logs archiveIRC Archive / Freenode / #exim / 2009 / November / 18 / 1
bmusson-work
if i created a sub domain bo.example.com, and i send an email to user@bo.example.com- the email server which the email goes to should be resolved by the NX record in dns, right?
petemc
it would try to look up the mx record for bo.example.com, and if one didnt exist it would then look up the A record for bo.example.com and attempt to deliver the mail there
bmusson-work
so either way, bo.example.com needs to ping, right?
or just the MX record's IP?
petemc
bo.example has to resolve to something, yes
CunningPike
bmusson-work: Neither needs to ping
bmusson-work
CunningPike, huh?
CunningPike
However, they need to resolve and answer on SMTP ports
bmusson-work
oh right
CunningPike
ICMP != DNS != SMTP
         

bmusson-work
so telnet bo.example.com 25 should show me something
CunningPike
Ideally, yes :)
bmusson-work
ah
tcrass
I'm having a bit of a problem and was wondering if someone could give me a bit of help. I have been attempting to block some really annoying spam sources, I was hoping to utilize sorbs and spamhaus for some of this but it seems that although things are being detected they are still coming through.
I have put the relevant information into a pastebin post so it would be easier to see things, I would greatly appreciate any help. -> http://pastebin.com/m1200ed43
brb
docelic
Hey folks, sending mail to mailman list on my setup doesn't work due to: R=system_aliases defer (-30): pipe_transport unset in system_aliases router
I don't understand the implication and what do I have to change. Hints appreciated.
Actually, I got it, thanks
bmusson-work
why dont i know how to debug mailman + exim?
ive been trying to get this stuff to work for literally days now.
i can telnet to my exim smtp bo.exampled.com (also mailman server) and send mail to user@exampled.com
ive searched all documents
i see no errors on exim or mailman
cannonball
With the new version of Exim, are there any changes to the code for libspf2? I should be able to use the same tarball to build it, I would think.
redbrain
hey guys i am about ready to rip my hair out with mailman and exim lol
when i send a mail to the list all i get in the log is
2009-11-18 20:28:20 1NAr8U-0003Ba-LI <= herron.philip@googlemail.com H=ey-out-2122.google.com [74.125.78.24] P=esmtp S=2428 id=4B0458DE.4090600@googlemail.com
2009-11-18 20:28:20 1NAr8U-0003Ba-LI == |/var/lib/mailman/mail/mailman post crules <crules@lists.crules.org> R=system_aliases defer (-30): pipe_transport unset in system_aliases router
i have the /etc/aliases setup too
Simon-
all you get?
it tells you what is wrong, fix it.
redbrain
dude wtf is wrong
lol
Simon-
read the error message
redbrain
system_aliases defer (-30): pipe_transport unset in system_aliases router
yes
what is it
some help this community is jezz
CunningPike
redbrain: When you Googled that error, what did you do next?
redbrain
read most of the things but its all just mailing lists
not exactly helpful
http://www.google.co.uk/#hl=en&source=hp&q=pipe_transport+unset+in+system_aliases+router&btnG=Google+Search&meta=&aq=f&oq=pipe_transport+unset+in+system_aliases+router&fp=408428c3d949fea1
all you get is people asking thingsl ike this http://www.opensubscriber.com/message/ubuntu-users@lists.ubuntu.com/2557291.html and its basically my questions
cannonball
redbrain: In that google page, look for the text "How do I resolve this" and click that link.
redbrain
Well, the error is telling you that the pipe_transport is unset in
system_aliases. Your file seems to indicate that that is not true --
it is set. So you need to debug it and make sure it is using the
config file you think it is using. Play detective based on what you
know. Restart exim (% kill -HUP <exim-pid>) as a starting point.
?
cause thats helpful
cannonball
redbrain: Again: In that google search, look for the text "How do I resolve this" and click that link.
redbrain
yeah i did
that was the text from it
cannonball
then you're on the wrong page. It goes to this: http://mail.python.org/pipermail/mailman-users/2005-February/042413.html
marcof
hi, could anyone please help me? I want to know about exim and chroot.
         

cannonball
I don't use exim in chroot, so no help here.
redbrain
see this is the thing i am using debian and for some reason there isnt a /etc/exim.conf only the template file
not sure how exim4_config works
jezzz exim.config.template is 2000 lines long!!!
a config file?
marcof
is it worth installing it in a chroot environment?
cannonball
redbrain: Your Debian system is configured to use the split config system. You can either convert it to a single file (don't remember the command, it's in the mailing list archive from a week or two ago) or grep and find the spot where the routers are defined, specifically the "system_aliases" router.
redbrain
yeah but i choose the single config setup in the exim4_config
since its reccomended
cannonball
then you should have an exim.conf somewhere. Run 'exim -bV' and look at the last line to see what configuration file it's using.
redbrain
i just added pipe_transport = address_pipe
to that config
and still get the error
cannonball
I assume you restarted exim...
redbrain
yeah
cannonball
and added it to the system_aliases router...
redbrain
?? i just added the line 'pipe_transport = address_pipe' to the "/var/lib/exim4/config.autogenerated"
cannonball
It has to be specifically in the definition for the system_aliases router.
redbrain
is that in that file?
the system_aliases: ....
cannonball
it should be. Your system is complaining specifically about that router, so it must exist int he config.
this guy pasted his config sections in an email tot he mailing list a few months ago. Might be helpful (or might just confuse you more) http://www.exim.org/lurker/message/20090819.140811.947173c7.ja.html
redbrain
a fu*k sake this still doesnt work added it to system_aliases: debug_print = "R: system_aliases for $local_part@$domain" driver = redirect domains = +local_domains pipe_transport = address_pipe allow_fail allow_defer data = ${lookup{$local_part}lsearch{/etc/aliases}} .ifdef SYSTEM_ALIASES_USER user = SYSTEM_ALIASES_USER
the config is 800+ lines long
for some reason
when i restart exim it takes that entry i add out
the 'pipe_transport = address_pipe'
i am about to crack up i have been at this all day
Simon-
try complaining to #debian
cannonball
So the init script must be regenerating the config file. Check the init script to see what it's doing would be my next guess. There is also a mailing list specficially for debian exim configs.
redbrain
emacs /etc/init.d/exim4
whoops
cannonball
If it makes you feel any better, I'm not a Debian fan WRT exim either. I run CentOS for our servers and build my own rpm (doing that with the new release right now).
redbrain
ah yeah sorry if i come across angry just i really like exim4 now i used to use postfix but i work as an erp sys admin and its painful enough lol but just exim is much faster so not really used to it plus with debian it worked as a basic mail server just out of the box really well postfix sucks balls
just never though this would be so awkward
cannonball
My favorite thing is perl support. I pull in memcache to various things using perl modules and I can track states of various things system wide: greylisting (implemented), throttle smtp auth senders (done in a sendmail milter, getting it working in exim now).
still looking for new uses.
redbrain
cools
ah just been looking to get a mailing list going i've been desiging my own programming language for the 8 or so months
and going to do a release in a few months
and hope maybe might get some people interested so a mailing list might be nice
cannonball
Good luck with that. What's it targeted at? server side? web browser? Is it scripting or compiled?
redbrain
well its a language designed to be as expressive as possible lol its like a mix of perl python and haskell but with my own ideas in
i've wrriten the interpreter from scratch so nothing really facy works yet other than basics like the memory model garbage collection functions and expressions
though i've setup alot of awkward things so implementing the rest shouldnt be that much work
« prev next »