#RefreshScope stops #Scheduled task - spring

I have a monitoring app wherein I am running a fixedRate task. This is pulling in a config parameter configured with Consul. I want to pull in updated configuration, so I added #RefreshScope. But as soon as I update the config value on Consul, the fixedRate task stops running.
#Service
#RefreshScope
public class MonitorService {
#Autowired
private AppConfig appConfig;
#PostConstruct
public void postConstRun() {
System.out.println(appConfig.getMonitorConfig());
}
#Scheduled(fixedRate = 1000)
public void scheduledMonitorScan() {
System.out.println("MonitorConfig:" + appConfig.getMonitorConfig());
}
}
AppConfig class just has a single String parameter:
#Configuration
#Getter
#Setter
public class AppConfig {
#Value("${monitor-config:default value}")
private String monitorConfig;
}
As soon as I update the value in consul, the scheduled task just stops running (display in sheduledMonitorScan method) stop showing up.

I'm successfully get & override the values from consul config server using RefreshScopeRefreshedEvent
import java.text.SimpleDateFormat;
import java.util.Date;
import org.springframework.beans.factory.annotation.Value;
import org.springframework.cloud.context.config.annotation.RefreshScope;
import org.springframework.cloud.context.scope.refresh.RefreshScopeRefreshedEvent;
import org.springframework.context.ApplicationListener;
import org.springframework.scheduling.annotation.Scheduled;
import org.springframework.stereotype.Component;
#Component
#RefreshScope
public class AlertSchedulerCron implements ApplicationListener<RefreshScopeRefreshedEvent> {
private SimpleDateFormat sdf = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
#Value("${pollingtime}")
private String pollingtime;
/*
* #Value("${interval}") private String interval;
*/
#Scheduled(cron = "${pollingtime}")
//#Scheduled(fixedRateString = "${interval}" )
public void task() {
System.out.println(pollingtime);
System.out.println("Scheduler (cron expression) task with duration : " + sdf.format(new Date()));
}
#Override
public void onApplicationEvent(RefreshScopeRefreshedEvent event) {
// TODO Auto-generated method stub
}
}

Here's how we've solved this issue.
/**
* Listener of Spring's lifecycle to revive Scheduler beans, when spring's
* scope is refreshed.
* <p>
* Spring is able to restart beans, when we change their properties. Such a
* beans marked with RefreshScope annotation. To make it work, spring creates
* <b>lazy</b> proxies and push them instead of real object. The issue with
* scope refresh is that right after refresh in order for such a lazy proxy
* to be actually instantiated again someone has to call for any method of it.
* <p>
* It creates a tricky case with Schedulers, because there is no bean, which
* directly call anything on any Scheduler. Scheduler lifecycle is to start
* few threads upon instantiation and schedule tasks. No other bean needs
* anything from them.
* <p>
* To overcome this, we had to create artificial method on Schedulers and call
* them, when there is a scope refresh event. This actually instantiates.
*/
#RequiredArgsConstructor
public class RefreshScopeListener implements ApplicationListener<RefreshScopeRefreshedEvent> {
private final List<RefreshScheduler> refreshSchedulers;
#Override
public void onApplicationEvent(RefreshScopeRefreshedEvent event) {
refreshSchedulers.forEach(RefreshScheduler::materializeAfterRefresh);
}
}
So, we've defined an interface, which does nothing in particular, but allows us to call for a refreshed job.
public interface RefreshScheduler {
/**
* Used after refresh context for scheduler bean initialization
*/
default void materializeAfterRefresh() {
}
}
And here is actual job, whose parameter from.properties can be refreshed.
public class AJob implements RefreshScheduler {
#Scheduled(cron = "${from.properties}")
public void aTask() {
// do something useful
}
}
UPDATED:
Of course AJob bean must be marked with #RefreshScope in #Configuration
#Configuration
#EnableScheduling
public class SchedulingConfiguration {
#Bean
#RefreshScope
public AJob aJob() {
return new AJob();
}
}

