jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jared Graber" <jared.gra...@IACcap.com>
Subject RE: Jackrabbit Caching?
Date Thu, 03 Jul 2008 14:20:25 GMT
Well I have 2 environments:
	The first one has an admin application that manages the data and
uses a Oracle bundle persistence manager.  Client applications connect
(only via a trigger of some kind) to the database (yes I know this is
not recommended) repository using the oracle persistence manager in a
read only capacity to export the data into a local cache (in memory
repository in this case).
	The other environment is completely read only.  The repository
data is stored in an oracle database that is periodically replaced by a
backup of the database used in the 1st environment.  There is no admin
in this case, but the clients connect in the same manner.

Because of database restores, I could not just use RMI with JCR because
I would then have to restart the RMI server application(s) which would
be less than ideal.  

The solution I have come up with, seems to be working well.  There are
no caching issues and since the clients are read only, I don't think
they can affect the repository.


-----Original Message-----
From: Alexander Klimetschek [mailto:aklimets@day.com] 
Sent: Thursday, July 03, 2008 7:29 AM
To: users@jackrabbit.apache.org
Subject: Re: Jackrabbit Caching?

I am still wondering in which way you actually replace the repository
data? Which persistence manager do you use? Just curious... ;-)

Regards,
Alex

On Thu, Jul 3, 2008 at 2:30 AM, Jared Graber <jared.graber@iaccap.com>
wrote:
> Found a solution for my use case:
> Use a transient repository.  Since I rarely connect to the repository,
I
> can take the performance hit of initialization and shutdown every time
I
> need to export the data.  On shutdown the cache is cleared so I always
> get the current data in the repository.
>
> -Jared
>
> -----Original Message-----
> From: Jared Graber [mailto:jared.graber@IACcap.com]
> Sent: Wednesday, July 02, 2008 1:00 PM
> To: users@jackrabbit.apache.org
> Subject: RE: Jackrabbit Caching?
>
> Thanks for the info.  I'm not using webdav though.
>
> I'm still looking for a way to turn off caching (preferably) or clear
> the cache.
>
> I found that setting the max sizes (setMaxMemory and
> setMaxMemoryPerCache) in CacheManager to zero doesn't turn off caching
> because the values are not enforced when new caches are added (or when
> the size is set after a cache exists already).  Those values only come
> into effect on a cache resize which can't be triggered directly.
>
> -Jared
>
>
> -----Original Message-----
> From: Conoly, Brett [mailto:Brett.Conoly@digitalinsight.com]
> Sent: Wednesday, July 02, 2008 12:04 PM
> To: users@jackrabbit.apache.org
> Subject: RE: Jackrabbit Caching?
>
> I don't know how you're actually retrieving the content but I had this
> problem retrieving data from webdav a while back.  I did learn that
the
> headers sent back by webdav are not updated unless you set the
> JCR_LASTMODIFIED property of the content node.  This prevented the
> caching for me, hope this helps.
> Brett
>
> -----Original Message-----
> From: Jared Graber [mailto:jared.graber@IACcap.com]
> Sent: Wednesday, July 02, 2008 8:55 AM
> To: users@jackrabbit.apache.org
> Subject: RE: Jackrabbit Caching?
>
>
> I understand that I shouldn't modify the underlying database directly,
> but I need to deal w/ database restores.
> So is there a way to turn off caching?
>
> -Jared
>
> -----Original Message-----
> From: Jukka Zitting [mailto:jukka.zitting@gmail.com]
> Sent: Tue 7/1/2008 5:32 PM
> To: users@jackrabbit.apache.org
> Subject: Re: Jackrabbit Caching?
>
> Hi,
>
> On Wed, Jul 2, 2008 at 12:13 AM, Jared Graber
<jared.graber@iaccap.com>
> wrote:
>> Is the in-memory cache preventing me from getting the new data?
>
> Only if you modify the underlying database directly, which you should
> never do.
>
> If you only access the content through the JCR API, Jackrabbit will
> automatically make sure that all caches are correctly invalidated and
> search indexes updated whenever content changes.
>
> Furthermore, since the JCR-RMI layer does no caching you can also use
> that without worrying about cache coherence issues.
>
> BR,
>
> Jukka Zitting
>
>



-- 
Alexander Klimetschek
alexander.klimetschek@day.com

Mime
View raw message