Develop Reference

Getting IgniteCheckedException: Default Ignite instance has already been started exception when enabling Persistence on single Node

I am deploying an application where I need to maintain some data in Ignite cache. I used in memory Ignite cache. Here is the Ignite configuration I have used:
<property name="cacheConfiguration">
<list>
<bean
class="org.apache.ignite.configuration.CacheConfiguration">
<property name="name" value="IGNITE_DATA" />
<property name="cacheMode" value="PARTITIONED" />
<property name="atomicityMode" value="ATOMIC" />
<property name="writeSync"
value="PRIMARY_SYNC" />
<property name="backups"
value="${IGNITE_CACHE_BACKUPS}" />
</bean>
</list>
</property>
Now when I deployed multiple instances of my application and stored data in Ignite cache. Its shared among all the application instances.
Even if any any instance goes down and comes up after sometime it has the latest data via Ignite cache sync.
But issue occurs when all the application instances go down. When they come up data is gone since it was not persisted. For persistence I used dataStorageConfiguration property and enabled the persistence. Here is the change I added to Ignite configuration:
<property name="dataStorageConfiguration">
<bean
class="org.apache.ignite.configuration.DataStorageConfiguration">
<!-- Enabling Apache Ignite Persistent Store. -->
<property name="defaultDataRegionConfiguration">
<bean
class="org.apache.ignite.configuration.DataRegionConfiguration">
<property name="persistenceEnabled" value="true" />
</bean>
</property>
<!-- Changing Write Ahead Log Mode. -->
<property name="storagePath" value="${IGNITE_BC_STORE_PATH}"/>
<property name="walMode" value="LOG_ONLY" />
</bean>
</property>
Now when I deploy my application and I try and start Ignite from Java code as mentioned below:
log.info("Initializing IGNITE...");
ignite = Ignition.start(getClass().getResource(CONF_FILE));
I get an exception every time stating the default instance has already started.Tried several things but didn't work. Even if I remove the CacheConfiguration from Ignite Configuration and just keep dataStorageConfiguration I still getting the same error. Error is :
Caused by: class org.apache.ignite.IgniteCheckedException: Default Ignite instance has already been started.
at org.apache.ignite.internal.IgnitionEx.start0(IgnitionEx.java:1141)
at org.apache.ignite.internal.IgnitionEx.startConfigurations(IgnitionEx.java:1076)
at org.apache.ignite.internal.IgnitionEx.start(IgnitionEx.java:962)
at org.apache.ignite.internal.IgnitionEx.start(IgnitionEx.java:881)
at org.apache.ignite.Ignition.start(Ignition.java:373)
Normally this error comes when we try and run multiple Ignite nodes under same JVM but here I am running single node per JVM. Then also getting the error.
Please do correct me if I am wrong.
Any help here will be appreciated.

Most probably, you have more than one IgniteConfiguration bean in your config file. If one configuration bean extends another one, then make sure, that the parent is abstract.

I have resolved the issue. Seems like the issue was not woth the Ignite configuration but was with Spring Framework configuration.
I was creating the bean for the Ignite class using the lazy-init=true. I switched that to the eager-init and that resolved my issue.
Not sure how exactly it solved this but it worked at least in my case.

How do you know the actually value for a property set by a map of beans on java?

Ive found this configuration on a applicationContext xml and I still dont understand it (even if I found info about it, is not clear) could someone give me an idea of how java chooses the right bean for its property? I did not find this bean in other file of the project context, seems like is "magic"
<bean id="mailCreator" class="com.project.ConfigurableXferEmailCreator">
<property name="mailCreatorMap">
<util:map id="mailCreatorMap" key-type="java.lang.String"
value-type="com.project.BaseMailCreator">
<entry>
<key>
<util:constant static-field="com.project.creatorType.TYPE1"/>
</key>
<bean class="com.project.creator1" parent="baseCreator">
<property name="service" ref="someService1" />
</bean>
</entry>
<entry>
<key>
<util:constant static-field="com.project.creatorType.TYPE2"/>
</key>
<bean class="com.project.creator1" parent="baseCreator">
<property name="service" ref="someService2" />
</bean>
</entry>
.... and so on
I really have no idea how java recognizes which one will use, I just know it uses the right service but I dont see where is being specifically set, could someone give me a hand?
I checked couple of sites like this but still no idea , does it call all services??
http://www.java2s.com/Tutorials/Java/Spring/0140__Spring_Map_Properties.htm

