Howdy,

Well, from looking @ the patch I can't see much delay:
- The code is bigger and a little more slower, if it is 0.001 of a
second slower(in the CPUs nowadays), in a system with 1000 msgs/seg it
will have a 3 seconds (store, local, remote) delay.... so why bother.
- What this patch does is allow for a bigger qmail queue, which IMHO is
the best.

I don't see what's the point in this discussion because qmail is not CPU
intensive, but a I/O intensive... soo loosing seconds in messages that
take sometimes 1 hour of delay is somewhat of a ">/dev/null", IMHO.

My Euro,
   Rui Lapa

PS: Not exact values, and not considering external conditions

On Fri, 2001-09-28 at 06:52, Henning Brauer wrote:
> [Franky, where did your linewrapping get lost?]
> 
> On Thu, Sep 27, 2001 at 10:05:25PM +0200, Franky Van Liedekerke wrote:
> > I experienced the same problem, and no, the big-todo patch won't help you 
> > a bit. 
> 
> In the situation the previous poster described it whould help.
> We won't know it for sure until he tries it ;-))
> 
> > What I did, and what I recommend to everyone, is that you buy a SCSI 
> > controller that has at least 64MB NVRAM cache on it (and preferably 
> > 128MB) with battery-backup. This results in mails being processed *much* 
> > quicker,
> 
> Another possibility is using a solid state disk. This thingy needs a battery
> backup too, of course. Usually your queue disk should not need to be too
> big, a ssd is an option here. 
> 
> Greetz
> 
> Henning
> 
> -- 
> * Henning Brauer, [EMAIL PROTECTED], http://www.bsws.de    *
> * BS Web Services, Roedingsmarkt 14, 20459 Hamburg, Germany *
> Unix is very simple, but it takes a genius to understand the simplicity.
> (Dennis Ritchie)
> 
-- 
---------------------------------------------------------------
Rui Lapa                                     DeSCT-OniSolutions
[EMAIL PROTECTED]
Fingerprint: 4C8F 2593 6813 55F5 8FA1  74B7 245F 9138 1C02 9331
---------------------------------------------------------------

Reply via email to