More 3.4 calendar woes

Homer Dokes hdokes at mail.inct.net
Tue Aug 18 05:26:21 CEST 2015


Hello,

I too am having this same issue as this link 
(http://lists.kolab.org/pipermail/users/2015-April/019163.html) and 
there seems to have been no definitive answer to it.

The symptom isn't happening to all kolab users within the domain and in 
fact... it is only happening to one user at this time and just recently 
manifested itself after 6 months of operation (no idea how).

The symptom is the user has lost all their contacts and calendar 
information.  Further investigation shows that the files still exist in 
the users cache directory but it appears as tho the cache itself has 
lost reference to them.  New events and contacts can be created and they 
do show up in the roundcubemail web client but the older entries do not 
show even tho the individual files exist under the users Calendar and 
Contacts directory on the server.    It should be noted that the new 
entries do sequence their file name properly after the existing files in 
the corresponding directories.

I have run /⁠usr/⁠share/⁠roundcubemail/plugins/libkolab/bin/modcache.sh 
clear -⁠u <user at domain.org> -h localhost which did not make a difference.

I have also found that the corresponding kolab_cache_event and 
kolab_cache_contact tables only have the 'new' entries in them 
pertaining to that user and that there was a butt load of entries in the 
kolab_folders table for that user.  When I say butt load... I mean 
thousands and thousands of entries for type Contact and Calendar with 
synclock values other than 0.  The thousands of entries were reduced to 
just one each for each type after running the above command for each 
user.  I notice the type entries for each user where added only upon 
each user logging back in to the client and then only after they went to 
the corresponding 'type' page. i.e. the user record for type Calendar 
was added back into the roundcube.kolab_folders only when the user 
opened the calendar page through their roundcubemail web client.  No 
where can I find ANY documentation otherwise to tell me it should do 
this.  As the Kolab package REQUIRES a developer to administer it's 
operation because it is NOT capable of doing so on it's own after the 
initial installation, it seems to me that having this information 
readily available on line would be highly desired.  Particularly for 
anyone that may have an ability to contribute to fixes and patches.  If 
I have missed this information some how.... please point me to a url 
that contains it.

I know from the users Contacts directory for example that there are 1486 
sequenced files (contacts) in there for this particular user however the 
kolab_folders Contact record is showing only 18 (newly added since the 
symptom presented itself) and that the kolab_cache_contact table shows 
only the new 18 contacts in it for this user.

It has been impossible for me to find any place on the internet that 
goes into detail on how all the files and the mysql tables relate... one 
of my biggest pet peeves of the Kolab email system.  Can ANYONE shine 
some light on how the kolab_cache_contact, or any other cache type 
folder can be reconstructed.  I have to believe this capability exists 
if all the original files still exist in the respective imap physical 
directories.

A side note.... to date... I have really regretted standardizing on the 
whole Kolab environment.  For my company and for our clientele. 
Information pertinent to the operation of this package is non existent 
other than how to install it.  For novices to the package, the mail 
system has been a very poor source of finding all ready answered 
questions to issues.  There doesn't seem to be one central repository 
for issues and solutions.  Email links like the one above seemingly go 
unanswered despite the request for help.  It begins to beg the question, 
is this a viable solution or just one's hobby?  I can think of a lot 
less frustrating hobbies to get into... like... golf as an example.... 
and I have no use for golf.

All that said.... being a developer for over 30 years I can see the 
potential in this package and still desire to give it a go. Please... 
somebody.... give me a reason to continue to believe that.

Thanks for any assistance on this matter and indulging me through my rant.


More information about the users mailing list