
even upgraded to 6.0-Release-p1 but no luck.  

It panics from time to time, but usually now it just
randomly reboots.  Often at least every 36 hours. 
Often 2-3 times a few minutes apart when it does.  I
did install a debug kernel 2 weeks ago to try to get a
crash dump, but no panic's since then.

My make.conf is pretty simple:
KERNCONF=crapbox
CPUTYPE?=p4
CFLAGS= -O -pipe
NO_BLUETOOTH=   true    # do not build Bluetooth
related stuff
NO_I4B= true    # do not build isdn4bsd package
NO_LPR= true    # do not build lpr and related
programs
NO_GAMES=        true    # do not build games (games/
subdir)
MAKE_IDEA=     YES     # IDEA (128 bit symmetric
encryption)
# added by use.perl 2005-09-29 11:26:41
PERL_VER=5.8.7
PERL_VERSION=5.8.7

Its not all that stressed of a box:
load averages: 0.14, 0.07, 0.06
Mem: 235M Active, 367M Inact, 160M Wired, 33M Cache,
99M Buf, 81M Free
Swap: 1024M Total, 1024M Free

I've tweaked all the kernel values mentioned in the
tuneable section of the manual, but the reboots
continue.  The box runs radiator (a perl based radius
server), apache+mod_ssl-1.3.x with php5, postfix-2.2,
amavisd-new-2.3, courier-imap-4.0, pure-ftp, nrpe
(nagios), net-snmp, ssh, and bind 9.3.1 all from ports
(except radiator).  

I run the same mail server setup on another server
with no problem, although it processes far less mail. 
I use radiator on another server with no problem. 
This is a unique combination of packages I'm running,
but no unique programs or configurations to this
server. Between amavis and radiator it does alot of
perl, and everything is a mysql lookup to another
server.  While all logic and experience points to a
hardware problem I feel like the complete hardware
replacement has pretty much laid the blame on
software.  

I'm open to any suggestion as to what I might need to
look at to clear this up. I'm about to have a nervous
breakdown and insteall debian in desperation heh. I've
tried a kernel without smp, I've tried disabling acpi
on boot.  If you need any more information feel free
to ask for it, I'll happily provide it.
Thanks!  

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com