- Previous message: Rob Tanner: "Postfix-2.3-20051229/Cyrus SASL & smtpd: bad command startup -- throttling"
- Maybe in reply to: Kev: "performance"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
|
|
|
From: Noel Jones (no email)
Date: Tue Jan 03 2006 - 12:23:15 EST
At 01:07 AM 1/3/2006, Kev wrote:
>On Tue, 3 Jan 2006 07:41:19 +0100
>Magnus Bäck <> wrote:
>
> > On Tuesday, January 03, 2006 at 06:05 CET,
> > Kev <> wrote:
> >
> > >
> http://www.sa-blacklist.stearns.org/sa-blacklist/sa-blacklist.current
> >
> > (File with 341k lines.)
> >
> > > i was thinking about using that list with
> my sender_access list,
> > > how much performance will i be loosing with this ?
> >
> > Probably very little, perhaps not even a
> measurable loss, but why don't you
> > measure it yourself? Build an indexed
> database using different map types
> > (hash, btree, cdb etc) and measure the lookup
> time with "postmap -q".
> > Compare with measurements of lookups against
> a very small database so you
> > know the constant-time overhead of postmap(1).
>
>i was thinking about using the list in a DNS
>server (rhsbl) so i can use
>the same list in few box's.
A local RBL will likely be some slower than a
local hash or cdb file, but the size of the
database will have little or no effect the lookup
speed either way.
I would use local indexed files rather than a
local RBL, but use whichever suits your needs.
> > > is it a good ida to use such a BIG list ?
> >
> > It's hardly a performance problem, but would
> it really stop that much
> > crap? Spammers often change their sender
> addresses, so I expect 99.9 %
> > of the domains in that list won't hit your
> server in a million years.
> > Static blacklists are not the best way to fight spam.
> >
>
>well, with this file and my logs im seeing like
>10-15% of the same kind
>a spam comming for weeks or less. thats why i
>was thinking to use this.
>
>thank you Magnus for the reply.
If you think it will be helpful to you, set it up and try.
-- Noel Jones
|
|
|