On 23 May 2017, at 13:43, Cecil Westerhof wrote:
2017-05-23 19:30 GMT+02:00 Viktor Dukhovni
<postfix-us...@dukhovni.org>:
On May 23, 2017, at 1:22 PM, Cecil Westerhof
<cldwester...@gmail.com>
wrote:
There are still some things I should look after, like:
warning: unable to create missing queue directories
But it is probably not the most urgent.
Actually, that *is* your most urgent problem. Without these
Postfix cannot work, regardless of the rest of the configuration
The $queue_directory must have various requisite subdirectories
with correct ownership. If the queue_directory has POSIX ACLs
or some other barrier to creation of the required sub-directories,
then Postfix is unusable.
I do not understand. It looks OK to me.
The queue directory is /var/spool/postfix and the way to it looks
OK (to
me).
drwxr-xr-x 31 root root 4.0K May 22 16:47 /
drwxr-xr-x 13 root root 4.0K May 22 12:03 /var
drwxr-xr-x 10 root root 4.0K May 22 13:59 /var/spool
drwxr-xr-x 20 root root 4.0K May 23 16:53 /var/spool/postfix
The directory itself looks also OK to me:
drwx------ 2 postfix root 4.0K May 23 19:10 active
drwx------ 2 postfix root 4.0K May 23 18:54 bounce
drwx------ 2 postfix root 4.0K May 22 14:00 corrupt
drwx------ 3 postfix root 4.0K May 23 17:36 defer
drwx------ 3 postfix root 4.0K May 23 17:36 deferred
drw-r--r-- 2 postfix root 4.0K May 23 17:10 dev
drwxr-xr-x 3 root root 4.0K May 23 19:10 etc
drwx------ 2 postfix root 4.0K May 22 14:00 flush
drwx------ 2 postfix root 4.0K May 23 16:39 hold
drwx------ 2 postfix root 4.0K May 23 19:10 incoming
drwxr-xr-x 3 root root 4.0K May 23 17:10 lib
drwx-wx--T 2 postfix postdrop 4.0K May 23 19:10 maildrop
drwxr-xr-x 2 root root 4.0K May 23 17:36 pid
drwx------ 2 postfix root 4.0K May 23 19:10 private
drwx--s--- 2 postfix postdrop 4.0K May 23 19:10 public
drwx------ 2 postfix root 4.0K May 22 14:00 saved
drwx------ 2 postfix root 4.0K May 23 16:39 trace
drwxr-xr-x 3 root root 4.0K May 23 16:53 usr
Or am I missing something?
Looks wrong to me. The presence of dev, etc, lib, and usr directories
implies that you are running some part of Postfix chrooted to that
directory. That would be consistent with processes looking for
directories and not finding them. Don't chroot Postfix. Fix master.cf
and you should eliminate your issues.
Note: The Debian customized Postfix has been packaged in the past (and
may still be) to run services chrooted. I assume this means that it CAN
work that way, but the place to seek help with HOW that works would be
the Debian wizards who decided to package it that way.
