| |
| |
| |
|
Comments:
<0> twkm: that is (can be) done via localhost. with port 25 totally blocked for the connections from outside, as sm suggessted.
<1> so then the point of the whole annoyance here, don't accept mail to local accounts, has gotten lost in the arguments progressing?
<2> What kind of mail server are we talking about here?
<2> It would save us the time arguing about off-topic cases
<0> sm: I don't think we are talking about any mail server.. I think we agreed with you on all (or most) points
<2> I may be wrong. Everyone has his own way of running a mail server
<0> twkm: no issue here, really.. All I meant to make a short comment that there are _legitimate_ situations when a particular host would be configured to send messages out but not receive any messages for local users from _outside_. And to make it clear: this is not a one host solution, - but where the host (e.g. smtp-out) is a part of the bunch in an ISP layout.
<0> sm: BTW, I am curious what solution you use for messages from .cn and such?
<1> i've yet to see that usefully, but whatever.
<2> str, there is no magic bullet. Filter
<0> sm: sure.. but I am just curious - what do you do with the messages that you filter out? you don't deny connect, don't use /dev/null.. Are you collecting them for a special Christmas gift to the FCC <grin> ?
<2> reject
<0> sm: ghm.. ok, it must be miscommunication here.. "<2> You can even send me mail from .cn:)" - I understood it as if you accept those...
<0> Now I see what you meant.
<2> I accept it if it is what the users define as valid
<2> If not, I tell you why it is not accepted
<0> sm: based on the per-user "whitelist" ?
<2> I don't do country blocking
<2> That and other rules
<0> sm: "what the users define as valid" -- is there anything in this part which is not per-user whitelist?
<2> content. No need to whitelisting like that as a whitelist ***umes that you know from whom the mail comes and from where it is sent
<0> ah, ok. but that is not exactly what "the users define as valid". As first you accept, and only then the user can look at the content. What your filters actually do is that based on what the users have defined as valid previously, they (filters) analyze and decide if the new message falls into the same category. Though, it is semantics
<2> No. :) It's a bit complex to explain
<0> sm: are you using anything that is by its logic completely different from "default configuration" spam******in, i.e. content (and header) pattern matching, calculating bayesian probability, and maybe looking up some well known BLs ?
<2> pattern matching is only one part. I don't rely on blacklists
<0> sm: besides pattern matching, what else do you do (if you don't mind). No need to get into details, - I am interested in the conceptual part.
<0> sm: ok. whatever... :)
<0> then no need to show off with it :-P
<3> exit
<4> bom dia
<5> who's up
<6> Ah, spork selling Viagra
<5> um.. so I just jumped on this box
<5> at of no were it stopped listening on it's external nic
<5> now i look and port 25 is bound to localhost
<5> any ideas?
<6> "at of no were it"
<6> You must be spending too much time on AOL
<1> heh.
<5> out of no where
<6> restart sendmail
<5> no
<1> sendmail.mc might explain it.
<1> rc.conf perhaps.
<5> before I do this I want to try and understand how this happened
<1> depends on the borx.
<6> What OS is it?
<5> freebsd
<6> What is in your cf?
<5> I've never seen this before
<6> check your cf
<1> rc.conf.
<5> nothing special, nothing to do with binding
<6> well, sendmail can listen on localhost
<5> yeah
<1> unless you make a specific and documented entry in rc.conf the rc.sendmail script will start sendmail bound only to localhost.
<5> that makes a little since but doesn't explain this
<5> I can understand if it starts up and binds localhost and nothing else because of entry
<5> what I don't understand is how it can start up and work, yet fail later
<6> You mean it fails just like that?
<6> When does it happen?
<5> randomly
<6> any errors in messages?
<5> thre could be
<5> this is a very busy server and pinning down the exact failure time is tough
<5> since theres no active monitoring on it
<5> um.
<5> also
<1> gremlins have been editing your rc.conf -- perhaps a portupchuck, err portupgrade, did it.
<5> defaults/rc.conf only binds localhost for submit
<5> otherwise it's open
<0> spork: what does 'fgrep sendmail /etc/rc.conf' show ?
<5> there are no sendmail entries in rc.conf, it uses all defaults
<5> I think im going to stick a monitoring script on this box
<5> and watch it
<0> spork: default doesn't enable the inbound sendmail daemon
<6> good idea, spork
<0> spork: if that is what you want to have, you need to have sendmail_enable="YES" in /etc/rc.conf
<0> sendmail_enable="YES"
<5> StR yes I know
<0> spork: then I am sorry, I don't understand what is your question? You are wondering why it stopped listening on the external IP, but to me it is a puzzle why it was listening there in the first place.
<0> However, I would try to restart sendmail and see if it gives any error messages in /var/log/maillog
Return to
#sendmail
or
Go to some related
logs:
_DjScrew_
#computers
#computers
put moving text on images
#beginner
#sex
#politics
#computers
#politics
#openbsd
|
|