is the following a general problem, or just one i am seeing?
note 2821 says
450 Requested mail action not taken: mailbox unavailable
(e.g., mailbox busy)
550 Requested action not taken: mailbox unavailable
(e.g., mailbox not found, no access, or command rejected
for policy reasons)
randy
Randy Bush writes on 12/1/2003 1:10 PM:
is the following a general problem, or just one i am seeing?
Verizon does SMTP callbacks, connecting back to the MX of the envelope sender and trying to verify that the user exists
2003-12-01 10:09:05 1APbBa-000Ork-DY == foo.user@verizon.net <foo@psg.com> R=lookuphost T=remote_smtp defer (0): SMTP error from remote mailer after MAIL FROM:<nikkiadamczyk@gerbangmail.com> SIZE=5365: host relay.verizon.net [206.46.170.12]: 450 Requested mail action not taken-Try later:sc004pub.verizon.net
So this would connect to the MX of gerbangmail.com and try to verify that whatever@gerbangmail.com exists.
gerbangmail.com. 5h55m6s IN MX 0 sitemail.everyone.net.
If the MX (sitemail.everyone.net) doesn't respond fast enough for verizon, their tester times out, and the message gets 4xx'd.
srs
I think you will find that people who want to reject the spam
but don't want to accidentally reject real mail will sometimes
use 45x instead of 55x error codes.
I know when i was rejecting spam at the SMTP layer
I first started rejecting with 45x and watched my logs
for those pesky false positives.
there might be some thought about punish the disk space of
those that are trying to deliver the spam to you ...
- Jared
Out of curiosity, would you know offhand how they do the
validation?
Neezam.
is the following a general problem, or just one i am seeing?
Verizon does SMTP callbacks, connecting back to the MX of the envelope
sender and trying to verify that the user exists2003-12-01 10:09:05 1APbBa-000Ork-DY == foo.user@verizon.net <foo@psg.com> R=lookuphost T=remote_smtp defer (0): SMTP error from remote mailer after MAIL FROM:<nikkiadamczyk@gerbangmail.com> SIZE=5365: host relay.verizon.net [206.46.170.12]: 450 Requested mail action not taken-Try later:sc004pub.verizon.net
So this would connect to the MX of gerbangmail.com and try to verify
that whatever@gerbangmail.com exists.gerbangmail.com. 5h55m6s IN MX 0 sitemail.everyone.net.
If the MX (sitemail.everyone.net) doesn't respond fast enough for
verizon, their tester times out, and the message gets 4xx'd.
interesting but utterly irrelevant. the question was not how
verison decided it was spam. the point was that their server
returned a 450 as opposed to a 5xx (550 looks good), and this
causes net damage.
randy
Neezam Haniff writes on 12/1/2003 1:46 PM:
Randy Bush writes on 12/1/2003 1:50 PM:
interesting but utterly irrelevant. the question was not how
verison decided it was spam. the point was that their server
returned a 450 as opposed to a 5xx (550 looks good), and this
causes net damage.
They haven't yet determined that it is spam. So, RFC nitpicking wise, they are right.
On the other hand, from a mail operations standpoint, I personally feel that sender verification, graylisting and other methods that rely on 4xx'ing email are a bad idea, as they makes things inconvenient for a whole lot of ISPs ... and because these emails have to be either 5xx'd or trashed sometime sooner or later.
I think he's saying that they were unable to perform
the validation hence the 450. If the validation was successful,
they'd return a 200 series code, if it was unsuccessful, they
would return a 500 series code.
