[Kolab-devel] [issue1779] kolab 2.1 private/kolabpolicy: Socket operation or non-socket

[ComCept Net GmbH Andrea Soliva]

Hi all

Now I found a solution but this solution is probably dirty. Troubleshooting
has brought up the issue meaning in main.cf.template is defined following:

## Kolab Policy Server
smtpd_recipient_restrictions = permit_mynetworks, permit_sasl_authenticated,
        reject_unauth_destination, reject_unlisted_recipient,
        check_policy_service unix:private/kolabpolicy
smtpd_sender_restrictions = permit_mynetworks,
        check_policy_service unix:private/kolabpolicy
kolabpolicy_time_limit = 3600
kolabpolicy_max_idle = 20

Here is the issue meaning after successful authentication the request would
be forwarded/hand over to unix:private/kolabpolicy and here we have the
issue that postfix can not talk to privat/kolabpolicy if the user comes from
outside world (funny that it works if the users comes from trusted network).
Probably this has something to do with the implemenation of zones under
Solaris 10. I changed also the position to inet but in this case all users
are not anymore able to deliver mails. At least I fully commented out the
positions meaning:

## Kolab Policy Server
smtpd_recipient_restrictions = permit_mynetworks, permit_sasl_authenticated,
        reject_unauth_destination, reject_unlisted_recipient
#        check_policy_service unix:private/kolabpolicy
smtpd_sender_restrictions = permit_mynetworks
#        check_policy_service unix:private/kolabpolicy
kolabpolicy_time_limit = 3600
kolabpolicy_max_idle = 20

If this would be done all users including the outside world user can deliver
mails through kolab2. I do not know what private/kolabpolicy is responsible
for etc. but I can imagine that from security point of view it would be
worse. Can somebody explain me what this position does? Is there a way to
change this meaning that unix:private/kolabpolicy would be handled in
another way?

In the meantime I will proceed with 2.2.0beta1 but I'm pretty sure I have
also with this version the issue. 

Any help really appriciated.


Andrea






Bernhard Reiter <bernhard at intevation.de> added the comment:

Thomas, Steffen,    
are we sure that this is working on non-solaris machines?   
What would be the next debugging step for Andrea?

----------
assignedto:  -> thomas
nosy: +bernhard, steffen, thomas, wilde
status: unread -> chatting
topic: +server
________________________________________________
Kolab issue tracker <kolab-issues at intevation.de>
<https://intevation.de/roundup/kolab/issue1779>
________________________________________________