On Fri, Jun 12, 2015 at 07:27:51PM +0300, Liutauras Adomaitis wrote:
> > > dict_ldap_get_values[1]: search returned 24 value(s) for requested
> result
> > > attribute associateddomain
> > > 2015-06-10T15:29:32.267793+02:00 mx05 postfix/smtpd[28480]:
> > > dict_ldap_get_values[1]: Leaving dict_ldap_get_values
> > > 2015-06-10T15:29:32.267798+02:00 mx05 postfix/smtpd[28480]:
> > > dict_ldap_lookup: Search returned <the list of mydestinations>
> >
> > What useful query would return a list of 24 domains? You're doing
> > something wrong perhaps? Your configuration is a mystery, so can't
> > help there.
>
> The config works, except one IP which has problems.
My instinct is nevertheless that something is wrong here. The only
way this makes sense is if you created a single LDAP object with
a multi-valued attribute that holds all your local domains, rather
than one object per domain:
dn: <object_dn>
associateddomain: example.com
associateddomain: example.net
associateddomain: example.org
...
and your configuration has something along the lines of:
main.cf:
mydestination = ldap:/some/mydest.cf
mydest.cf:
search_base = <object_dn>
scope = base
query = associateddomain = %s
result_attribute = associateddomain
which has the side-effect of returning all the domains when matching
any one of them. This is a bit ugly. Separate objects would be
cleaner.
>
> The main point of pasting logs was to show the gap of 30s. Every debug
> log output has the gap of 30s. It can't be just a coincidence.
Not a coincidence for sure, but nothing to do with any quota on
logging. Most likely an actual delay in communication between the
SMTP client and server. If logs are lost, that would be the result
of syslog protocol issues, not Postfix settings.
--
Viktor.
