You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by "Rob McEwen (PowerView Systems)" <ro...@powerviewsystems.com> on 2006/11/17 21:13:20 UTC

Could THIS have doubled my SA Speed...

RE: Could THIS have doubled my SA Speed...

First, I'm using a windows Port of SA... and I use this as a helper application in addition to my own custom programmed spam filter. Along these lines, I purposely have RBL checks and URI checks disabled in SA because I do these myself. But I **do** have Razor2 and DCC enabled.

Anyways, I was trying to see what I could do to speed SA up as it seemed slower than it used to be.

I tried adding a "resolv.conf" file (which wasn't previously there) and entered my local DNS caching server there.

Then, I restarted SpamD and ran a corpus of 50 test files through SA (using a batch file, processing them one-by-one)... and this 2nd time it processed twice as fast. I ask if these results sound correct because I figure that my results might be anidotal. Does this type of speedup sound correct?

I know that using a local DNS caching server can speed things up, but I was only specifying the SAME one what was already the default DNS server in my NIC card setup... so I would have thought that this would have already been the one chosen.

But I have another question:

It stands to reason that, even though I have RBLs and URI-checked turned off, there must be something ELSE that is getting checked across the network (via DNS)... or OTHER DNS traffic besides just RAZOR and DCC. Any ideas what that might be?

I guess I was a bit surprised at this speedup since I have most of these DNS-type checks disabled. (But maybe there is still more going on via DNS that I realize?)

Thanks!

Rob McEwen
PowerView Systems
rob@PowerViewSystems.com



RE: Could THIS have doubled my SA Speed...

Posted by Darren Cockburn <DC...@cigionline.org>.
Yes,

I have tested this by pointing my entries in resolve.conf to our
non-caching server. . .  definitely slowed it down ... about 1/2 the
speed
(could not do r-DNS as quickly) :)

- Darren.

-----Original Message-----
From: Jeff Chan [mailto:jeffc@surbl.org] 
Sent: Saturday, November 18, 2006 1:36 AM
To: users@spamassassin.apache.org
Subject: Re: Could THIS have doubled my SA Speed...

On Friday, November 17, 2006, 12:13:20 PM, Rob Systems) wrote:
> I tried adding a "resolv.conf" file (which wasn't previously there)
and entered my local DNS caching server there.

> Then, I restarted SpamD and ran a corpus of 50 test files through SA
(using a batch file, processing them one-by-one)... and this 2nd time it
processed twice as fast. I ask if these results sound
> correct because I figure that my results might be anidotal. Does this
type of speedup sound correct?

> I know that using a local DNS caching server can speed things up, but
I was only specifying the SAME one what was already the default DNS
server in my NIC card setup... so I would have thought that
> this would have already been the one chosen.

Under *nix/nux Net::DNS wants to use the first resolv.conf record
as it's resolver.  I'd imagine it could get confused or delayed
if there were no resolv.conf.  To compound the wild guess,
perhaps it works that way under Windows too.

> It stands to reason that, even though I have RBLs and URI-checked
turned off, there must be something ELSE that is getting checked across
the network (via DNS)... or OTHER DNS traffic besides just
> RAZOR and DCC. Any ideas what that might be?

A packet sniffer might tell you.

Jeff C.
-- 
Jeff Chan
mailto:jeffc@surbl.org
http://www.surbl.org/


Re: Could THIS have doubled my SA Speed...

Posted by Jeff Chan <je...@surbl.org>.
On Friday, November 17, 2006, 12:13:20 PM, Rob Systems) wrote:
> I tried adding a "resolv.conf" file (which wasn't previously there) and entered my local DNS caching server there.

> Then, I restarted SpamD and ran a corpus of 50 test files through SA (using a batch file, processing them one-by-one)... and this 2nd time it processed twice as fast. I ask if these results sound
> correct because I figure that my results might be anidotal. Does this type of speedup sound correct?

> I know that using a local DNS caching server can speed things up, but I was only specifying the SAME one what was already the default DNS server in my NIC card setup... so I would have thought that
> this would have already been the one chosen.

Under *nix/nux Net::DNS wants to use the first resolv.conf record
as it's resolver.  I'd imagine it could get confused or delayed
if there were no resolv.conf.  To compound the wild guess,
perhaps it works that way under Windows too.

> It stands to reason that, even though I have RBLs and URI-checked turned off, there must be something ELSE that is getting checked across the network (via DNS)... or OTHER DNS traffic besides just
> RAZOR and DCC. Any ideas what that might be?

A packet sniffer might tell you.

Jeff C.
-- 
Jeff Chan
mailto:jeffc@surbl.org
http://www.surbl.org/