debugging exim(4)

[graemev2]

I had a issue with exim4. I asked a number of question of the exim4 mailing list, several of which ended up being related to debugging, which I'd already tried . The upshot was all the clever debugging options they suggested and ones I made up, did not work.

Many resulted in no debug, one ...killing exim and starting by hand with debug options on the command line produced a little debug then stopped (it looked as if it had forked of a child to do the and child was not talking debug, or maybe had stderr pointing elsewhere).

Anyhow their advice was "ask on the Debian forum!" , so here I am:

So it's a question of packaging, processes are kicked off by systemd (cron?) etc. They have defaults in /etc/defaults and of course the Debian script update-exim4.conf(8) . So ...

What is a good/typical debug workflow for exim4 ?

Now I have a very small system and can afford to just stop the exim service and have debug in the real files but I guess in terms of an answer it would be good to consider also a big (production?) system.

If it helps with a "use case" what I was trying to debug (for several days, but now in the past) was. I was getting failures back from smarterhost (bad email addresses somewhere) but it was not clear what "line" they were objecting to. It became evident the mails in question had gone via a users ~/.forward...so I could not be sure what was in the data sent. So I did not know what the issue was (from smarthost) and I did not know what I had sent. Ideally what I would have like to have seen was the dialogue between the two mail systems (e.g SMTP traffic)

[dilberts_left_nut]

What does the log say?
What extra info do you think you'll need from a debug log?

[graemev2]

>What does the log say?
Sorry , say about what? It has stuff like this

Code: Select all

2024-03-07 12:33:01.568 [1205489] cwd=/var/spool/exim4 2 args: /usr/sbin/exim4 -q
2024-03-07 12:33:01.573 [1205489] Start queue run: pid=1205489
2024-03-07 12:33:01.573 [1205489] End queue run: pid=1205489
2024-03-07 13:03:01.569 [1210188] cwd=/var/spool/exim4 2 args: /usr/sbin/exim4 -q
2024-03-07 13:03:01.574 [1210188] Start queue run: pid=1210188
2024-03-07 13:03:01.574 [1210188] End queue run: pid=1210188

Do you mean, when I tried their various debug options ? Well nothing changed in the logs ... I don't believe that is where the debugging info goes. As I said the only debug I managed to get out came via stderr/stdout and that was just "startup" info.


>What extra info do you think you'll need from a debug log?
Well the guys in the exim4 mailing list suggested various options. Right now I don't need anything ... I spent about 6 days of 12 hours working on its and by brute force found the problem. What I'd like to do is to be able to try their suggestions .


But this is not my question. My question is HOW do people debug exim4 on Debian .... e,g

• 1: I add flags to /etc/default/exim4
  2: I add options in /etc/exim4/update-exim4.conf.conf and do a update-exim4.conf(8)
  3: edit files in /etc/exim4/conf.d/ and run update-exim4.conf(8)
  4: Kill the exim4 process and start by hand adding debug options
  5: ....

The guys on the exim4 mailinglist can offer advice on what options I can use to debug , but don't know how to apply these in a Debian environment.

[dilberts_left_nut]

Was there any traffic during that log snippet?
I would expect to see log lines showing mail passing through, with addresses, server responses etc.
Sorry I'm not an Exim user (postfix is the gods chosen MTA).

[Aki]

Moved to "General Question" sub-forum.