- jared
I think he's saying that they were unable to perform the
validation hence the 450. If the validation was successful,
they'd return a 200 series code, if it was unsuccessful, they
would return a 500 series code.
nice words, but crap. due to needs to spool mail for sites in
countries with very poor connectivity, mail spool time here is
quite long. if verizon and others seem unable to decide in weeks,
why should i pay the penalty?
but, i guess the problem is easily solved with exim config. i have
set it so that if it can not deliver to verizon in say one hour, it
dumps the mail.
verizon.net * F,1h,5m
life is simple, except for verizon users i guess.
randy
> I think he's saying that they were unable to perform the
> validation hence the 450. If the validation was successful,
> they'd return a 200 series code, if it was unsuccessful, they
> would return a 500 series code.nice words, but crap. due to needs to spool mail for sites in
countries with very poor connectivity, mail spool time here is
quite long. if verizon and others seem unable to decide in weeks,
why should i pay the penalty?
you should likely queue those other countries on
a seperate machine dedicated to that purpose. this way one
user/host site doesn't unduly cause significant impact to other
sites/users.
it's interesting you view the interpertation (which at least one
other person views as correct) as "crap". this behaviour does
seem to fit strict interpretation of the rfc in question.
but, i guess the problem is easily solved with exim config. i have
set it so that if it can not deliver to verizon in say one hour, it
dumps the mail.verizon.net * F,1h,5m
life is simple, except for verizon users i guess.
this is the ability of a single host operator to make
their own local policy decisions. you've both done what
you feel is appropriate.
- jared
I personally haven't seen ANY validation, just an arbitrary block that's been in place for over a month without cause, reason, or even any ability to contact them. It appears nobody at verizon is at the helm anymore. I've tried several times to contact abuse, postmaster, etc, and even a couple people from this list gave me or forwarded my plight to internals with no results. Modwest is still being blocked. Perhaps not very operational in content though here...
this is the ability of a single host operator to make
their own local policy decisions.
which leads to the heat death of the net
randy
or allows the net to prosper, since policy is distributed rather than centralised.
FWIW, there are now fake smtp tarpits out there that behave this way (4xx deferrals of suspected spam).
http://www.benzedrine.cx/relaydb.html
The idea is to "punish" spammers by filling up their queues, although honestly I don't know of any spammers who actually *have* queues. They just borrow other people's of course. I'm not sold on the wisdom of this course.
this is the ability of a single host operator to make
their own local policy decisions.which leads to the heat death of the net
or allows the net to prosper, since policy is distributed rather
than centralised.
consider verifying, or making any assertions about, or having any
confidence in, the correctness of a universe of distributed and
opaque policy.
i once worked for an m&a isp which acquired some sixty+ small isps.
my comment on the non-global thinkers in the acquisitions was
"think locally, act globally."
randy
Tell anybody who suffered through 69/8 how that was an example of the net prospering.
Correct. More and more, anti-spammers are annoying me more than
the spammers. Anti-spammers tend to "make my problem YOUR problem"
thinking. Be it mangled sender addresses (this "NOSPAM" nonsense),
be it 450 to suspected spam.
Antispanners seem to be very easy in accepting collateral damage
to the net.
Regards,
Daniel
jared:
this is the ability of a single host operator to make
their own local policy decisions.
randy:
which leads to the heat death of the net
joe:
or allows the net to prosper, since policy is distributed rather than
centralised.
randy:
consider verifying, or making any assertions about, or having any
confidence in, the correctness of a universe of distributed and
opaque policy.
to further amplify and make clear whose ideas i am stealing, let
me quote tim griffin:
The Interdomain routing system will enter a state of non-
convergence that is so disruptive as to effectively bring down
large portions of the Internet. The problem will be due to
unforeseen global interactions of locally defined routing
policies. Furthermore, no one ISP will have enough knowledge
to identify and debug the problem. It will take nearly a week
to fix and cost the world economy billions of dollars. The
world press will learn that the internet engineering community
had known about this lurking problem all along....
[ tim goes on to suggest how we might keep from getting in
such a mess ]
i suggest that we take this seriously.
randy
telling spammers 4xx or 5xx doesn't matter, they don't listen.
Exactly this is the flawed point about returning 4xx. They produce
only collateral damage, but don't hit their target at all.
Regards,
Daniel