DUE TO SPAM, SIGN-UP IS DISABLED. Goto Selfserve wiki signup and request an account.
While we are refactoring CloudStack architecture, to lay out a better component oriented foundation, be open to tools that most developers are familiar with in Java community, we started to experience the usage of Apache Spring Framework, switching to Spring touches a lot of existing CloudStack codebase, however, it does not bring any business logic changes, but it does introduce changes that CloudStack developers need to be aware of.
1. CloudStack Framework components and business logic components
With old component container managed by ComponentLocator class, we don't distinguish framework components and regular components that implement varios CloudStack business logic, we started to separate framework components and business logic components, there are separated context configuration file, these two files are applicationContext.xml.in and componentContext.xml.in, following is some sample content from Javelin
applicationContext.xml.in
<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:context="http://www.springframework.org/schema/context"
xmlns:tx="http://www.springframework.org/schema/tx"
xmlns:aop="http://www.springframework.org/schema/aop"
xsi:schemaLocation="http://www.springframework.org/schema/beans
_ http://www.springframework.org/schema/beans/spring-beans-3.0.xsd_
_ http://www.springframework.org/schema/tx _
_ http://www.springframework.org/schema/tx/spring-tx-3.0.xsd_
_ http://www.springframework.org/schema/aop_
_ http://www.springframework.org/schema/aop/spring-aop-3.0.xsd_
_ http://www.springframework.org/schema/context_
http://www.springframework.org/schema/context/spring-context-3.0.xsd">
<context:annotation-config />
<context:component-scan base-package="org.apache.cloudstack, com.cloud" />
<!--
@DB support
-->
<aop:config proxy-target-class="true">
<aop:aspect id="dbContextBuilder" ref="transactionContextBuilder">
<aop:pointcut id="captureAnyMethod"
expression="execution(* *(..))"
/>
<aop:around pointcut-ref="captureAnyMethod" method="AroundAnyMethod"/>
</aop:aspect>
<aop:aspect id="actionEventInterceptorAspect" ref="actionEventInterceptor">
<aop:pointcut id="captureEventMethod"
expression="execution(* *(..)) and @annotation(com.cloud.event.ActionEvent)"
/>
<aop:around pointcut-ref="captureEventMethod" method="AroundAnyMethod"/>
</aop:aspect>
</aop:config>
<bean id="transactionContextBuilder" />
<bean id="actionEventInterceptor" />
<!--
RPC/Async/EventBus
-->
<bean id="onwireRegistry"
init-method="scan" >
<property name="packages">
<list>
<value>org.apache.cloudstack.framework</value>
</list>
</property>
</bean>
<bean id="messageSerializer">
<property name="onwireClassRegistry" ref="onwireRegistry" />
</bean>
<bean id="transportProvider" init-method="initialize">
<property name="workerPoolSize" value="5" />
<property name="nodeId" value="Node1" />
<property name="messageSerializer" ref="messageSerializer" />
</bean>
<bean id="rpcProvider" init-method="initialize">
<constructor-arg ref="transportProvider" />
<property name="messageSerializer" ref="messageSerializer" />
</bean>
<bean id="eventBus" class = "org.apache.cloudstack.framework.eventbus.EventBusBase" />
<bean id="apiServlet" class = "com.cloud.api.ApiServlet" />
</beans>
Components that offer framework level services are collected in applicationContext.xml.in, these framework components usually help define the runtime environment for all other business logic components, for example, middle-ware service components that are for inter-component communication, CloudStack Database, general DB transaction management etc. CloudStack customers may have various combination of activated business logic components in a particular setup, but most of time, a same set of framework components is usually used.
componentContext.xml.in
<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:context="http://www.springframework.org/schema/context"
xmlns:tx="http://www.springframework.org/schema/tx"
xmlns:aop="http://www.springframework.org/schema/aop"
xsi:schemaLocation="http://www.springframework.org/schema/beans
http://www.springframework.org/schema/beans/spring-beans-3.0.xsd
http://www.springframework.org/schema/tx
http://www.springframework.org/schema/tx/spring-tx-3.0.xsd
http://www.springframework.org/schema/aop
http://www.springframework.org/schema/aop/spring-aop-3.0.xsd
http://www.springframework.org/schema/context
http://www.springframework.org/schema/context/spring-context-3.0.xsd">
<!--
Compose a CloudStack deployment with selected components here
-->
<bean id="databaseUpgradeChecker" />
<bean id="management-server" class ="com.cloud.server.ManagementServerExtImpl" />
<bean id="configuration-server" />
<bean id="clusterManagerImpl" />
<bean id="clusteredAgentManagerImpl" />
<bean id="clusteredVirtualMachineManagerImpl" />
<bean id="highAvailabilityManagerExtImpl" />
<!-- bean id="bareMetalVmManagerImpl" / -->
<bean id="userVmManagerImpl" />
<bean id="consoleProxyManagerImpl" />
<bean id="securityGroupManagerImpl2" />
<bean id="premiumSecondaryStorageManagerImpl" />
<bean id="randomlyIncreasingVMInstanceDaoImpl" />
<!--
Network Elements
-->
<bean id="Ovs">
<property name="name" value="Ovs"/>
</bean>
<bean id="ExternalDhcpServer">
<property name="name" value="ExternalDhcpServer"/>
</bean>
<bean id="BareMetal">
<property name="name" value="BareMetal"/>
</bean>
<bean id="SecurityGroupProvider">
<property name="name" value="SecurityGroupProvider"/>
</bean>
<bean id="VirtualRouter">
<property name="name" value="VirtualRouter"/>
</bean>
<bean id="VpcVirtualRouter">
<property name="name" value="VpcVirtualRouter"/>
</bean>
<bean id="NiciraNvp">
<property name="name" value="NiciraNvp"/>
</bean>
</beans>
2. How to declare to be a CloudStack component
In Spring term, CloudStack uses both singleton components and prototype components, Spring component is nothing else more than a Java POJO object, you can either declare it with @Component annotation or in <bean> declaration in applicationContext.xml or componentContext.xml. CloudStack uses a lot of singleton components, for example, managers, DAOs, following is an example of such components
@Component
@Local(value=ClusterDao.class)
public class ClusterDaoImpl extends GenericDaoBase<ClusterVO, Long> implements ClusterDao {
protected final SearchBuilder<ClusterVO> PodSearch;
protected final SearchBuilder<ClusterVO> HyTypeWithoutGuidSearch;
protected final SearchBuilder<ClusterVO> AvailHyperSearch;
protected final SearchBuilder<ClusterVO> ZoneSearch;
protected final SearchBuilder<ClusterVO> ZoneHyTypeSearch;
private static final String GET_POD_CLUSTER_MAP_PREFIX = "SELECT pod_id, id FROM cloud.cluster WHERE cluster.id IN( ";
private static final String GET_POD_CLUSTER_MAP_SUFFIX = " )";
@Inject
protected HostPodDao _hostPodDao;
public ClusterDaoImpl()
Unknown macro: {
}
// more content..
