LMTP with unix socket (core 8.7.8, Debian 9.6)
-
The directory appears to be still unreadable by postfix: there is no “+” shown for the ACL that was supposedly set.
root@srv3:/var/log/kopano# ls -al /var/spool/kopano/ total 8 drwx------ 2 kopano kopano 4096 Nov 29 19:44 . drwxr-xr-x 6 root root 4096 Nov 29 19:09 .. srw-rw-rw- 1 root root 0 Nov 29 19:44 dagent.sock
The other idea is that, because postfix reports “No such file or directory”, that it might be running in some kind of mount namespace. You can use
nsenter -r -m -w -t P
with the postfix pid P to get into a shell that shows what postfix sees. -
Thank you for your help.
Ok, I fixed ACLs (my fault). But still no luck. :(
Directory has full access (777) now (for testing):root@srv3:/var/spool/kopano# ls -al total 8 drwxrwxrwx+ 2 kopano kopano 4096 Nov 30 21:56 . drwxr-xr-x 6 root root 4096 Nov 30 21:37 .. srwxrwxrwx 1 root root 0 Nov 30 21:56 dagent.sock
Tried also with nsenter - spawned shell from postfix PID and directory is accessibile & readable under postfix PID.
Will try to play around with this over the weekend. -
For my part, I ran the commands as listed in the documentation on Ubuntu 18.x, and the socket file NEVER gets created. I had to revert to *:2003 config style.
Whether using “run_as” kopano or root, the socket file never gets created.
Any details on how to make that happen?
-
Why not…
dpkg-statoverride --add kopano postfix 770 /var/spool/kopano
and you on Debian with systemd.see: systemctl cat kopano-dagent
fix it with : systemctl edit kopano-dagent
add:
[Unit] # optional, but helps is the start up process. After=kopano-server.service Wants=kopano-server.service [Service] LimitNOFILE=8192:16384 User=kopano Group=kopano
and try again
-
I have the same problem as mapo.
Dec 30 20:32:07 servername postfix/error[11706]: DB1844C1F50: to=<max@mustermann.de>, relay=none, delay=0.01, delays=0.01/0/0/0, dsn=4.4.1, status=deferred (delivery temporarily suspended: connect to servername.server.de[/var/spool/kopano/dagent.sock]: No such file or directory)
The problem is not fixed after the suggested doings from thctlo.
-
Ubuntu may be using AppArmor.
-
@jengelh and what do you want to tell me with this information?
I’am using Debian 9 as mapo. -
Well, that the LSM may interfere? Investigation needs shells, since no one has crystal balls (and won’t for the—ha—foreseeable future!)
-
Hi,
similar problem on ubuntu 18.04
The postfix/lmtp process runs chrooted in /var/spool/postfix and can therefore not access the socket in /var/spool/kopano.
my solution:
create the kopano spooldir in /var/spool/postfix instead of /var/spool (same chown, chmod and setfacl as in original instructions apply)in dagent.conf:
lmtp_listen = unix:/var/spool/postfix/kopano/dagent.sock
In postfix/main.cf:
virtual_transport = lmtp:unix:/kopano/dagent.sock
worked for me…
-
similar problem on ubuntu 18.04
The postfix/lmtp process runs chrooted in /var/spool/postfix and can therefore not access the socket in /var/spool/kopano.Yep. Just as I postulated on 2018-11-30:
The other idea is that, because postfix reports “No such file or directory”, that it might be running in some kind of mount namespace.
-
@pks57 said in LMTP with unix socket (core 8.7.8, Debian 9.6):
/var/spool/postfix/kopano/dagent.sock
I have the same problem with a fresh install on Ubuntu 18.
connect to kopano[/var/spool/kopano/dagent.sock]: No such file or directory
So as suggested, I created:
/var/spool/postfix/kopano/
And set permissions and facl, and updated main.cf and dagent.cfg etc etc and rebooted.
The dagent.sock file is never created and postfix now errors with:
connect to kopano[/var/spool/postfix/kopano/dagent.sock]: No such file or directory
Any suggestions ??
-
Look at the dagent log and see what it says about dagent.sock.