pykolab smtp access policy

Paul Boddie paul at boddie.org.uk
Tue Apr 29 21:56:33 CEST 2014


On Tuesday 29. April 2014 20.41.08 Thomas Spuhler wrote:
> On Saturday, December 21, 2013 04:44:47 PM you wrote:
> > It looks like some unconfigured caching stuff. You find some generic
> > cache db user settings in kolab.conf
> > 
> > Afaik it shouldn't be causing any errors. Maybe there's a way to disable
> > SMTP policy cache or configure caching right.

setup-kolab will set this up, I believe:

http://git.kolab.org/pykolab/tree/pykolab/setup/setup_mta.py#n325

However, it may not set it up completely.

> > > Am 21.12.2013 um 16:23 schrieb Thomas Spuhler 
<thomas.spuhler at btspuhler.com>:
> > >> On Saturday, December 21, 2013 11:19:32 AM Christian Hügel wrote:
> > >> Hi,
> > >> 
> > >> i get a bunch of errors in /var/log/kolab/pykolab.log
> > >> 
> > >> 2013-12-19 20:21:00,520 pykolab.smtp_access_policy ERROR Operational
> > >> Error in caching: (OperationalError) (1045, "Access denied for user
> > >> 'user'@'localhost' (using password: YES)") None None

[...]

> > >> Does anyone know where these errors come from?

[...]

> > > I have the same errors in my log. Would help if we could fix this.
> 
> Any progress on this?

Well, as you can see from the defaults in /etc/kolab/kolab.conf, it's likely 
that this is just a case of the policy credentials not being set up:

http://git.kolab.org/pykolab/tree/conf/kolab.conf#n273

So, 'user'@'localhost' is what the kolab_smtp_access_policy program sees in 
the configuration and thus uses to connect to MySQL.

I'm actually not sure what the procedure is to set this up properly, but there 
is a mention of it in the documentation:

http://docs.kolab.org/architecture-and-design/mta.html#and-mta-kolab-smtp-
access-policy

And there's some kind of description of its Postfix-related configuration:

http://docs.kolab.org/howtos/build-kolab-from-
source.html?highlight=smtp%20access%20policy#kolab-smtp-access-policy

But nothing I could instantly find about the MySQL part. Maybe I'll have to 
read the code.

Paul


More information about the users mailing list