Kolab 3.2 Freebusy

Paul Boddie paul at boddie.org.uk
Sun Mar 2 01:23:23 CET 2014


On Thursday 27. February 2014 20.30.33 Brady, Mike wrote:
> On 2014-02-28 06:48, Christian Hügel wrote:
> > Nevermind, found it. But I get a buch of errors here:
> > 
> > 
> > (18:45:13) findkolabfoldersjob.cpp(111):        "GetMetaData failed,
> > server replied: A000111 BAD Missing arguments to Getmetadata "
> > (18:45:13) findkolabfoldersjob.cpp(111):        "GetMetaData failed,
> > server replied: A000112 BAD Missing arguments to Getmetadata "
> > (18:45:13) findkolabfoldersjob.cpp(111):        "GetMetaData failed,
> > server replied: A000113 BAD Missing arguments to Getmetadata "
> > (18:45:13) findkolabfoldersjob.cpp(111):        "GetMetaData failed,
> > 
> > Looks like a bug: https://issues.kolab.org/show_bug.cgi?id=2027

I found a duplicate here...

https://issues.kolab.org/show_bug.cgi?id=2862

...presumably newly created because the above issue was supposedly fixed but 
has obviously come back to trouble us once again.

> I experienced the same issue with 3.2 and just ended up disabling the
> freebusyd cron job. Hardly a solution I know, but this problem was
> putting so much load on my test server with a continual stream of
> failures that it was impacting my ability to do anything else.

That's another issue, I think. I found that the daemon would repeatedly 
reconnect, but recent libcalendaring code seems to fix that even when getting 
the usual "The certificate is self-signed, and untrusted" error.

Indeed, the Debian packages (and presumably others) provide rather out-of-date 
code, and I think there's something to be said for just rebuilding them with 
newer code in order to shake off some of the problems being experienced here.

> I am assuming that is due to changes in Cyrus IMAP 2.5?

Cyrus 2.5 seems to support the METADATA extension to IMAP and libcalendaring 
will use that in preference to the ANNOTATEMORE extension. It appears that 
when confronted with METADATA support, libcalendaring comes up rather short 
and cannot formulate commands that Cyrus wants to understand. I imagine that 
somehow forcing use of ANNOTATEMORE (GETANNOTATION commands) with 2.5 would 
probably restore the working state as seen in 2.4, but I've submitted a patch 
to bug #2862 that should at least get libcalendaring working with Cyrus again.

It was quite a tour of Kolab's C++ library stack and involved a lot of 
rebuilding, but I got there in the end.

Paul


More information about the users mailing list