Performance Problem
by Harald Entner
Hello everyone,
i have a problem with the performance of hibernate. I execute the
following code during server startup about 5000 times.
public PoTextModule findTextModuleByNameAndLanguage(String name, String
uid) {
long st = System.currentTimeMillis();
Object[] keyValues = {name, uid};
try {
List l = getHibernateTemplate().find("from PoTextModule tm" +
" where tm.name = ? and tm.language.UID = ?",keyValues);
return (PoTextModule) (l.size()>0 ? l.get(0) : null);
} catch (Exception e) {
logger.error("Could not determine textmodule with name '" +
name + "' and language '" + uid +"'");
logger.error(e,e);
return null;
}
finally {
logger.info("ftmbnal: " + (System.currentTimeMillis()-st) +
".");
}
}
It should return a textmodule with the given name and language. When the
server starts, the query is pretty fast. But the more queries have been
executed, the more time it needs to execute.
The startup looks like this
start transaction
for all modules
for all textmodules in the modules
if textmodule is existent
if textmodule hasChanged
update module
else
create textmodule
// end for all textmodulesmodules
// end for all modules
findTextModuleByNameAndLanguage(String name, String uid) needs 0 to 30
ms in the beginning, but after a while sometimes 500ms.
What is wrong? The table has no more than 5000 entries, and there is
almost no difference if its empty or not.
Maybe too many queries in one transaction?
Here is my hbm:
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE hibernate-mapping PUBLIC "-//Hibernate/Hibernate Mapping DTD
3.0//EN" "http://hibernate.sourceforge.net/hibernate-mapping-3.0.dtd">
<hibernate-mapping default-cascade="none" default-access="property"
default-lazy="true" auto-import="true">
<class table="PoTextModule"
name="at.workflow.webdesk.po.model.PoTextModule" mutable="true"
polymorphism="implicit" dynamic-update="false" dynamic-insert="false"
select-before-update="false" optimistic-lock="version">
<cache usage="read-write" include="all" />
<id name="UID" type="string" column="TEXTMODULE_UID" length="32">
<generator class="uuid.hex" />
</id>
<property name="name" unique="false" optimistic-lock="true"
lazy="false" generated="never">
<column name="NAME" unique-key="nameAndLanguage" />
</property>
<property name="value" length="1000" unique="false"
optimistic-lock="true" lazy="false" generated="never" />
<many-to-one cascade="none" foreign-key="FK_TEXTMODULE_LANGUAGE"
not-null="true" name="language" unique="false" update="true"
insert="true" optimistic-lock="true" not-found="exception" embed-xml="true">
<column name="LANGUAGE_UID" unique-key="nameAndLanguage" />
</many-to-one>
<many-to-one column="ACTION_UID" cascade="none"
foreign-key="FK_TEXTMODULE_ACTION" not-null="false" name="action"
unique="false" update="true" insert="true" optimistic-lock="true"
not-found="exception" embed-xml="true" />
<many-to-one column="TEXTMODULE_PARENT_UID" cascade="none"
foreign-key="FK_TEXTMODULE_TEXTMODULE" not-null="false" name="parent"
unique="false" update="true" insert="true" optimistic-lock="true"
not-found="exception" embed-xml="true" />
<property name="allowUpdateOnVersionChange" unique="false"
optimistic-lock="true" lazy="false" generated="never" />
<set inverse="true" lazy="true" name="childs" sort="unsorted"
mutable="true" optimistic-lock="true" embed-xml="true">
<key column="TEXTMODULE_PARENT_UID" on-delete="noaction" />
<one-to-many class="at.workflow.webdesk.po.model.PoTextModule"
not-found="exception" embed-xml="true" />
</set>
<many-to-one column="MODULE_UID" cascade="none"
foreign-key="FK_TEXTMODULE_MODULE" not-null="false" name="module"
unique="false" update="true" insert="true" optimistic-lock="true"
not-found="exception" embed-xml="true" />
</class>
</hibernate-mapping>
The transaction is managed via spring.
Thanks in advance,
Harald
16 years, 10 months
- 1
- 0
Hibernate Search 3.0.0 Beta3
by Emmanuel Bernard
Hibernate Search 3.0.0 Beta3 has been released. It primarily fix a
severe bug introduced in beta2 (HSEARCH-66).
Besides bug fixes, this release brings some new features:
- Lucene index optimization (searchFactory.optimize())
- configurable batch size limit for indexing (drastically
simplifying complete data (re-)indexations)
- cleaner access to native Lucene (incl shared IndexReaders)
Download available at http://www.hibernate.org/6.html
16 years, 10 months
- 1
- 0
- ← Newer
- 1
- Older →