logs archiveIRC Archive / Freenode / #exim / 2015 / September / 9 / 1
mutante: On our systems, ignore_bounce_errors = 2d, timeout_frozen_after = 4d
hs12: thanks! i'm currently deleting a bunch with exiqgrep and a pipe. (.racing and .date TLD seem to be popular with spammers :p)
Freezing often has some reason, you should check your setup. Lots of frozen bounces indicate, that you accept messages you can't deliver. This should be avoided.
Aha... interesting!
RFC 5321 is the name of the game?
one of many
But basically that one, isn't it?
sudo yum update

why do i have /var/spool/exim AND /var/spool/exim4 and they are _not_ just symlinks or something
this is from the exim4 .deb package on ubuntu
damn, somebody made that setup really confusing then, since the package only creates /var/spool/exim3
now i'm not even sure which data i was working on with my commands to remove from queue ..arr
when rsyncing a mail queue/spool, do i really just need the "input" dir or do i need both, input and "msglog" too
so turns out i can't "exim -M" a single message while the exim service is not running
it just sits there but the command doesnt finish and nothing in mainlog
how can i test delivery with just a single message without starting the service and making it work on the entire queue
hs12, yes thats me. Im still without a proper fix :/
any idea how to do an ACL for line length?
In the body?
Rubin: see his comment on https://bugs.exim.org/show_bug.cgi?id=1687
In the body it's difficult, as I'm not aware of an expansion item containing the whole body (and I'm not sure if I'd like to see such an item in the RAM)
no in the headers..
but not the continued header length, but the individual lines
Rubin: yes, it's what my ACL example should check, the length of the individual physical lines
sweet let me look
from my testing, long lines in the body did not cause hangups
Ah, good news :)
I will put this into place, and see how it goes
Limiting the length of the header field's body isn't sensible, since IMHO there is not limit at all.
(Read: the RFCs do not specify a limit there.)
that was my understanding as well
we should link this bug to mine so people can find it
i also opened a debian bug, so maybe they can put this acl rule in their template
i'm having some trouble with this acl in acl_check_data
if i put it under 'deny' with the single condition..
it cant be overridden later, right?
i want it just to stop and be done no matter what later rules
hrm musta been a typeo. nvm
I have 2 questions that is hard to get an answer on, should probably try mailing list, but will try one more time here:
1) Is acl_c and acl_m variables from original message available to autoreplies and bounces?
2) Is it possible to identify if the current rcpt beeing routed is the original recipient?
bjornar_: for 2) $sender_address contains the incoming envelope sender, $return_path the outgoing sender. At the start of of routing both are the same. /^\$return_path/ in spec.txt
for 1) I'd say, as long as the original message is concert, the $acl_?_* are available.
hs12, yeah sure.. but what about the autoreplies and bounces generated _from_ the original message
I have a website on server A and the email is at Server B email.domain.com aliases server B (rackspace mail) But exim on Server A thinks the address is local because the web domain is pointing to the local web server.
so mail appears to exim as unroutable
what do I need to do to get Server A's exim to understand the mail needs to route to Server B.
oasisfleeting: i'd say an MX record in DNS
I have provided the correct dns records.
and looking at mxtoolbox they are pointing to server b
digging points correctly
points to mx.emailsrvr.com
which is rackspace
« prev next »