Your question:
"how java chooses the right bean for its property"
I assume you mean for example this line:
<property name="service" ref="someService1" />
if you do not see another XML bean element that defines "someService1"
fr example:
<bean id="someService1" class="com.project.SomeService1Impl">
Then it exists in the Java Code
try to find an annotation like below
#Service("someService1")
public class SomeService1Impl {
...
How?:
Spring loads "someService1" in the Spring Context on initialization , so the XML can reference it.

Switch LDAP connection at runtime in Spring

I am new to spring. Admins of my spring based web app want to configure settings from the web interface, so users can authenticate against LDAP server with their company username and password.
Change in LDAP settings should be possible without restarting the application. This might happen during a 'migration' or whatever reason. I have a couple beans, which need to be refreshed after the admin saves new settings for the LDAP server:
<bean id="ldapServer" class="org.springframework.security.ldap.DefaultSpringSecurityContextSource">
<constructor-arg>
<list>
<value>${ldap.url1}</value>
...
</list>
</constructor-arg>
<constructor-arg value="${ldap.basedn}"</constructor-arg>
<property name="referral" value="${ldap.referral}" />
<property name="baseEnvironmentProperties">...</property>
<property name="userDn" value="${ldap.username}" />
<property name="password" value="${ldap.password}" />
</bean>
I am using Springframework 3.1.2. The problem is, there are constructor arguments, which I want to change and not affect other running jobs. I tried playing with Scoped proxy, but not to much success yet:
<bean id="ldapServer" scope="prototype" ...>
<aop:scoped-proxy/>
I was successful though to get ldapServer to reinstantiate, when using prototype scope by running this piece of code:
#Controller
public class LDAPSettingsController implements ApplicationContextAware {
public ModelAndView handleRequest(...) {
DefaultSpringSecurityContextSource ldap;
ldap = context.getParentBeanFactor().getBean("ldapServer");
System.out.println(ldap.hashCode());
return new ModelAndView(new RedirectView('login.jsp'));
}
...
}
Are scopes and proxies here the way to go, or is the another mechanism in Spring to reflect configuration changes into a running program instance?
UPDATE: Clear up the question.
UPDATE: The root problem with the AOP proxies was following root exception:
java.lang.IllegalArgumentException: Superclass has no null constructors but no arguments were given

What worked was adding proxy-target-class="false" attribute to the <aop:scoped-proxy/> tag. I created a new scope, which works better than prototype - It destroys beans on settings update. Now I have this in my beans.xml:
<bean class="org.springframework.beans.factory.config.CustomScopeConfigurer">
<property name="scopes">
<map>
<entry key="ldap">
<ref bean="ldapScope" />
</entry>
</map>
</property>
</bean>
<bean id="ldapScope" class="com.myapp.SettingsScope" />
<bean id="ldapServer" scope="ldap" ...>
<aop:scoped-proxy proxy-target-class="false"/>
<constructor-args>
<list><value>${ldap.url1}</value> .. </list>
</constructor-args>
...
</bean>
I also have a controller for LDAP settings into which I inject ldapScope and I call a method which destroys current life-cycle objects and starts a new life-cycle every time, user presses the apply button.
PS: Not sure if I handle the life-cycle "re-start" in the right way - people my way to look for auto-start beans and start them after such event happens (i.e.: Setting -> Apply)

Conditional initialization of classes in spring

I have a service which refers to a single source.
<bean id="XYZService" class="com.services.impl.DataService1">
<constructor-arg ref="DataSource1" />
</bean>
<bean id="DataSource1" class="com.source.impl.DataSource1">
<constructor-arg ref="DBDataSource"/>
<constructor-arg value="xyz"/>
</bean>
<bean id="DataSource2" class="com.source.impl.DataSource2">
<constructor-arg ref="MsgDataSource"/>
<constructor-arg value="xyz"/>
</bean>
Now if i want to perform a conditional check and my service should be able listen to particular source based on a input variable something like below.
<bean id="XYZService" class="com.services.impl.DataService1">
<constructor-arg ref=" $VARIABLE == true ? DataSource1 : DataSource2" />
</bean>
I did tried SPEL however no luck. I am beginner in spring. Any help will be appreciated.
Thanks.