}
For most of low-level built-in components like DAOs, in Javelin practice, we use @Component annotation, since it saves us a lot of typing and can take advantage of Eclipse IDE's powerful refactoring and type-infer feature to help us coding. However, if you are developing high level components and you want to give flexibility for customer to choose to compose a particular CloudStack deployment, declare in componentContext.xml would be a better choice.
3. Auto-wiring
One of the biggest advantage of switching to Spring is that Auto-wiring in Component has now become consistent. With previous ComponentLocator, there are a lot of places that have to use run-time wiring, for example, using ComponentLocator.getManager() to wire a reference to a manager component. The reason for developers to do so is that ComponentLocator does not fully resolved dependent injection for components, so when inter-component relationship becomes complex in a large system like CloudStack, lots of hacking ways rise up inside CloudStack codebase. Although it solves the immediate needs, but it also creates a lot of confusion for developers, the more components being involved into injection business, the more the system is tightly coupled with ComponentLocator itself.
Following is an example of such hacking way, inside BaseCmd class, developer tried to resolve all the references for every used components into static variables at BaseCmd. It also means that, when you add a new Command class to the system, you will have to remember to do something about your new service at BaseCmd class
public abstract class BaseCmd
Unknown macro: {
static public NetworkACLService _networkACLService;
}
With Spring, developer can now always use @Inject annotation to declare such reference. One of the top two reasons for us to try Spring experiments in Javelin, since with more complete dependency injection, it gives us a cleaner component coding practice. (the other top reason is that we have broader integration supports from many other third-party vendors, i.e., JUnit)
4. CloudStack Spring component coding conventions
4.1 Be aware of injection auto-wiring time and runtime
With previous ComponentLocator, we have many places that use run-time resolution for fields that can't be resolved automatically by ComponentLocator's injection, and in many of these cases, these run-time resolutions happen at object construction time. Under Spring, such way won't work any more. The reason is that when Spring constructs the component object, the added runtime logic running inside the constructor can not feed information back into Spring injection framework, it may cause the process to be broken. The solution to this issues is to always leave the object construction and auto-wiring work to Spring, and leave the component initialization later
4.2 Public constructors
Protected or private constructors may be a good coding practice to enforce certain usage pattern of the class. However, protected or private constructors are not friendly to Spring injection process. Please always use public constructors, for the same reason, if you POJO class wants to be Spring component, don't seal the class by putting final modifier to your class
4.3 Component independent self initialization
It is a good practice to separate component construction and initialization, always try to make your component be self-dependable in initialization, self-dependable initialization means the component is able to initialize itself with the very minimal dependency to other component's initialization status. Use @PostConstruct to mark such initialization method for Spring to automatically call for you. Following is an example.
@Component
public class ConfigurationDaoImpl extends GenericDaoBase<ConfigurationVO, String> implements ConfigurationDao {
@PostConstruct
void initComponent()
Unknown macro: {
}
}
4.4 Be aware of Spring proxy mode
One side effect of using Spring is that component object created by Spring are not real object instantiated from you java class, but a proxy object in front of it. If your code heavily relies on runtime reflection info, your should be aware of such impact. During Javelin refactoring period, we've corrected a few cases that falls into this category. Examples can be found in ApiServer/ApiDispatcher.
For regular business logic component developers, there are very few chances that you should be aware of, however, if do you encounter some problems and suspect that you have fallen into such traps, checkout ApiServer/ApiDispatcher for a reference solution.
4.5 Avoid creating run-time relationship with component container
Although we've tried the best to cut the relationship to a component container inside component code, we still have very few places in framework components that need to be aware of existence of the component container, to avoid any strong binding to a particular container like Spring, we introduced a class called ComponentContext, it is responsible to bridge CloudStack component with a chosen component container (for now, it is Spring).
As a business logic component developer, you should avoid using any of the functions provided by ComponentContext, this can make the business component neutral to component container, since we now switched to standard javax @Inject auto-wiring annotation, it is always possible for your component to run in containers that are not CloudStack
5. Component lifecyle
In CloudStack, some components are lifecyle sensitive, examples are those manager objects, adapter objects, to make lifecyle management more general and flexible, there are some TODO works to make component life-cycle management easier. We'll keep this topic updated in the community.
Overview
Content Tools