Re: [Kolab-devel] event triggered freebusy data generation
Paul Boddie <[email protected]>
| Newsgroups | gmane.comp.kde.devel.kolab |
|---|---|
| Message-ID | <201402252306.04999.paul__14817.1866000213$1393365076$gmane$org@boddie.org.uk> |
On Tuesday 25. February 2014 20.34.27 hede wrote: > Hi, > > I asked for documentation to the kolabs freebusy system on the users list. > If it is really needed to run a cron job every x minutes to generate the > data (kolab-freebusyd). I wondered similar things a while ago and wrote a blog post that briefly mentioned the generation of free/busy data amongst other things: http://blogs.fsfe.org/pboddie/?p=606 It probably doesn't explicitly mention the cron job that runs the daemon, though. > I didn't find any configuration option to let the freebusy data get > generated on demand and because no one answered, I implemented it myself, > in "my way. (The "my way" means I expect it to be an ugly hack and not > accepted upstream. I'm pretty sure there are better ways to implement > this. At least my chosen names will hurt any english native speaker...) > > Instead of running kolab-freebusyd every x minutes, for my use-case it's > better to start it on demand. So I do so. The freebusy web script triggers > kolab-freebusyd to regenerate and aggregate the data for the corresponding > user, if needed, e.g. if the cached ifb-file is too old. If it's not too > old, the old one gets used and no kolab-freebusyd is triggered. This way > even on high load systems it won't hurt to use this (with a triggertime of > 900 it would be like a regular cron job every 15 minutes). That sounds like a nice idea, and it's probably a common convention in the Web applications world. I know that I do similar stuff to cache remote resources in a Web application component I've written (that happens to also have something to do with calendars). As for the daemon, I did wonder whether the data was being regenerated for all users every time, but I didn't get round to investigating what it actually does. However, I imagine that the eventual aim is to trigger free/busy updates via an event delivery mechanism upon storing or updating an object in the IMAP store. [...] > Everything using the "kolab-freebusy" web service should trigger the > generation. I tested roundcube and kontact so far. Clients using the imap > Freebusy-folder won't trigger anything. That's the drawback. > > Btw: Why differ between aggregate and regenerate? And why generate inside > imap and not directly to /var/lib/kolab-freebusy/? Are there plans to get > rid of the ifb-files and use imap instead? I wonder what the CalDAV infrastructure does. Perhaps it's worth looking at that to see. Paul