You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Nigel Frankcom <ni...@blue-canoe.com> on 2010/12/14 17:58:27 UTC

Comment - GFI/SORBS

Hi All,

Is sorbs going to be continued as a scoring option in SA?

Having hit yet more problems with them I've zeroed their scoring.

I found this a couple of days ago, maybe it can add weight.
http://blog.wordtothewise.com/2010/12/gfi-sorbs-considered-harmful/

Best to all

Nigel

Re: Comment - GFI/SORBS

Posted by Nigel Frankcom <ni...@blue-canoe.com>.
This is a long and somewhat complex story. I've been running my own
mail for 15+ years or so, always on a fixed IP. A few years ago
business picked up so I got some additional IP's from my supplier
(BT); it turned out that they were "decommissioned" DUL's renewed as
statics. Initially we jumped the hoops (both BT & I) and after several
fraught weeks the issue was resolved.

Now we hit November 27th this year, suddenly I'm in SORBS again.
Nothing changed this end, same IP, same RIPE entry, same everything...
apart from SORBS, who, apparently, redid their db at the end of
November. Happily I am now clean and clear.

How did I really end up there? I've no real idea, I suspect the
reload. 

I really do appreciate the work RBL's do, mostly; it's a thankless
task and if the same wit were applied adversely a lot of money could
be made. That they are moral and work as they do makes the life of all
legit server admins much easier.... until they get too rabid.

For those of you that supply reliable rbl's, please accept my profound
thanks. Some maybe "could do better", perhaps those should be
carefully judged before inclusion into sa, or perhaps made an
optional?

All that said, SA isn't the direct problem. Admins blocking purely on,
for example, SORBS, should maybe rethink their strategy and adjust
scoring on rules within SA.

All of the above is my opinion only; I don't think SORBS do a bad job,
I just think they could do it better, and maybe accept that we all get
it wrong sometimes... Just my 2.5p worth :-D

Kind regards

Nigel



On Tue, 14 Dec 2010 22:41:40 -0500, Jason Bertoch <ja...@i6ix.com>
wrote:

>On 12/14/2010 8:06 PM, Bart Schaefer wrote:
>> http://blog.wordtothewise.com/2010/12/gfi-sorbs-considered-harmful-part-5/
>
>I've seen the headaches of getting off SORBS, but how did you really end 
>up there?
>
>While I agree that SORBS is not reliable enough for use at the MTA 
>level, I've not seen one complaint from my customers over using SORBS in 
>SA.  Isn't the beauty of SA the fact that you can score gray areas and 
>not be stuck with black or white?
>
>In case it's a mystery, SA scores are automatically generated based on 
>results from the corpus.  If those results weren't productive, the rules 
>would either be disabled or their scores adjusted even lower.  However, 
>if the corpus isn't representative, the generated scores are in error, 
>and that means we need more trusted submitters.  Or maybe your traffic 
>is relatively unique and you should already be generating your own scores?
>
>Ultimately, this seems to be more of a witch hunt against SORBS than a 
>SA issue.  Although I'm not opposed to a SORBS witch hunt, I don't think 
>it belongs here.
>
>/$.02

Re: Comment - GFI/SORBS

Posted by Nigel Frankcom <ni...@blue-canoe.com>.
On Wed, 15 Dec 2010 07:04:18 +0000, "corpus.defero"
<co...@idnet.com> wrote:

>
>> Ultimately, this seems to be more of a witch hunt against SORBS than a 
>> SA issue.  Although I'm not opposed to a SORBS witch hunt, I don't think 
>> it belongs here.
>
>Indeed, and it's Lynford and his money grabbing cronies mostly behind it
>- hence it lacks sophistication.

I guess we all have our opinions based on our experiences. Personally,
I've had no issue with zen, though cbl does seem sometimes to have an
issue with back-scatter. That said, proper spf should help stop
back-scatter.

Kind regards

Nigel

Re: Comment - GFI/SORBS

Posted by "corpus.defero" <co...@idnet.com>.
> Ultimately, this seems to be more of a witch hunt against SORBS than a 
> SA issue.  Although I'm not opposed to a SORBS witch hunt, I don't think 
> it belongs here.

Indeed, and it's Lynford and his money grabbing cronies mostly behind it
- hence it lacks sophistication.


Re: Comment - GFI/SORBS

Posted by Jason Bertoch <ja...@i6ix.com>.
On 12/14/2010 8:06 PM, Bart Schaefer wrote:
> http://blog.wordtothewise.com/2010/12/gfi-sorbs-considered-harmful-part-5/

I've seen the headaches of getting off SORBS, but how did you really end 
up there?

While I agree that SORBS is not reliable enough for use at the MTA 
level, I've not seen one complaint from my customers over using SORBS in 
SA.  Isn't the beauty of SA the fact that you can score gray areas and 
not be stuck with black or white?

In case it's a mystery, SA scores are automatically generated based on 
results from the corpus.  If those results weren't productive, the rules 
would either be disabled or their scores adjusted even lower.  However, 
if the corpus isn't representative, the generated scores are in error, 
and that means we need more trusted submitters.  Or maybe your traffic 
is relatively unique and you should already be generating your own scores?

Ultimately, this seems to be more of a witch hunt against SORBS than a 
SA issue.  Although I'm not opposed to a SORBS witch hunt, I don't think 
it belongs here.

/$.02

-- 
/Jason


Re: Comment - GFI/SORBS

Posted by Bart Schaefer <ba...@gmail.com>.
http://blog.wordtothewise.com/2010/12/gfi-sorbs-considered-harmful-part-5/

Re: Comment - GFI/SORBS

Posted by "corpus.defero" <co...@idnet.com>.
On Tue, 2010-12-14 at 16:58 +0000, Nigel Frankcom wrote:
> Hi All,
> 
> Is sorbs going to be continued as a scoring option in SA?
> 
> Having hit yet more problems with them I've zeroed their scoring.
...
I hope so. I find SORBS wonderful in dealing with those troublesome
mailers that have managed to by passage from the likes of $pamhau$$ and
Barracuda myself.

That said, I'd like to see the total removal of rules that favour that
haven of transactional spammers - Return Path.