I have done workaround for this kind of scenario by implementing SchedulingConfigurer interface.
Here I am dynamically updating "scheduler.interval" property from external property file and scheduler is working fine even after actuator refresh as I am not using #RefreshScope anymore.
Hope this might help you in your case also.
public class MySchedulerImpl implements SchedulingConfigurer {
#Autowired
private Environment env;
#Bean(destroyMethod = "shutdown")
public Executor taskExecutor() {
return Executors.newScheduledThreadPool(10);
}
#Override
public void configureTasks(final ScheduledTaskRegistrar taskRegistrar) {
taskRegistrar.setScheduler(this.taskExecutor());
taskRegistrar.addTriggerTask(() -> {
//put your code here that to be scheduled
}, triggerContext -> {
final Calendar nextExecutionTime = new GregorianCalendar();
final Date lastActualExecutionTime = triggerContext.lastActualExecutionTime();
if (lastActualExecutionTime == null) {
nextExecutionTime.setTime(new Date());
} else {
nextExecutionTime.setTime(lastActualExecutionTime);
nextExecutionTime.add(Calendar.MILLISECOND, env.getProperty("scheduler.interval", Integer.class));
}
return nextExecutionTime.getTime();
});
}
}

My solution consists of listening to EnvironmentChangeEvent
#Configuration
public class SchedulingSpringConfig implements ApplicationListener<EnvironmentChangeEvent>, SchedulingConfigurer {
private static final Logger LOGGER = LoggerFactory.getLogger(SchedulingSpringConfig.class);
private final DemoProperties demoProperties;
public SchedulingSpringConfig(DemoProperties demoProperties) {
this.demoProperties = demoProperties;
}
#Override
public void configureTasks(ScheduledTaskRegistrar taskRegistrar) {
LOGGER.info("Configuring scheduled task with cron expression: {}", demoProperties.getCronExpression());
taskRegistrar.addTriggerTask(triggerTask());
taskRegistrar.setTaskScheduler(taskScheduler());
}
#Bean
public TriggerTask triggerTask() {
return new TriggerTask(this::work, cronTrigger());
}
private void work() {
LOGGER.info("Doing work!");
}
#Bean
#RefreshScope
public CronTrigger cronTrigger() {
return new CronTrigger(demoProperties.getCronExpression());
}
#Bean
public ThreadPoolTaskScheduler taskScheduler() {
return new ThreadPoolTaskScheduler();
}
#Override
public void onApplicationEvent(EnvironmentChangeEvent event) {
if (event.getKeys().contains("demo.config.cronExpression")) {
ScheduledTasksRefresher scheduledTasksRefresher = new ScheduledTasksRefresher(triggerTask());
scheduledTasksRefresher.afterPropertiesSet();
}
}
}
Then I use the ContextLifecycleScheduledTaskRegistrar to recreate the task.
public class ScheduledTasksRefresher extends ContextLifecycleScheduledTaskRegistrar {
private final TriggerTask triggerTask;
ScheduledTasksRefresher(TriggerTask triggerTask) {
this.triggerTask = triggerTask;
}
#Override
public void afterPropertiesSet() {
super.destroy();
super.addTriggerTask(triggerTask);
super.afterSingletonsInstantiated();
}
}
Properties definition:
#ConfigurationProperties(prefix = "demo.config", ignoreUnknownFields = false)
public class DemoProperties {
private String cronExpression;
public String getCronExpression() {
return cronExpression;
}
public void setCronExpression(String cronExpression) {
this.cronExpression = cronExpression;
}
}
Main definition:
#SpringBootApplication
#EnableConfigurationProperties(DemoProperties.class)
#EnableScheduling
public class DemoApplication {
public static void main(String[] args) {
SpringApplication.run(DemoApplication.class, args);
}
}

Based on previous answers I added the following interface and used it on #RefreshScope annotated beans:
public interface RefreshScopeScheduled {
#EventListener(RefreshScopeRefreshedEvent.class)
default void onApplicationEvent() { /*do nothing*/ }
}

Related

Component bean is not injected (autowired) by spring when running AWS lambda function using Spring Cloud function AWS Adapter

