[kronolith] Event modification are not synced to remote device using CALdav if horde cache enabled

[Giorgio Paolucci]

  Thanks Jan,
I'm starting digging into this....

Bests
Giorgio
Quoting Jan Schneider <jan at horde.org>:

> Zitat von Giorgio Paolucci <giorgio.paolucci at unipd.it>:
>
>> Quoting Jan Schneider <jan at horde.org>:
>>
>>> Zitat von Giorgio Paolucci <giorgio.paolucci at unipd.it>:
>>>
>>>> Hello everybody,
>>>> I have a horde webmail installation 5.2.11 on a clustered debian
wheezy
>>>> platform.
>>>> Horde is installed on a 4 node cluster (2 clusters of 2) and clients
>>>> are
>>>> balanced on a round robin DNS basis; to keep sessions sticky, each
node
>>>> is
>>>> configured with an apache reverse proxy that forwards (based on cookie
>>>> values) requests to apache2 backend instances listening on non
standard
>>>> ports on each node. Php caching system APC is installed.
>>>>
>>>> If I enable horde cache (local file system on each node), users who
>>>> subscribe calendars through CALdav don't receive updates on their
>>>> smartphones, or better:  they receive new events or deletion of
events,
>>>> but not modifications of existing events made through horde web
>>>> interface.
>>>> Apache logs show that caldav clients sessions are sticky and each
>>>> smarphone
>>>> query always the same node of the cluster.
>>>>
>>>> With horde cache disabled, everything works fine. Any clues to
>>>> investigate
>>>> this issue?
>>>
>>> This observation doesn't make much sense to me, because we don't use
any
>>> caching in the CalDAV code or in Kronolith, except for external
calendar
>>> resources like remote calendars of free/busy information.
>>
>> Hi Jan,
>> that sounds odd to me too, also because the issue is repeatable: you
>> enable
>> horde cache and event modification are not caught bye the phone, you
>> disable horde cache and events are caught again....
>> Css and js can be cached without any issues.
>> So, according to you there aren't possibly other implications in
>> activating
>> horde cache....
>
> Correct. And needless to say, that nobody else has reported this so far,
> and I cannot reproduce it either.
>
> --
> Jan Schneider
> The Horde Project
> http://www.horde.org/
>
> --
> kronolith mailing list
> Frequently Asked Questions: http://wiki.horde.org/FAQTo unsubscribe,
> mail: kronolith-unsubscribe at lists.horde.org