I have a spring app which do minimal operations with DB.
And I have a requirment that my application should run under the absence of DB(or when db is down).Below is my datasource configuration.
<bean id="dt31DataSource" class="com.mchange.v2.c3p0.ComboPooledDataSource" destroy-method="close" depends-on="systemPropertyInitializer"
p:driverClass="${dt31.driver_class}"
p:jdbcUrl="${dt31.url}"
p:user="${dt31.username}"
p:password="${dt31.password}"
p:idleConnectionTestPeriod="1000"
p:maxPoolSize="4"
p:minPoolSize="2"
p:maxIdleTime="2000"
p:unreturnedConnectionTimeout="600"
p:contextClassLoaderSource="library"
p:privilegeSpawnedThreads="true"
p:initialize=false
/>
<bean id="dt31SessionFactory" class="org.springframework.orm.hibernate4.LocalSessionFactoryBean">
<property name="dataSource" ref="dt31DataSource"/>
<property name="packagesToScan" value="com.t22.dt31"/>
<property name="hibernateProperties">
<props>
<prop key="hibernate.dialect">
${dt31.dialect}
</prop>
<prop key="hibernate.show_sql">
false
</prop>
<prop key="hibernate.hbm2ddl.auto">
update
</prop>
</props>
</property>
</bean>
I found a page in google,saying to use "initialize: false" in spring data source configuration.But I am using "ComboPooledDataSource" datasource which does not have this property.Is there any other way to achieve this ?
The application service layer can operate with two DAO layers (file system and database), so when the DB is down you catch the connection acquire exception and switch to the file system instead.
Having two sources of truth is going to make it difficult to preserve consistency across two different data sources, especially if one resource is not available.
When both resources are available you can use XADisk and Bitronix for XADataSource and JTA transaction management.
Related
I have a java application in which I'm using Spring and Hibernate frameworks. On my production environment, it would take too long (upto an hour) to start the application properly, as in, within 10 minutes. I could see the login page but if I try to login, it would just keep processing. Only after an hour or so, I would be able to login.
After some research I found that by default hibernate.temp.use_jdbc_metadata_defaults is set to true, in which case it takes time, so I set it to false. Now my application is starting up quickly in my local system when I point it to production db to test. Even in test environment it performs as expected. But when I rolled it out to production, it worked fine initially but now it again is taking too long to start up.
Any idea how to approach it now?
EDITED: adding more details:
initial startup time is huge. once it is up then pages would load fine. we are using postgres db and have multiple schemas. Please suggest what other details are needed
Below is session factory onfiguration:
<bean id="sessionFactory" class="org.springframework.orm.hibernate4.LocalSessionFactoryBean">
<property name="dataSource" ref="mydb"/>
<property name="entityInterceptor" ref="nullToEmptyStringInterceptor" />
<property name="hibernateProperties">
<props>
<prop key="hibernate.dialect">org.hibernate.dialect.PostgreSQLDialect</prop>
<prop key="hibernate.multiTenancy">SCHEMA</prop>
<prop key="hibernate.tenant_identifier_resolver">path_to.DataBaseResolver</prop>
<prop key="hibernate.multi_tenant_connection_provider">path_to.MultiTenantProvider</prop>
<prop key="hibernate.show_sql">false</prop>
<prop key="hibernate.temp.use_jdbc_metadata_defaults">false</prop>
</props>
</property>
<property name="packagesToScan">
<list>
<value>path_to_model_package</value>
</list>
</property>
</bean>
Below is entityManagerFactory configuration
<bean id="entityManagerFactory" class="org.springframework.orm.jpa.LocalContainerEntityManagerFactoryBean">
<property name="dataSource" ref="mydb" />
<property name="jpaVendorAdapter" ref="jpaVendorAdapter" />
<property name="jpaDialect" ref="jpaDialect" />
<property name="persistenceUnitName" value="myPersistenceUnit" />
<property name="packagesToScan" >
<list>
<value>path_to_model_package</value>
</list>
</property>
<property name="jpaProperties">
<props>
<prop key="hibernate.dialect">org.hibernate.dialect.PostgreSQLDialect</prop>
<prop key="hibernate.temp.use_jdbc_metadata_defaults">false</prop>
<prop key="hibernate.jdbc.use_get_generated_keys" >true</prop>
</props>
</property>
</bean>
I don't think we can tell you what the problem is. It will be something to do what the applications is doing during startup, and we can only guess what that would be.
However ...
What you should do is to profile the application during startup, and use the profiling to identify where most of the time is being spent. Then look to see if you can work out what it its doing ... and optimize it.
It would also be work turning on GC logging to see if memory usage / garbage collection is part of the problem. Ditto for checking if database activity is a significant bottleneck.
