| |
| |
| |
|
Page: 1 2
Comments:
<0> hello, why is my mail able to send out, but not recieve?
<0> is there a port needing open
<1> probably.
<1> did you bother to read your vendor's documentation?
<0> there is a vendor? where.. heh I downloaded fedora free
<1> fedora is your vendor.
<1> did you read their documentation?
<1> (say no, we all know you didn't)
<0> know about smtp port 587 getting that setup for auth and all
<0> cuz my setup here blocks 25
<1> what makes you think 587 will work if 25 won't?
<1> anyway, the smtp auth section of cf/README has the basics. also i think the sial.org site has something.
<0> twkm, there has been found on forums to set to 587, when 25 is blocked
<1> if your own setup blocks 25 i doubt you can figure out how to use 587.
<0> twkm, I wont with a negetive additude. I can do it, if people will help me find the solution or point me in the direction
<1> have you considered reading anything i've suggested?
<0> I just see only read vendor info, and also what makes you think u can use 587
<1> "fedora is your vendor" ... "did you read their documentation?"
<1> and, "the smtp auth section of cf/README".
<1> and, "i think sial.org has something".
<2> Whats the difference between libmilter.a and libmilter.so?
<1> the usual difference.
<2> and for the uninitiated?
<2> I'm having build failures and I think it's because of a very old .so file
<1> one is a static library, the other is shared.
<2> but I can't point sendmail to look exclusively at the .a file
<1> you probably can if you know how to bribe your compiler.
<2> http://www.technoids.org/libmilter.so.html <-- actually this looks like what I need.
<3> eeei
<3> any ldap user here?
<4> is it possible to send incoming mail to a policy system before the data command is issued? i.e. after mail from+rcpt to?
<1> uh.
<1> before the data command there is no mail.
<4> so?
<4> i want to drop mail that isn't destined for any user before i waste time processing it
<4> although, i want more matching functionality than virtusertable has to offer
<4> i think postfix does regex matching, which is all i really want.. just not sure if i want to migrate to postfix.. i have mostly everything with sendmail already
<1> sendmail should always do that.
<4> to some degree, sure, but it doesn't work so hot when you have a catchal
<1> well duh.
<1> how can you reject anything with a catch-all?
<4> well, i don't want a catchall
<1> bah.
<4> i want user.*@domain, not *@domain
<1> then why did you mention it?
<1> user.* -- wtf is that?
<4> zemmiphobia.dell@domain
<1> sendmail already has + addressing. user+whatever@fqdn
<4> you're right
<4> i wish i had known about that functionality many years ago
<4> although, does that really work for spam segregation? seems to me spammers would recognize that anything before + is the valid user id, and then strip the rest to avoid filtering
<1> ahh, so you hacked something into some mta to provide you user.whatever?
<4> i suppose it would work for ignorant spammers
<4> no
<4> i've just always used myname.company@domain when registering for things
<1> oh. then what's the years ago reference? i'm a little lost.
<1> and so you have lots of virtusertable entries?
<4> hah, well, no.. when i first started it was a friend's colo and he hosted my mail, and he gave me @domain so i thought, hey! cool! i can see which companies are spamming me! ... but, i was young and didn't now anything about smtp or rfcs
<4> but, yeah, i do have quite a few virtusertable entries
<1> and you cannot now recall them all, and don't want something to bounce if you forgot it?
<4> i've tried to do user.badcompany@domain nouser:550 No such user here, but sendmail still usually delivers the mail for some reason
<4> exactly
<1> yeah, that can be a dilemma.
<4> i'm trying to document all of my 'known good ones'
<4> and specify them
<1> so, you need a catch-all after all.
<4> that way i can move the catch-all to something after that
<4> and if i ever did suspect i lost it, i could just search through the moved stuff
<1> then after a while you get rid of it and pray you got them all.
<4> yeah
<1> okay, so, lets say you could forward to a policy server ... just what would policies would you test against?
<4> but, what i'm seeing currently, is a bunch of random letter mails... asdsadgasdg@domain, which i am certain doesn't exist
<4> first i'd drop any mail i knew was bad, i think it would be more effective than parsing To: and CC: headers
<4> more efficient, anyway
<4> also, i would drop anything that didn't have my username in it
<4> i wish i could do user*@domain user, @domain dev-null... but alas, limited pattern matching is available with hash tables (read: none)
<1> okay, well, there are two possibilities. one is a milter, which is intimately involved in the smtp transaction. the other is a bit of forwarding, which is not and has issues deciding what to do with failures.