I am implementing AWS lambda function creating handler using Spring Cloud function AWS Adapter SpringBootRequestHandler. The functional bean registered in GenericApplicationContext is invoked, but autowiring of component class is giving NullPointer Exception.
I have tried #ComponentScan for base package at Spring Application.
Application class:
#Slf4j
#SpringBootApplication
#ComponentScan({ "com.poc.evthub" })
#EnableConfigurationProperties(EventHubProperties.class)
public class EventHubServerlessApplication implements ApplicationContextInitializer<GenericApplicationContext> {
public EventHubServerlessApplication() {
}
public static void main(String[] args) throws Exception {
FunctionalSpringApplication.run(EventHubServerlessApplication.class, args);
}
#Bean
public KinesisEventFunction ingestEvents() {
return new KinesisEventFunction();
}
#Override
public void initialize(GenericApplicationContext context) {
log.debug("======== initialize ========");
context.registerBean("ingestEvents", FunctionRegistration.class,
() -> new FunctionRegistration<Function<KinesisEvent, ApiResponse>>(ingestEvents())
.type(FunctionType.from(KinesisEvent.class).to(ApiResponse.class).getType()));
}
}
Handler:
public class KinesisEventHandler extends SpringBootRequestHandler<KinesisEvent, ApiResponse> {
}
Functional Bean:
package com.poc.evthub.function;
import com.amazonaws.services.lambda.runtime.events.KinesisEvent;
import com.poc.evthub.beans.ApiResponse;
import com.poc.evthub.constant.Constants;
import com.poc.evthub.service.IngestionServiceFactory;
import lombok.extern.slf4j.Slf4j;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Component;
import java.util.function.Function;
#Slf4j
#Component
public class KinesisEventFunction implements Function<KinesisEvent, ApiResponse> {
private Context context = null;
#Autowired
private IngestionServiceFactory ingestionServiceFactory;
#Autowired
#Qualifier("targetExecutionContext")
public void setContext(Context context) {
log.info("Context: {}", context);
this.context = context;
}
#Override
public ApiResponse apply(final KinesisEvent kinesisEvent) {
log.info("KinesisEventFunction apply called...");
String sourceDomain = System.getenv(Constants.SYSENV.SOURCE_DOMAIN);
log.info("Source Domain = {}", sourceDomain);
if(null == kinesisEvent || null == kinesisEvent.getRecords()) {
log.error("Event contains no data. {}", System.lineSeparator());
//TODO build response NOT FOUND
return null;
}
else
log.info("Received {} records from {}. {}",
kinesisEvent.getRecords().size(),
kinesisEvent.getRecords().get(0).getEventSourceARN(),
System.lineSeparator());
log.info("ingestionServiceFactory = {}",ingestionServiceFactory);
ingestionServiceFactory.ingest();
return null;
}
}
Complete code and pom is uploaded at:
https://github.com/rjavaria/eventhub-serverless
KinesisEventFunction apply called...
Also able to read the environment value(source_domain) from lambda, and receiving Kinesis event record.
But #Autowired ingestionServiceFactory is null. I am injecting this component bean to delegate the business logic.
What is missed here, so spring is not able to inject this component bean?
Thanks in advance!
You could try manual injection of your IngestionServiceFactory bean into your function class via constructor or setter injection.
In your function class, add a constructor and remove #Autowired from your IngestionServiceFactory field, like:
...
public class KinesisEventFunction implements Function<KinesisEvent, ApiResponse> {
...
// No #Autowired here...
private final IngestionServiceFactory ingestionServiceFactory;
...
// The new constructor here...
public KinesisEventFunction(final IngestionServiceFactory pIngestionServiceFactory) {
this.ingestionServiceFactory = pIngestionServiceFactory;
}
...
}
Then in your main class (the one implementing ApplicationContextInitializer<GenericApplicationContext>), pass the reference to the factory when registering the function bean, like:
...
public class EventHubServerlessApplication implements ApplicationContextInitializer<GenericApplicationContext> {
...
#Autowired
private IngestionServiceFactory ingestionServiceFactory;
...
#Bean
public KinesisEventFunction ingestEvents() {
return new KinesisEventFunction(this.ingestionServiceFactory);
}
#Override
public void initialize(GenericApplicationContext context) {
log.debug("======== initialize ========");
context.registerBean("ingestEvents", FunctionRegistration.class,
() -> new FunctionRegistration<>(ingestEvents())
.type(FunctionType.from(KinesisEvent.class).to(ApiResponse.class).getType()));
}
}
Or, as you are already manually registering the function bean with context.registerBean(), you could just:
...
public class EventHubServerlessApplication implements ApplicationContextInitializer<GenericApplicationContext> {
...
#Autowired
private IngestionServiceFactory ingestionServiceFactory;
...
#Override
public void initialize(GenericApplicationContext context) {
log.debug("======== initialize ========");
context.registerBean("ingestEvents", FunctionRegistration.class,
() -> new FunctionRegistration<>(new KinesisEventFunction(this.ingestionServiceFactory))
.type(FunctionType.from(KinesisEvent.class).to(ApiResponse.class).getType()));
}
}
Please, let me know if it works!