Finally, an important step in finding the cause of this problem is likely to be understanding the internal architecture of the application. Especially if this is code that has been passed from developer to developer over a number of years.
As you did not posted full details of your problem, like hibernate, spring and other app configurations so suggesting you a solution is bit difficult. The following could be possibilities:
Somewhere there could be deadlock during application load time. For example, so batch job running in background before the container completely starts.
If your db schema is always checked to be updated after the container start and your db is having a large number of tables and columns with a lot of data inside then it would definitely take time before that application loading completes.
Check server configurations like if there is another instance of application running in parallel which is causing delay.
After upgrading Hibernate version from 4.3.9.Final to 5.0.2.Final, Tomcat Server start-up time has increased.
After debugging, I realized that hibernate takes too much time in adding mapping locations (*.hbm.xml files) in its metadata sources.
I have added mapping location in session factory using following code, and in my project there are around 1000 hbm.xml files.
<bean id="baseSessionFactory" class="org.springframework.orm.hibernate5.LocalSessionFactoryBean">
<property name="hibernateProperties">
<props>
<prop key="hibernate.dialect">org.hibernate.dialect.Oracle10gDialect</prop>
<prop key="hibernate.show_sql">false</prop>
<prop key="hibernate.cache.use_query_cache">true</prop>
<prop key="hibernate.jdbc.fetch_size">300</prop>
<prop key="net.sf.ehcache.configurationResourceName">/ehcache.xml</prop>
<prop key="hibernate.cache.region.factory_class">org.hibernate.cache.ehcache.SingletonEhCacheRegionFactory</prop>
</props>
</property>
<property name="dataSource" ref="dataSource" />
<property name="mappingLocations">
<list>
<value>classpath*:com/*/**/*.hbm.xml</value>
</list>
</property>
</bean>
While starting tomcat server, I debugged and fournd that in method
org.springframework.orm.hibernate5.LocalSessionFactoryBean.afterPropertiesSet
following for loop takes too much time to add all mapping locations in metadata sources object of configuration class.
if (this.mappingLocations != null) {
// Register given Hibernate mapping definitions, contained in resource files.
for (Resource resource : this.mappingLocations) {
sfb.addInputStream(resource.getInputStream());
}
}
Is there any solution to improve performance here? Anyone noticed such issue after upgrading to Hibernate-5?
It seems to be a common problem but I am not able to resolve it.
my Spring 3.1.1 configuration is
<bean id="sessionFactoryEditSolution" class="org.springframework.orm.hibernate4.LocalSessionFactoryBean">
<property name="dataSource" ref="editSolutions-pool"/>
<property name="mappingResources">
<list>
<value>/editsolutions.hibernate.cfg.xml</value>
</list>
</property>
<property name="hibernateProperties">
<props>
<prop key="hibernate.dialect">org.hibernate.dialect.SQLServerDialect</prop>
<prop key="hibernate.cache.use_second_level_cache">true</prop>
<prop key="hibernate.cache.provider_class">net.sf.ehcache.hibernate.SingletonEhCacheProvider</prop>
<prop key="net.sf.ehcache.configurationResourceName">/ehcache.xml</prop>
<prop key="hibernate.max_fetch_depth">6</prop>
<prop key="hibernate.default_schema">dbo</prop>
</props>
</property>
</bean>
And this how I am trying to get the configuration object
Configuration editSolutionsConfiguration = `(Configuration)AppContext.getBean("&sessionFactoryEditSolution");`
Initially I have Hibernate 3 for my application but to integrate it with Spring 3.1.1 I have upgraded it to Hibernate 4 .That's why I have to keep hibernate3.jar in my lib folder to support few hibernate 3 specific code lines.
A couple of notes:
you can't work with two versions of hibernate on the classpath. Your old code should be updated
&x returns the factory bean, it doesn't return the produced object. Even without the ampersand, this would return the SessionFactory, rather than Configuration.
with spring you don't need Configuration actually, it is handled behind the scene
we are working with hibernate 3 and spring 3.1. 3.1.1 might be different, but either keep the lower versions of both, or upgrade both (including code)
I need your help in order to solve connection pool problem with Spring.
I’m using Spring with Java 1.4 (no annotation).
