[nSLUG] FreeBSD 4.11 vs Redhat 5
D G Teed
donald.teed at gmail.com
Fri Mar 7 00:01:19 AST 2008
On 3/6/08, Stephen Gregory <nslug at kernelpanic.ca> wrote:
> On Thu, Mar 06, 2008 at 10:01:29PM -0400, D G Teed wrote:
> > Server A: Twin 3.2 Ghz Xeon with Redhat EL 5, postfix, amavisd and clamav
> > Server B: Twin 1.4 Ghz Xeon PIII with FreeBSD 4.11, postfix, amavisd and clamav
> > Server A took 45 minutes to process the active queue.
> > Server B took 20 minutes to process the active queue.
> When I see numbers that skewed I suspect that there is something wrong
> with one of the servers, or the test setup. The fact that FreeBSD 4.11
> does not have great SMP performance furthers my suspicion that
> something is not right. Historically the difference in performance
> between FreeBSD and Linux have been slight with each taking the lead
> under different tests.
> If these numbers are accurate you should call your RedHat support rep
> and ask why the RHEL server is so slow.
Ha! There is no way we could afford that or anything
similar. I've also heard that Redhat won't support any
non-Redhat packages, and amavis and friends are
not of Redhat packages.
The versions of spamassassin and amavis are slightly newer
on Redhat, and we might have something like SPF checks
in the mix which are not on FreeBSD, but I can't imagine this
doubling the processing time. Otherwise, the config settings
are the same, as much as possible with the different versions.
I could experiment with exactly duplicated SA plugins,
at the cost of allowing the spam scores to fall a little.
In that particluar set of batchs, I don't know what the content was.
There could have been some larger messages to scan on one side
than the other, but I'd think with a large random sample like
that it would be not too different.
More information about the nSLUG