Spring-Boot Failing to #Schedule tasks for List of beans created by #Configuration class

I am attempting to create List of beans of the same type in a class annotated by #Configuration. What I need is to execute #Scheduled function declared in those beans.
#SpringBootApplication
#EnableScheduling
public class DemoApplication
{
public static void main(String[] args)
{
SpringApplication.run(DemoApplication.class, args);
}
}
#Configuration
public class Config
{
#Bean
public List<Monitoring> mon()
{
List<Monitoring> list = new ArrayList<>();
for (int x = 0; x < 5; ++x)
{
list.add(new First());
}
return list;
}
}
public class First implements Monitoring
{
private static final Logger logger = LoggerFactory.getLogger(First.class);
#Override
public void doSth()
{
logger.info("first monitoring bean");
}
#Scheduled(fixedRate = 50)
private void init()
{
logger.info("scheduled task");
}
}
What am I expecting from these code snippets is for my 5 Beans of Monitoring to print "scheduled task" every 50ms, but I never see this output.
You are only creating a list as a bean with unmanaged Monitoring instances. That won't work.
You need to properly create the Monitoring beans (so that Spring can do its magic):
#Configuration
public class Config {
#Bean
public Monitoring first() {
return new First();
}
#Bean
public Monitoring second() {
return new First();
}
}

Field created in spring component in not initialized with new keyword

I have spring component class annotated with #Component and in it I have field ConcurrentHashMap map, which is init in constructor of component and used in spring stream listener:
#Component
public class FooService {
private ConcurrentHashMap<Long, String> fooMap;
public FooService () {
fooMap = new ConcurrentHashMap<>();
}
#StreamListener(value = Sink.INPUT)
private void handler(Foo foo) {
fooMap.put(foo.id, foo.body);
}
}
Listener handle messages sent by rest controller. Can you tell me why I always got there fooMap.put(...) NullPointerException because fooMap is null and not initialzied.
EDIT:
After #OlegZhurakousky answer I find out problem is with async method. When I add #Async on some method and add #EnableAsync I can't anymore use private modificator for my #StreamListener method. Do you have idea why and how to fix it?
https://github.com/schwantner92/spring-cloud-stream-issue
Thanks.
Could you try using #PostConstruct instead of constructor?
#PostConstruct
public void init(){
this.fooMap = new ConcurrentHashMap<>();
}
#Denis Stephanov
When I say bare minimum, here is what I mean. So try this as a start, you'll see that the map is not null and start evolving your app from there.
#SpringBootApplication
#EnableBinding(Processor.class)
public class DemoApplication {
private final Map<String, String> map;
public static void main(String[] args) {
SpringApplication.run(DemoRabbit174Application.class, args);
}
public DemoApplication() {
this.map = new HashMap<>();
}
#StreamListener(Processor.INPUT)
public void sink(String string) {
System.out.println(string);
}
}
With Spring everything has to be injected.
You need to declare a #Bean for the ConcurrentHashMap, that will be injected in you Component. So create a Configuration class like:
#Configuration
public class FooMapConfiguration {
#Bean("myFooMap")
public ConcurrentHashMap<Long, String> myFooMap() {
return new ConcurrentHashMap<>();
}
}
Then modify your Component:
#Component
public class FooService {
#Autowired
#Qualifier("myFooMap")
private ConcurrentHashMap<Long, String> fooMap;
public FooService () {
}
#StreamListener(value = Sink.INPUT)
private void handler(Foo foo) {
fooMap.put(foo.id, foo.body); // <= No more NPE here
}
}

