jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alexander Klimetschek (JIRA)" <j...@apache.org>
Subject [jira] Closed: (JCR-1794) This is performance Issue while inserting nodes in repository
Date Thu, 09 Oct 2008 11:44:44 GMT

     [ https://issues.apache.org/jira/browse/JCR-1794?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Alexander Klimetschek closed JCR-1794.
--------------------------------------

    Resolution: Invalid

You can configure the cache manager to the specific environments (see http://wiki.apache.org/jackrabbit/CacheManager
just in case you haven't already).

Otherwise, as this isn't a very clear and specific bug description, could you please re-send
it on the jackrabbit users list (see http://jackrabbit.apache.org/mailing-lists.html for subscription)
for a discussion of your issue. I'll close the bug for now.

> This is performance Issue while inserting nodes in repository
> -------------------------------------------------------------
>
>                 Key: JCR-1794
>                 URL: https://issues.apache.org/jira/browse/JCR-1794
>             Project: Jackrabbit
>          Issue Type: Improvement
>          Components: jackrabbit-core
>    Affects Versions: core 1.4.4
>         Environment: Tomcat and Weblogic10.2 on Windows 2003 and Solaris
>            Reporter: Rajesh Upadhyay
>             Fix For: core 1.4.4
>
>
> The issue related to the performance while creating more then 10000 nodes in the repository.
> When I am using tomcat 5.5 to run my application which is using jackrabbit repository,
then the cache of LocalItemStateManager and SharedItemStateManager Classes refreshed when
the size of ItemStaeMap object goes over about 6000.
> And when i will run the same code on the Weblogic 10.2, then cache of above classes will
not remains constant some times it will refreshed when its size is only 5000 and time it will
refreshed when the size of map will goes over 150000.
> The map size refresh process also depends on the process that we have started to created
nodes in repository. if we will insert the nodes in small numbers like when we will start
the process for 1000 node then cache will refreshed, and if we will start the process to created
nodes 50000 then cache refresh process will not remains constant. some times cache has more
then 100k entries. This process reduce the performance of creating nodes in repository.
> The issue on solaris system with weblogic 10.2 is quite different.
> On this system cache will never refresh while we will not restart the repository.
> I want to say by all the above description of problem that I am facing a big problem
because i want to create about 2-3lak nodes in repository and this process will takes too
much time,.
> my main problem concern is about different behavior of repository on different environment
while the max cache size will remain same 4*1024*1024 in all the cases.
> If any body has any suggestion then he can also sent it on my personal mail id "raj.upa8@gmail.com".
> Thanks in advance.
> Rk

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message