<4> after mail p***ed the primitive yet effective first line, then i could parse it with more in-depth, processor-intensive applications
<4> can milters be used before the data command though?
<1> ugh.
<1> it is *PART OF* the smtp transaction.
<4> oh
<1> the milter decides at every point whether to allow, reject, etc....
<1> connect, helo/ehlo, auth, mail from, each rcpt to, data, end of headers, end of data, end of message.
<1> anyway, one milter that might help you, a little, is milter-regex.
<4> excellent
<1> or you might just use a milter to rewrite from user.* to user+* then let normal virtusertable rules handle the mapping.
<4> yes, that would be interesting...
<1> or, you might attack the +detail rules and make them into .detail rules -- which means you must be careful of the gecos stuff, but that is normally disabled anyway.
<1> a local rule 3 could easily change user.* into user+*, no milter needed, come to think of it.
<4> oh?
<4> hmm
<4> looking at sendmail.cf (freebsd), i don't see where the + stuff is parsed? i'm looking around "Ruleset 3 -- Name Canonicalization"
<4> wait, i think i found it
<1> +detail isn't "parser" in the canonicalization rules.
<1> look in ruleset 0, in particular in parse1.
<4> hmm
<4> what about ruleset 5? "Ruleset 5 -- special rewriting after aliases have been expanded"?
<4> it seems to be checking for +details
<1> that is what it is called.
<1> +detail
<4> hmm
<1> e.g., a series of virtusertable lookups are done where the user+detail is changed from user+detail to user+ to just user in the hopes of finding a matching key in the map.
<4> is it under "# handle virtual users"? looks pretty complicated :/
<4> right
<1> so, you could change it to use . instead of +.
<1> it would be a largish hack, but mostly you would be changing single characters and not trying to write a rule (which isn't all that hard, but it scares some right off).
<4> it looks pretty scary, but doesn't look like it has many components .. so i agree, it probably isn't that time once you read about it
<4> time consuming*
<4> i think i will use the regex thing first ;)
<4> at least in the meantime
<4> do you know any good documentation for those rules? i'd like to learn more about it
<1> well, doc/op/op.* talks about them, as does the bat book.
<1> also there is a little in cf/README, dealing mostly with how to put them into an mc file.
<4> sweet, i'll have to check it out
<4> thanks for your insight
<1> sure thing.
<5> Say I have a text file with a complete email. How can I feed this to sendmail easily?
<6> Q: sendmail was turned off for some reason few months ago, now clientmqeue is huge, can't even ls the directory, possible if I can grep out a specific email in queue?
<1> usually.
<6> well : bash: /usr/bin/grep: Arg list too long
<1> sigh, basic sysadmin course.
<1> what were you looking for?
<1> to, from, subject, relay machine, content-type?
<6> yeha to email.
<6> I guess I need to throw it in a loop
<1> mailq -qRrecipient@address
<1> err, mailq -Ac -qRrecipient@address
<6> so no spaces between -qR and address?
<1> nope.
<7> hello, i have an mx configured with sendmail, with sends email to another box where imap is installed, now the problem is that you can send user to an account that does not exist and email gets bounced back to postmaster, and ther are a log of emails, any ideas how i can fix this?
<1> don't do it that way.
<1> give the sendmail machine a way to validate the addresses, then deliver via the network.
<7> so lets say, i create virtusertable and add all valid emails in that file, will this work?
<7> and @domain.com error:nouser User unknown
<7> ?
<1> yes.
<7> well it doesnt, i just added adrian@domain.com adrian@domain.com \n @domain.com error:nouser User unknown, and it still p***es the email to the other box
<7> email sent to some.name@domain.com
<1> then domain.com is not considered a virtual domain.
<1> is it in local-host-names?
<7> twkm: yes
<1> then sendmail should treat @domain.com as a catch-all, matching if the address doesn't match any other keys that end with @domain.com.
<1> did you remember to makemap it?
<7> twkm: yes
<1> then it should work. you should debug it.
<8> everything is working correctly making a telnet of the mail server
<8> BUT
<8> when i send a mail from my ISP adress to a user who belongs to my linux mail server (which sendmail is running) i can't get it
<8> and a full TCP test with TELNET works
<8> is someone up to have a look of my config ?
Return to
#sendmail
or
Go to some related
logs:
mb_internal_encoding wamp
#perl
xorg xgl displaymanager_xserver
liblame0 ubuntu
could not init font path element, xubuntu
crystal report outer join criteria
kopete with jingle debian
#debian
no local.conf
javascript var length undefined
|
|