Question wrt. dovecot replicator
remko at FreeBSD.org
Thu Oct 30 18:55:14 UTC 2014
> On 30 Oct 2014, at 17:31, Jan Behrend <jbehrend at mpifr-bonn.mpg.de> wrote:
> On Fri, 2014-10-10 at 14:45 +0200, Jan Behrend wrote:
>> On Fri, 2014-10-03 at 23:37 +0200, Remko Lodder wrote:
>>>>>> How can I determine why there are duplicated emails?
>> Same problem here!
>>>>>> What kind of messages should I specifically look for?
>>>>> Look for any errors and warnings in the Dovecot log. You could also enable mail_debug (ref.
>>>>>> Can I set this up for a few selected accounts instead of all accounts like it was currently? To make sure I do not make things worse for others then needs to be :-)
>>>>>> The service had been disabled for the time being to prevent the other users from getting duplicated emails.
>>>>> I do not know what kind of userdb you are running, but there is a newish patch that enables per user replication via the
>>>>> mail_replica setting. It is not yet included in the newest (2.2.13) release of Dovecot, but is available via the enterprise
>>>>> version. There are no FreeBSD builds for that, though. ref: http://hg.dovecot.org/dovecot-2.2/rev/c1c67bdc8752
>>>> my userdb consists of local users (Which are fed through LDAP at the backend). perhaps I can setup a mailAttributes setting or something so that the replica can be
>>>> set, although I prefer that I have control over that in the config itself :-)
>> With the latest Debian jessie version 1:2.2.13-5 you can actually have a
>> per user mail_replica setting taken from a (LDAP) directory. This keeps
>> the duplicate mail issue away from other users but a few willing to
>> test ...
>>> For what it’s worth:
>>> replication_dsync_parameters = -f -d -N -l 30 -U
>> I read in Peer Heinlein's Dovecot book
>> that replicating a public namespace gives you trouble :-(. So keep the
>> "-N" option away for now.
>> However I would like examples for the "-n" and "-x" options, which are
>> neither given in the wiki nor in the nonexistant man page.
>> I think the replication feature is very, very cool, but right now it
>> gives me a hard time to implement flawlessly ;-)
>> Thanks for any help or light shed on this issue ...
> Found it ;-)
> All working beautifully now!
For what it’s worth it is working fine here with debugging enabled. I will upgrade to 2.2.15 and keep testing that to see how that works.
Thanks for the help and suggestions so far!
> Cheers Jan
> MAX-PLANCK-INSTITUT fuer Radioastronomie
> Jan Behrend - Rechenzentrum
> Auf dem Huegel 69, D-53121 Bonn
> Tel: +49 (228) 525 359, Fax: +49 (228) 525 229
> jbehrend at mpifr-bonn.mpg.de http://www.mpifr-bonn.mpg.de
More information about the dovecot