jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From ndesk1900 <ndesk1...@gmail.com>
Subject performance issue;
Date Fri, 31 Jul 2009 02:57:03 GMT

Hello All,

Here the subtree that i'm having performance issues with;

Features

-- Feature
    x = 10
    y = 100
    val = [ multi value property ];
    NNode
    -- Feature
        reference property to some Feature node
    -- Feature
        reference property to some Feature node
    (around 30 such Feature nodes under NNode)

-- Feature
    x = 2661
    y = 100
    val = [ multi value property ];
    NNode
    -- Feature
        reference property to some Feature node
    -- Feature
        reference property to some Feature node
    (around 30 such Feature nodes under NNode)

(thousands of such Feature nodes under the Features Node; maybe approx
150,000)

Issues:-

1. When i execute an XQuery like this-
"Features/Feature[x=2661]" i get NO results. I am quite confident that there
exists some Feature nodes with this value for X, i wonder if this issue has
to do with the 'large amount of data' in the repository.

2. When i ran an XQuery like this "Features/Feature" to get all feature
nodes i get messages like these.

25297 [main] INFO
org.apache.jackrabbit.core.persistence.bundle.util.BundleCache - num=2436
mem=8191k max=8192k avg=3443 hits=46666 miss=23334
26974 [main] INFO
org.apache.jackrabbit.core.persistence.bundle.util.LRUNodeIdCache -
num=0/10240 hits=0 miss=80000

I read this post: - 
http://www.nabble.com/newb-performance-question-td17629427.html#a17629427
http://www.nabble.com/newb-performance-question-td17629427.html#a17629427 
but the solution included something to do with WEBDAV which i am not using.
The post also had something to do with adding these lines to do the
repository.xml.
<DataStore class="org.apache.jackrabbit.core.data.FileDataStore"> 
     
     
</DataStore> 

"The data store is optionally used to store large binary values.", i don't
think i am dealing with any large single binaries here, so i am not sure if
this will do the trick. Also, i hope there is a solution without need to
RE-index everything. I have loaded tremendous amount of data in the
repository which will take me days to re-index.

Thanks
Neville

(PS:- Also, i was wondering is there is some kind of GUI where you can
'view'/edit your repository real time?)
-- 
View this message in context: http://www.nabble.com/performance-issue--tp24750320p24750320.html
Sent from the Jackrabbit - Users mailing list archive at Nabble.com.


Mime
View raw message