jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Antonio Gioia <antonio.gioi...@gmail.com>
Subject indexing configuration performance
Date Wed, 09 May 2012 12:10:29 GMT
Hi to all,

I would wont run a search operation only on specific properties of my nodes.
Follow my node type definition (.cnd file):

*<nt = 'http://www.jcp.org/jcr/nt/1.0'>
<nthl = 'http://ip-server:port/hl/nthl'>
<hl = 'http://ip-server:port/hl'>

[nthl:user] > nt:base
  - hl:portalLogin (String) mandatory
  - hl:uuid (String) mandatory
  - hl:scope (String)
[nthl:workspaceItem] >
  - hl:lastAction (String) mandatory
  + *

[nthl:workspaceLeafItem] > nthl:workspaceItem, nt:file
  - hl:workspaceItemType (String)
  - hl:workflowId (String)
  -	hl:workflowStatus (String)
  -	hl:workflowData (String) 
  + hl:owner(nthl:user)
  = nthl:user
  mandatory autocreated*

and below there is the my indexing_configuration.xml file

*<?xml version="1.0"?>
<!DOCTYPE configuration SYSTEM
<configuration xmlns:nt="http://www.jcp.org/jcr/nt/1.0" 
xmlns:nthl = "http://ip-server:port/hl/nthl"
xmlns:hl = "http://ip-server:port/hl">
 <index-rule nodeType="nthl:workspaceItem">
    <property nodeScopeIndex="false">.*:.*</property> 
  <index-rule nodeType="nthl:workspaceLeafItem">
    <property nodeScopeIndex="false">.*:.*</property> 
 <index-rule nodeType="nt:file">
    <property nodeScopeIndex="false">.*:.*</property> 

finally follow my repository.xml file ... in particular the search index

<SearchIndex class="org.apache.jackrabbit.core.query.lucene.SearchIndex">


The problem is that I don't see difference of performance to regard CPU
usage if I don't specified the param indexingConfiguration. I mean, I would
use this configuration to avoid full text indexing that I know use much CPU
....but I don't see difference, the CPU usage go up over 100%. Instead, if I
disable completely the search index section adding xml comment (), 

the CPU usage is acceptable, about 60%, and the add node operations are five
times faster.....obviously with this last configuration I can't execute none
search operation :).

How I can obtain comparable performance between this last configuration
(search index disable) and a search index for only those properties shown
above ("hl:workspaceItemType" and "jcr:title") ? I mean, I would wont search
my nodes through only these properties but I would execute the add node
operation using the CPU reasonably...

best regards.

View this message in context: http://jackrabbit.510166.n4.nabble.com/indexing-configuration-performance-tp4620246.html
Sent from the Jackrabbit - Users mailing list archive at Nabble.com.

View raw message