dovecot as alternative

[Mihai Badici]

On Friday 31 May 2013 10:42:26 Henning H wrote:
> Essentially kolab just works with any saslauth-capable imap-server or am
> I wrong?

Yes but in order to store additional info's ( address book, calendar) you need 
metadata support in IMAP server. Without it will work, but only as a classical 
mail server
> 
> The IMAP-Server in the kolab environment only works as normal
> imap-service. I think you would not have to configure additional things.
> 
> 1st: Just make dovecot run with you 389-ds and there you go.
> 
> 2nd: Remember all necessary steps and write some python-routines for
> setup-kolab (or kolab-setup in the future).
> 
> Henning

When installing kolab, we need to add the access account to 389-ds and the 
corresponding dn in dovecot configuration files.

In fact, in my openldap setup I used before meet kolab I created an root dn ( 
dc=machinet") and put all data in.
Like this, I need to change nothing in dovecot config.
Well, it is true your dn will be "cn=john Doe, dc=kolab,dc=com,dc=machinet" 
but I don't care. Probably searches will be a little slower but I don't think 
is significant.




> Am 31.05.2013 09:37, schrieb Christian Hügel:
> > This is a very interesting topic. We use dovecot as well and I´d love to
> > see if there is a way to make it run with kolab 3.
> > 
> > Chris.
> > 
> > Am 31-05-2013 9:18, schrieb Mihai Badici:
> >>> It would be quite easy to allow dovecot as an alternative imap
> >> 
> >> server.
> >> 
> >>> I'm pretty sure setup-kolab doesn't work with dovecot, and to the
> >> 
> >> best
> >> 
> >>> of my knowledge setup-kolab does not check whether or not cyrus is
> >>> 
> >>> installed, it just assumes that it is. Which is correct with the
> >> 
> >> current
> >> 
> >>> installation dependencies.
> >>> 
> >>> 
> >>> 
> >>> So it gets a little bit more complex. I'm thinking we could have
> >> 
> >> kolab
> >> 
> >>> depend on an imap server (either cyrus, which would be the default,
> >> 
> >> or
> >> 
> >>> dovecot, or anything else). In combination, we could move the
> >> 
> >> dependency
> >> 
> >>> on cyrus to the kolab-conf package, so cyrus would be required for
> >>> 
> >>> setup-kolab - at least for now. The end result being that you can
> >>> 
> >>> install any old imap server, but if it isn't cyrus, you'd be on your
> >> 
> >> own
> >> 
> >>> in terms of setting up your server.
> >> 
> >> If Cyrus will be required by setup-kolab I can install all kolab
> >> metapackage, run setup-kolab and eventually remove setup-kolab and
> >> cyrus.
> >> 
> >> After that, i can install dovecot. I will manage with configuration,
> >> in fact i will ned only one or two templates. If I will install more
> >> than two servers probably I will wrote a pach for setup-kolab :)
> >> 
> >>> If you could file a bug, we could work against that to see what can
> >> 
> >> be
> >> 
> >>> done.
> >>> 
> >>>> Well, I could compile from source, but I already done that for
> >>>> 
> >>>> Slackware, on Debian it is supposed to be easy :)
> >>> 
> >>> It is, provided the package maintainer makes it easy :-D.
> >> 
> >> Well, everybody agree, Debian is great for production use.
> >> 
> >> But when you try to abandon the "right way" it's problematic. That's
> >> way I use slackware who let me broke my system if I want to .. and
> >> sometime even if I don't want to :)
> >> 
> >> So I will fill a bug ( I'm not sure where, because is not a kolab
> >> issue but a package one) and perhaps will have a nice result at the
> >> end.
> >> 
> >>> Cheers,
> >>> 
> >>> 
> >>> 
> >>> Paul
> > 
> > _______________________________________________
> > Kolab-users mailing list
> > Kolab-users at kolab.org
> > https://www.intevation.de/mailman/listinfo/kolab-users
> 
> _______________________________________________
> Kolab-users mailing list
> Kolab-users at kolab.org
> https://www.intevation.de/mailman/listinfo/kolab-users
-- 
Mihai Badici
http://mihai.badici.ro
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.kolab.org/pipermail/users/attachments/20130531/9e8159b7/attachment.html>