Hi Arthur,
first of all - I won't beat anyone here, as i asked for others'
opinions. :)
If you ask me, i would prefer to let all go through, but my server
wouldn't survive that high load then. Without spamdyke my server load
goes up from a load average of 0.10 to over 30.0 and more.
I tried that... unvolunteerly... You're right, if you say i allow
my users a lot, but the way i did it, i always get reports,
when the list updates are done. If something goes wrong, it's on me
to fix it. Thats always 30 minutes for me to react (theoretically). I
didn't figure out a way to configure spamdyke to use per-domain-lists,
so i've chosen this one.
I think i'll give it a month or two as testing period, to see what my
customers "kill" or not, do or not. Another thought is just to let them
maintain the whitelist. hehe.
All in all it's not only my decision. My boss wanted that. :-)
Thanks for your statements about this.
Arthur Girardi schrieb:
Heya,
I can't say what you are trying to do is a good thing, giving this
kind of power to your customers is in my humble opinion, as Sam use to
say, a solution looking for a problem. I think that in your case you
will likely run over two or more customers disagreeing in the choice
of filters, sooner than it may seem.
I think that it would require of spamdyke to work on a much more
user-level kind of configuration than what it is capable of today. (I
may be completely wrong in this matter, tho, as I haven't configured
spamdyke to its deepest maximum usefulness).
Anyway, letting spam go thru and putting the responsability of
deciding what to block on the customer, that and considering most
customers do not have a good technical knowledge, looks wrong to me.
You said that blocking e-mails without the consent of the customer is
borderline illegal in Germany, but what if a customer end up putting
by mistake one of those big providers in a blacklist? That would
affect other customers as well, and you would end up taking the blame
the same way!
My advice is that you should declare your anti-spam policy when a
customer signs in, exempting yourself from criminal responsability in
case of legitimate mail being rejected. I don't know if that is ever
acceptable in Germany though, and most importantly, in the market you
serve.
But those are just my 2 cents. Please don't beat me. :)
Cheers
Arthur
Hi all,
i've written a Spamdyke GUI for Plesk for my customers, so that they all
have their own responsibility,
if they want to use greylisting and are able to maintain their black-and
whitelists. It's nice, as they all can see, what's
really happening in the mailsystem and keep away spammers and welcome
their customers... Rejecting
mails without letting the customers know, is near the border to being
illegal in Germany, because the customers
can make me, or my company responsible for missing mails.
But one problem i have is the logic of where i keep those lists. At the
moment i just save them to the Plesk
database and dump them regularly by cron-job to special files called
customer_blacklist_ip, customer_blacklist_rdns,
and so on, which are used by spamdyke. That's a good way to write them
with root and keep all privileges healthy
and i can let it send a report to me, what has been done.
Do you think it's a good politic to activate them globally? I understand
it in the way that every whitelisted entry
should be a possible "good" sender for the others too. The critial point
are the blacklists:
Of course i avoided that they add known IP's, i.e. my mail server's
network and local IP's and also created
a button to check the reverse data. As far as i know, thats a way the
"big" providers do it, i mean tagging
mails manually as spam or ham.
Am i right?
Greetz,
David
_______________________________________________
spamdyke-users mailing list
[email protected]
http://www.spamdyke.org/mailman/listinfo/spamdyke-users
_______________________________________________
spamdyke-users mailing list
[email protected]
http://www.spamdyke.org/mailman/listinfo/spamdyke-users