There are many solutions. Here are two: You can use profiles for this. Define two profiles, define the DataSource beans with the same name but different profiles. (docs)
Alternatively, you can use a single bean and a static factory method (docs).
<bean id="DataSource" class="com.source.impl.DataSourceFactory"
factory-method="createInstance"/>
Inside of DataSourceFactory.createInstance(), you can check the flag and then create the correct data source in plain Java.
The latter is a bit easier to understand, IMO. Using profiles allows you to keep everything in XML (but you should really consider switching to the Java Configuration). The drawback with profiles is that you must not forget to activate at least one of the bean won't be defined.
A third option is to use three XML files and then modify the list of XML files that should be parsed when you pass it to the ApplicationContext. But that only works if you have control over this part of the code.

Assuming you are using Spring 3.1 or later, Spring Profiles may be the best solution.
Using the example of Production and Dev/QA environments, common bean declarations go in a shared file
<beans>
<bean id="XYZService" class="com.services.impl.DataService1">
<constructor-arg ref="DataSource" />
</bean>
</beans>
A separate configuration contains production references
<beans profiles="prod">
<bean id="DataSource" class="com.source.impl.DataSource1">
<constructor-arg ref="DBDataSource"/>
<constructor-arg value="xyz"/>
</bean>
</beans>
Another contains dev references
<beans profile="dev">
<bean id="DataSource" class="com.source.impl.DataSource2">
<constructor-arg ref="MsgDataSource"/>
<constructor-arg value="xyz"/>
</bean>
</beans>
To activate the given profile add -Dspring.profiles.active=prod to your JVM arguments
You can find more info here
Another approach uses factory methods.
<bean id="DataSource" class="com.source.impl.DataSourceFactory" factory-method="getInstance">
<constructor-arg value="#{VARIABLE}" />
</bean>
The above fragment assumes that you want your factory method to explcitly invoke the constructor of each of your services. If you dead set on using Spring to create the instances you can pass each datasource implementation as constructor arguments and use the constructor method as a simple dispatcher.

You need something like this:
<constructor-arg
ref="#{systemProperties.variable == 'true' ? 'DataSource1' : 'DataSource2'}" />
where "variable" is set like -Dvariable=true.

Building custom Spring config tags for a framework

I have a framework which currently requires pretty verbose setup in Spring:
<bean id="dpHibernateRemotingAdapter"
class="org.springframework.flex.core.ManageableComponentFactoryBean">
<constructor-arg value="org.dphibernate.adapters.RemotingAdapter" />
<property name="properties">
<value>
{"dpHibernate" :
{
"serializerFactory" : "org.dphibernate.serialization.SpringContextSerializerFactory"
}
}
</value>
</property>
</bean>
<bean id="dataAccessService" class="org.dphibernate.services.SpringLazyLoadService"
autowire="constructor">
<flex:remoting-destination />
</bean>
<bean id="dpHibernateSerializer" class="org.dphibernate.serialization.HibernateSerializer"
scope="prototype">
<property name="pageSize" value="10" />
</bean>
<bean id="dpHibernateDeserializer" class="org.dphibernate.serialization.HibernateDeserializer"
scope="prototype" />
I'd like to look at providing a more elegant configuration tag, similar to user-friendly tags used elsewhere in spring:
<context:annotation-config />
<mvc:annotation-driven />
<tx:annotation-driven />
<flex:message-broker/>
etc.,
However, I don't really know where to start.
How does this approach work? What are these tags called? What's their base class?
If someone could point me to the class names in the source (ideally, the <flex:message-broker />, as that's the closest problem set to my project), then I can go from there. I just don't really know where to start!

Custom XML namespaces are certainly possible (see Appendix D), but in my experience a royal pain to get working properly.
I strongly recommend that instead you use #Bean-style configuration. This lets you use Java to compose your bean graphs, instead of XML. Not only can it be much more concise in certain situations, it's properly type-safe, and more easily re-used.
Either way, you'll end up writing some Java that wires objects together. It's a question of how you want to expose that.

See Appendix D. Extensible XML authoring.