New speed benchmark

[Glenn Steen]

On 04/02/06, Res <res at ausics.net> wrote:
> On Sat, 4 Feb 2006, shuttlebox wrote:
>
> > Is that with SA or just virus checking? I find that SA with all its
> > network checks adds a lot more time than the virus scan.
>
> With SA it loads to 4 but it gave us too many false alarms so we disabled
> it, even when we ran it we disabled most checks, the only thing  it did
> was indicate spam content, we use qmail and sendmail to test RBL's etc, no
> point in accepting the full msg passing it to MS to reject/drop when we
> can reject on header only at MTA

That in the greater part explains the difference in load avgs. Not
that I know exactly what network tests Jules ran in this case, but
your low figures are simply due to you not doing 1) SA, and 2) SAs BL
lookups. As mentioned, these two tend to add some "real" load and (in
the latter case) significant "unreal" load;-).

> But qmailscan has a bad habbit of not being able to handle alot of stuff
> gracefully, which is why I was after a clear cut guide on how to install
> MS on a qmail system,  because until the sendmail consortium can produce a
> copy of sendmail that works identical to qmail in relation to like with
> vpopmail for virtuals there is no beating that combination, be it for
> visp's or hosting.

Did someone mention postfix ....:-):-)

--
-- Glenn
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se