Here is the datasource and the connection pool definition:
<bean id="dataSource" class="org.apache.commons.dbcp.BasicDataSource" destroy-method="close">
<property name="driverClassName">
<value>${database.jdbcName}</value>
</property>
<property name="url" value="jdbc:sybase:${database.url}">
</property>
<property name="username">
<value>${database.username}</value>
</property>
<property name="password">
<value>${database.password}</value>
</property>
<property name="maxActive">
<value>${database.maxActive}</value>
</property>
<property name="maxIdle" >
<value>${database.maxIdle}</value>
</property>
<bean id="sessionFactory" class="org.springframework.orm.hibernate3.LocalSessionFactoryBean">
<property name="dataSource">
<ref bean="dataSource" />
</property>
<property name="hibernateProperties">
<props>
<prop key="hibernate.dialect">org.hibernate.dialect.SybaseDialect</prop>
<prop key="show_sql">true</prop>
<prop key="hibernate.connection.release_mode">after_transaction</prop>
<prop key="hibernate.dbcp.maxWait">100000</prop>
<prop key="hibernate.connection.autocommit">false</prop>
<prop key="cache.provider_class">org.hibernate.cache.NoCacheProvider</prop>
</props>
</property>
...
When all the connections are active, if the system need a new connection I got a message saying “opening JDBC connection” and all the process are stopped. I don’t understand why all the threads are locked.
All services and DAO classes are defined as singletons in Spring application context.
Any idea?
Thanks for your help.
Béranger
What values are you using for maxActive and maxIdle? I can give you a more definitive answer if I know those values, but in the meantime you can also try changing the value of hibernate.dbcp.maxWait from 100000 to 1 and test your application again.
Also make sure you haven't you accidentally made some of your DAO methods synchronized.
This was a weird one but I finally solved my problem.
The reason of this connection locking was that I used both declarative and programmatic transaction management with Spring. I think that my programmatic transaction management was wrong.
Here is the code used to begin a transaction:
//Return Spring context
ApplicationContext context = ApplicationContextHolder.getContext();
// transactionManager
PlatformTransactionManager transactionManager = (PlatformTransactionManager) context.getBean("txManager");
//Set propagation required
DefaultTransactionDefinition td = new DefaultTransactionDefinition();
td.setPropagationBehavior(TransactionDefinition.PROPAGATION_REQUIRED);
// Begin
TransactionStatus transactionStatus = transactionManager.getTransaction(td);
As you can see I set the programmatic transactions as “required”. But when those transactions were started in a transactional context, Spring didn’t recognized the current transaction and started a new one. It leaded to a connection locking...
I solved this mystery by removing the programmatic transaction management.
I currently use spring for depency injection. Hibernate uses a postgres dialect for the normal run, but I want to use HSQL for the DBUnitils Databank-Testing.
My Spring-Configuration contains this:
<!-- Hibernate session factory -->
<bean id="sessionFactory"
class="org.springframework.orm.hibernate3.annotation.AnnotationSessionFactoryBean">
<property name="dataSource" ref="dataSource"/>
<property name="hibernateProperties">
<props>
<prop key="hibernate.dialect">${hibernate.dialect}</prop>
<prop key="hibernate.show_sql">${hibernate.show_sql}</prop>
<prop key="use_outer_join">${hibernate.use_outer_join}</prop>
<prop key="hibernate.cache.use_second_level_cache">${hibernate.cache.use_second_level_cache}</prop>
<prop key="hibernate.cache.use_query_cache">${hibernate.cache.use_query_cache}</prop>
<prop key="hibernate.cache.provider_class">${hibernate.cache.provider}</prop>
<prop key="hibernate.connection.pool_size">10</prop>
<prop key="hibernate.jdbc.batch_size">1000</prop>
<prop key="hibernate.bytecode.use_reflection_optimizer">true</prop>
</props>
</property>
<property name="annotatedClasses">
<list>
<value>de.dbruhn.relations.model.Relation</value>
<value>de.dbruhn.relations.model.RelationData</value>
<value>de.dbruhn.relations.model.RObject</value>
<value>de.dbruhn.relations.model.Type</value>
</list>
</property>
<property name="schemaUpdate" value="${hibernate.schemaUpdate}"/>
</bean>
The fields get replaced by maven resource-filtering.
The Spring-Configruation for DBUnitils contains this:
<bean id="dataSource" class="org.unitils.database.UnitilsDataSourceFactoryBean"/>
</beans>
and so overrides the dataSource from my run configuration with the UnitilsDataSource.
The Problem: I cant run the Tests using Postgres-Dialect against the HSQL-Test-Database because some commands dont work.
The only solution which came to my mind: Switching the resource-filter in maven, but I have to do this by hand (by provining a "-P TEST" on every maven call). So isn't there a possibilty to override the hibernate.dialect?
Thanks!
You normally don't need to specify the dialect at all, Hibernate will figure it out by looking at the underlying datasource. You only need to specify the dialect if you want to force Hibernate to use a specific one.
In your case, you should just be able to remove the dialect property completely, and it should work in postgres and hsql without config modification.
You should possibly look at using the PropertyPlaceHolderConfigurer in spring to change the config. That way you need only supply a different config file for the different environments, the spring xml stays the same.
And you can load the config file like this.