Jon Haugsand wrote:
>We have a table with approx 1500 rows and a finder that picks one or
>two rows during each query like this:
>
>for (id=0;id++;id<1000) {
> // Start transaction
> Collection c=findByStatus(id,status);
> for(i=c.iterator();i.hasNext();) {
> ((EntityLocal) (i.next())) . setStatus(newStatus);
> }
> // End transaction
>}
>
Why not package the whole in one transaction (including the outher
for-statement), too see if this speeds things up, if it doesn't change
the semantics of your operation. Or you might start a nested transaction
for the inner for.
>
>Each such transaction takes around one second, which means that around
>1000 seconds to fulfill the whole operation.
>
>Looking at the server.log, we have those large SQL queries I have asked
>about earlier. However, we do not need those and thinking we have
>found the problem with the performance, we tried to eliminate them.
>
>1. Is it correct then to edit the conf/standardjbosscmp-jdbc.xml to tune
> this?
>
> We tried to set
> <strategy>on-find</strategy>
>
> and remove completely
> <page-size>1000</page-size>
>
> but alas, this did not help. What do we really do?
>
>2. In the CMP-manual, the examples on how to tune this for each finder
> only shows it when you use JBOSS-finders and not EJB-QL. Is it
> possible ot use EJB-QL and tune this per finder?
>
>
>
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user
