jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eugeny N Dzhurinsky <b...@redwerk.com>
Subject performance question
Date Wed, 28 Nov 2007 17:54:31 GMT
Hello everybody!

We are trying to adapt JackRabbit to our application which needs to store
large set of nodes, we already implemented it using basic repository settings,
however we are not satisfied with performance and we would like to know what
are the ways to configure JackRabbit to achieve maximum performance. 

Due to the nature of our application we have a set of nodes which form tree
with such structure:

root node
- content
-- domain object name
--- date
----- data node 1
----- data node 2
----- data node 3
----- data node 4
--- date
----- data node 1
----- data node 2
----- data node 3
----- data node 4
-- domain object name
--- date
----- data node 1
----- data node 2
----- data node 3
----- data node 4

So in general the entire tree has several domain nodes, and each domain node
hash set of a nodes which hold the content of domain object for given date,
and each date node has a lot (up to several millions) of child nodes.

Each child node has several simple properties and a large BLOB associated with
it.

We need to:

1) speed up the XPATH queries to content of child nodes by a certain property,
which is just a string object( probably there is some kind of indexes, similar
to SQL ones? we may add to the content node's property we will use to search
on?)
2) speed up retrieval of the BLOB object of each of content node
3) speed up removal of entire domain object node with all of it's children

Could somebody please advice some generic ways to achieve such benefits from
JackRabbit, since now we ended with very poor performance of search queries
and removal operations?

Thanks a lot in advance!

-- 
Eugene N Dzhurinsky

Mime
View raw message