Hi Marc, This is getting even more strange.In short: the wrong behaviour disappeared, and it seems I'm not able to reproduce the bug anymore.
I left the office after my last email, and come back this morning. Follows detailed explanation.
From: "Marc Haber"
I am kind of out of ideas now. Can you dump the output of echo foo | /usr/sbin/exim -d root@devil... here after looking for private data in the debug output?
I am still running with Osamu Aoki's workaround (dc_other_hostnames='devilserver.deviltechnologies.com' in update-exim4.conf.conf).
[WORKAROUND ON] I executed the command, dumped the output and received the email message. The details are in "tests_with-workaround.tbz", attached.So I removed the workaround, executed update-exim4.conf, restarted exim, and executed echo bar | /usr/sbin/exim -d root@devil....
The details are in "tests_without-workaround.tbz", attached.To my surprise, now all is working as expected! The message does not get sent to the smarthost.
root@devilserver:/etc/exim4# exim -v -bt r...@devilserver.deviltechnologies.com
----- R: system_aliases for r...@devilserver.deviltechnologies.com R: system_aliases for dgue...@devilserver.deviltechnologies.com R: userforward for dgue...@devilserver.deviltechnologies.com R: procmail for dgue...@devilserver.deviltechnologies.com dgue...@devilserver.deviltechnologies.com <-- r...@devilserver.deviltechnologies.com router = procmail, transport = procmail_pipe -----So, I looked around and found that the router rebooted itself during the weekend. I started thinking of a DNS issue, so I tried to reproduce by generating a DNS problem in the network:
----- root@devilserver:/etc/bind# service bind9 stop Stopping domain name service...: bind9 waiting for pid 1437 to die. root@devilserver:/etc/bind# dig devilserver.deviltechnologies.com ; <<>> DiG 9.7.3 <<>> devilserver.deviltechnologies.com ;; global options: +cmd ;; connection timed out; no servers could be reachedroot@devilserver:/etc/bind# exim -v -bt r...@devilserver.deviltechnologies.com
R: system_aliases for r...@devilserver.deviltechnologies.com R: system_aliases for dgue...@devilserver.deviltechnologies.com R: userforward for dgue...@devilserver.deviltechnologies.com R: procmail for dgue...@devilserver.deviltechnologies.com dgue...@devilserver.deviltechnologies.com <-- r...@devilserver.deviltechnologies.com router = procmail, transport = procmail_pipe root@devilserver:/etc/bind# service exim4 stop Stopping MTA: exim4_listener. root@devilserver:/etc/bind# service exim4 start Starting MTA: exim4.root@devilserver:/etc/bind# exim -v -bt r...@devilserver.deviltechnologies.com
R: system_aliases for r...@devilserver.deviltechnologies.com R: system_aliases for dgue...@devilserver.deviltechnologies.com R: userforward for dgue...@devilserver.deviltechnologies.com R: procmail for dgue...@devilserver.deviltechnologies.com dgue...@devilserver.deviltechnologies.com <-- r...@devilserver.deviltechnologies.com router = procmail, transport = procmail_pipe root@devilserver:/etc/bind# -----??? It seems exim does not try to forward the mail to the smarthost, even without a DNS reply!!
Now, I purposefully misconfigured the local DNS server (bind).It is in forward-only configuration; all the requests are forwarded to the router. I entered a wrong IP address in /etc/bind/named.conf.options, then started bind:
----- root@devilserver:/etc/bind# service bind9 start Starting domain name service...: bind9. root@devilserver:/etc/bind# dig devilserver.deviltechnologies.com ; <<>> DiG 9.7.3 <<>> devilserver.deviltechnologies.com ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 39461 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 0 ;; QUESTION SECTION: ;devilserver.deviltechnologies.com. IN A ;; ANSWER SECTION: devilserver.deviltechnologies.com. 86400 IN A 192.168.200.249 ;; AUTHORITY SECTION:devilserver.deviltechnologies.com. 86400 IN NS devilserver.deviltechnologies.com.
;; Query time: 0 msec ;; SERVER: 127.0.0.1#53(127.0.0.1) ;; WHEN: Mon Apr 28 09:37:53 2014 ;; MSG SIZE rcvd: 81root@devilserver:/etc/bind# exim -v -bt r...@devilserver.deviltechnologies.com
R: system_aliases for r...@devilserver.deviltechnologies.com R: system_aliases for dgue...@devilserver.deviltechnologies.com R: userforward for dgue...@devilserver.deviltechnologies.com R: procmail for dgue...@devilserver.deviltechnologies.com dgue...@devilserver.deviltechnologies.com <-- r...@devilserver.deviltechnologies.com router = procmail, transport = procmail_pipe root@devilserver:/etc/bind# ----- Still, working as expected.I am now running with a correct DNS server configuration, and with a correct exim configuration (the same that worked since some years).
Now I'm out of ideas, too.
Greetings Marc
Thanks, Diego
tests_without-workaround.tbz
Description: Binary data
tests_with-workaround.tbz
Description: Binary data
