From: (no name) (no email)
Date: Thu May 02 2002 - 02:26:50 EDT
Mail server vs web server in regards to valid PTR record is going to
be a big difference...you have to remember tons of people are hitting
your web server where as only other mail servers are talking to your
mail server (or should be)...I would have to throw in a "on the fence"
judgement on this subject. My own personal mail server will not talk
to other mail servers that don't have a PTR setup...I can be very
rigid with it if I wish and ALL mail servers really should have a PTR
record setup properly. However, my mail servers that I run at my
company do not have the restriction...many many many idiot admins out
there do not even know what a PTR record is and my business customers
get down right pissed when they can't get email from some other
business. I am having a heck of a time just using blacklists because
of all the idiot admins with open relays. I am having to whitelist
quite a few servers because I have had some large business customers
threaten to take their business elsewhere if I didn't let some open
relay (client of theirs) get mail through. So I think the whole PTR
record thing is similar...certain servers you can be very rigid...some
you have to be more lax on and the points about delegation below are
very valid as some ISPs don't even know how to delegate it out.
On Wed, 1 May 2002 16:06:00 -0400 (EDT), you wrote:
>Yes, you are correct about the webserver part. I was just using that as
>an example for argument's sake. However, I would be curious to see how
>that does stack up when run against a maillog... maybe I'll try that and
>see where it goes.
>
>-Gary
>
>Running Windows is kinda like playing blackjack:
>User stays on success, reboots on failure
>
>On Wed, 1 May 2002, Doug Clements wrote:
>
>> Just a small something I noticed..
>>
>> on 5/1/02 12:00 PM, Gary D. Margiotta at wrote:
>> > To chime in on this as well, in running reverse lookups for our analog
>> > reports, almost 65% of the hosts hitting one of the sites were
>> > unresolveable. This server receives over 50,000 distinct hosts requesting
>> > pages per day, and 65% of 50,000 is just around 35,000 hosts. That's a
>> > lot of damn hosts which don't resolve. And while they're mostly part of
>> > dhcp pools, just setting up a blanket reverse zone shouldn't be all that
>> > hard.
>>
>> It sounds like you're talking about a web server, which talks to a very
>> different class of machines than normal smtp servers. Barring the minority
>> that send mail from their dialup/dsl/cable modems, most smtp traffic comes
>> from other ISP's mail servers. I would hope that administrators give greater
>> attention to proper configuration of their servers than to their dial pools.
>>
>> While I get (and tend to agree with) the point you're trying to make, I
>> think the numbers could be adjusted to be a little more accurate. Maybe run
>> log analysis on your smtp logs, and compare them to what you got for web
>> logs? That might be interesting to see.
>>
>> --Doug
>>
>> -
>> To unsubscribe, send mail to with content
>> (not subject): unsubscribe postfix-users
>>
>
>-
>To unsubscribe, send mail to with content
>(not subject): unsubscribe postfix-users
-
To unsubscribe, send mail to with content
(not subject): unsubscribe postfix-users
|
|
|