Stop a scheduling job from rest endpoint

I am doing a Spring Boot Project
This is the main class
#SpringBootApplication
#ComponentScan(basePackages="blabla.quartz")
#EnableScheduling
public class App
{
public static void main( String[] args )
{
ConfigurableApplicationContext context =SpringApplication.run(App.class, args);
}
}
This is the controller
#RestController
public class Controller {
#Autowired
private SampleTask m_sampletask;
#Autowired TaskScheduler taskScheduler;
ScheduledFuture scheduledFuture;
int jobid=0;
#RequestMapping(value = "start/{job}", method = RequestMethod.GET)
public void start(#PathVariable String job) throws Exception {
m_sampletask.addJob(job);
Trigger trigger = new Trigger(){
#Override
public Date nextExecutionTime(TriggerContext triggerContext) {
org.quartz.CronExpression cronExp=null;
CronSequenceGenerator generator = new CronSequenceGenerator("0 * * ? * *");
Date nextExecutionDate = generator.next(new Date());
System.out.println(nextExecutionDate);
return nextExecutionDate;
}
};
scheduledFuture = taskScheduler.schedule(m_sampletask, trigger);
}
}
This is the ScheduleConfigurer implementation
#Service
public class MyTask implements SchedulingConfigurer{
#Bean
public TaskScheduler taskScheduler() {
ThreadPoolTaskScheduler scheduler = new ThreadPoolTaskScheduler();
scheduler.setThreadNamePrefix("somegroup-");
scheduler.setPoolSize(10);
scheduler.setWaitForTasksToCompleteOnShutdown(true);
scheduler.setAwaitTerminationSeconds(20);
return scheduler;
}
#Override
public void configureTasks(ScheduledTaskRegistrar taskRegistrar) {
}
}
This is the class which I am calling from controller as scheduled job
#Component
public class SampleTask implements Runnable{
private List<String> jobs=new ArrayList<String>();
private String jobName;
public void addJob(String job){
jobName=job;
}
#Override
public void run() {
System.out.println("Currently running "+jobName);
}
}
How to stop the schedule job by a rest endpoint(Suppose "/stop/{jobname}").. When I have started the job using the "/start/{jobname}" rest endpoint?
You will probably need to use the quartz scheduler (if not already), and add a service with the required methods, then inject that service into your controller.
There's a decent example here: https://github.com/javabypatel/spring-boot-quartz-demo
If you want an in-memory job store (that isn't a database), checkout the RAMJobStore: http://www.quartz-scheduler.org/documentation/quartz-2.x/configuration/ConfigRAMJobStore.html
Stop Example
This is an excerpt from the demo project. Credit goes to Jayesh Patel: https://github.com/javabypatel
/**
* Stop a job
*/
#Override
public boolean stopJob(String jobName) {
System.out.println("JobServiceImpl.stopJob()");
try{
String jobKey = jobName;
String groupKey = "SampleGroup";
Scheduler scheduler = schedulerFactoryBean.getScheduler();
JobKey jkey = new JobKey(jobKey, groupKey);
return scheduler.interrupt(jkey);
} catch (SchedulerException e) {
System.out.println("SchedulerException while stopping job. error message :"+e.getMessage());
e.printStackTrace();
}
return false;
}

Spring : #Scheduled task triggered through the #Controller and Websocket

I have an #Scheduled task which send data to a client every sec throught a websocket.
My need is to start running my scheduled task only when the client ask for it.
Instead of, my task starts when my server starts. it's not the behavior i want.
currently, I have a bean of my scheduled task which is declared in my SchedulingConfigurer :
#Configuration
#EnableScheduling
public class SchedulingConfigurer implements org.springframework.scheduling.annotation.SchedulingConfigurer {
#Bean
public ThreadPoolTaskScheduler taskScheduler() {
return new ThreadPoolTaskScheduler();
}
#Bean
public ScheduledTask scheduledTask() {
return new ScheduledTask();
}
#Override
public void configureTasks(ScheduledTaskRegistrar taskRegistrar) {
taskRegistrar.setTaskScheduler(taskScheduler());
}
}
Here is my spring controller code :
#MessageMapping("/hello")
public void greeting() throws Exception {
//How do I start my scheduled task here ?
}
Maybe isn't possible to do that with #Scheduled annotation and i have to use the TaskScheduler interface ?
remove #Scheduled declaration from ScheduledTask class
implements Runnable interface instead of
#Component
//#Scope(ConfigurableBeanFactory.SCOPE_PROTOTYPE)
public class ScheduledTask implements Runnable {
private static final Logger log = LoggerFactory.getLogger(ScheduledTask.class);
private static final SimpleDateFormat dateFormat = new SimpleDateFormat("HH:mm:ss");
public void doWork() {
printMessage();
// TODO real work
}
private void printMessage() {
log.info("time to work: {}", dateFormat.format(new Date()));
}
#Override
public void run() {
doWork();
}
}
schedule Your task in controller area like this
#Controller
public class ScheduledTaskController {
#Autowired
private TaskScheduler taskScheduler;
#Autowired
private ScheduledTask scheduledTask;
#RequestMapping(value = "/task/run", method = RequestMethod.GET)
public String runTask() {
// start to run task every 5 sec.
taskScheduler.schedule(scheduledTask, new CronTrigger("0/5 * * * * ?"));
// ok, redirect
return "redirect:/task";
}
}
#Schedule is the declarative way, so not the point you're trying to achieve here.
You could create a Bean using one of the TaskScheduler implementations, such as ThreadPoolTaskScheduler and inject that bean in your application.
It has all the necessary methods to dynamically schedule tasks.

Resources