jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From euangreen <euangr...@gmail.com>
Subject Re: reading from repository quickly
Date Thu, 25 Jun 2009 13:39:16 GMT

Hi,

I'm trying various combinations of PersistenceManager , but have had no luck
getting them to actually run. I keep getting a file lock error when trying
to change from SimpleDBPersistenceManager to BundleDbPersistenceManager. But
will persevere and look into changing the cache size to see if it makes a
difference.

I'm using Session.exportDocumentView() to read the information from
repository into an xml object then into an object in the java app. This
certainly seems from profiling where the main problem is. 

I'm trying to speed this up but not sure how basically

Any ideas?

Euan




Marcel Reutegger wrote:
> 
> I'd say increase the bundleCacheSize in the persistence manager
> configuration section of your workspace.xml and soon the reads from
> the repository will be served entirely from cache.
> 
> are you sure the repository is the bottleneck? Do you have profiling
> data? XML transformations are usually also expensive.
> 
> speaking of XML, what exactly do you mean, when you say:
> 
>> reading the item info into java by way of an xml stream.
> 
> is that a Session.exportDocument/SystemView()?
> 
> regards
>  marcel
> 
> On Wed, Jun 24, 2009 at 16:51, Euan Green<euan@code9.co.uk> wrote:
>> Hi,
>>
>> What is the best/quickest way to display items from the repository on a
>> web
>> page?
>> I have an app that needs to display approx 250 items (using a list of
>> UUID's) from a repository containing 10000 items. At the moment we
>> display
>> the list by locatingthe item in the repository by the UUID, then reading
>> the
>> item info into java by way of an xml stream. Then we use the details to
>> populate a jsp. The whole process for displaying approx 250 items seems
>> very
>> long. The only way we can speed things up is to use our own cache. I was
>> wondering if there is a better more effeicent way of querying jackrabbit
>> for
>> specific uuids and getting that info displayed on a web page?
>>
>> Many thanks
>>
>> Euan
>>
> 
> 

-- 
View this message in context: http://www.nabble.com/reading-from-repository-quickly-tp24199279p24203457.html
Sent from the Jackrabbit - Dev mailing list archive at Nabble.com.


Mime
View raw message