本篇涉及技术栈
动态代理设计模式
AOP 自定义注解
Spring-retry
guava-retrying
需求
接口异常需要重试
需求的不断提出, 怎么迭代升级?
更多接口需要加重试逻辑怎么处理?
需求导入(逐步完善)
基础接口异常重试
循环的方式, 可以实现简单重试的功能, 但是重试间隔等更多策略较麻烦
代码: com.zlc.studying.retry.simpleRetry.SimpleRetry
public class SimpleRetry {
private final static int MAX_TIMES = 5;
//调用处
public void test(){
int time = 0;
while(time < MAX_TIMES){
try {
doSth();
} catch (Exception e) {
time++;
if(time >= MAX_TIMES){
e.printStackTrace();
}
}
}
}
// 实际工作的接口
private void doSth(){
System.out.println("doSth.................");
throw new RuntimeException("测试重试");
}
public static void main(String[] args) {
new SimpleRetry().test();
}
}
加入动态代理
上述方式代码不好维护, 耦合性太高, 加入动态代理, 因为需要的地方肯定有没有实现接口的, 故采用cglib代理
代码: com.zlc.studying.retry.ProxyRetry.RetryProxyFactory
public class RetryProxyFactory implements MethodInterceptor {
private Object target;
public RetryProxyFactory(Object target) {
this.target = target;
}
public Object getProxyInstance(){
Enhancer enhancer = new Enhancer();
enhancer.setSuperclass(target.getClass());
enhancer.setCallback(this);
return enhancer.create();
}
@Override
public Object intercept(Object o, Method method, Object[] objects, MethodProxy methodProxy) throws Throwable {
int times = 0;
Object returnValue = null;
while(times < 5){
try {
//参数
returnValue = method.invoke(target, objects);
} catch (Exception e) {
times++;
if(times >= 5){
e.printStackTrace();
}
}
}
return returnValue;
}
}
public class ProxyRetry {
public void doSth(){
System.out.println("ProxyRetry doSth..........");
throw new RuntimeException("测试ProxyRetry");
}
public static void main(String[] args) {
ProxyRetry retry = new ProxyRetry();
ProxyRetry proxyFactory = (ProxyRetry) new RetryProxyFactory(retry).getProxyInstance();
proxyFactory.doSth();
}
}
AOP实现
策略不够灵活, 比如重试次数啥的, 当然也可以实现
代码: com.zlc.studying.retry.aop.RetryAspect 不好使 //TODO
spring-retry
spring已经提供了比较完善的模块, 不需要自己再造轮子了
依赖:
<dependency>
<groupId>org.springframework.retry</groupId>
<artifactId>spring-retry</artifactId>
<version>1.1.2.RELEASE</version>
</dependency>
<dependency>
<groupId>org.aspectj</groupId>
<artifactId>aspectjweaver</artifactId>
<version>1.5.4</version>
</dependency>
注解式使用
优缺点
代码简洁方便
@Recover 多个方法时无法指定方法
注解说明
@EnableRetry 需要开启注解
@Retryable 标注此注解的方法在发生异常时会进行重试
@Backoff 重试间隔策略
@Recover 熔断机制 全部重试失败调用标有该注解的方法
代码实现
错误示例: 由于retry用到了aspect增强,所有会有aspect的坑,就是方法内部调用,会使aspect增强失效,那么retry当然也会失效
@RestController
public class Test {
private Logger logger = LoggerFactory.getLogger(getClass());
@RequestMapping("test")
public void test() {
logger.info("==============");
try {
process();
} catch (Exception e) {
logger.error("eeeeeeeeeeeeeeeee");
e.printStackTrace();
}
}
// TODO: 由于retry用到了aspect增强,所有会有aspect的坑,就是方法内部调用,会使aspect增强失效,那么retry当然也会失效
@Retryable(value = Exception.class,
maxAttempts = 3,
backoff = @Backoff(delay = 2000L, multiplier = 1))
public void process() {
logger.info("test.........................");
//这里写 1 / 0 下面的 @Recover 都不生效
throw new RuntimeException("抛出异常");
}
@Recover
public void recover(Exception e) {
logger.info(" --------------------------- ");
logger.info(e.getMessage());
}
}
正确示例:
@RestController
public class Test {
private Logger logger = LoggerFactory.getLogger(getClass());
@Autowired
Test2 test2;
@RequestMapping("test")
public void test() {
logger.info("==============");
try {
test2.process();
} catch (Exception e) {
logger.error("eeeeeeeeeeeeeeeee");
e.printStackTrace();
}
}
}
@Component
public class Test2 {
private Logger logger = LoggerFactory.getLogger(getClass());
@Retryable(value = Exception.class,
maxAttempts = 3,
backoff = @Backoff(delay = 2000L, multiplier = 1))
public void process() {
logger.info("test.........................");
//这里写 1 / 0 下面的 @Recover 都不生效
throw new RuntimeException("抛出异常");
}
@Recover
public void recover(Exception e) {
logger.info(" --------------------------- ");
logger.info(e.getMessage());
}
}
方法式使用
优缺点
编码方式更加灵活, 而且可以在一个类里调用, 不会有注解式方法内部调用增强失效的问题
代码实现
RetryTemplate template = new RetryTemplate();
SimpleRetryPolicy policy = new SimpleRetryPolicy();
FixedBackOffPolicy policy1 = new FixedBackOffPolicy();
policy.setMaxAttempts(5);
policy1.setBackOffPeriod(1000);
//可以指定多个policy, 上面一个指定重试最大次数, 一个指定了时间间隔
template.setRetryPolicy(policy);
template.setBackOffPolicy(policy1);
boolean flag = template.execute(new RetryCallback<Boolean, Throwable>() {
@Override
public Boolean doWithRetry(RetryContext retryContext) throws Throwable {
return null;
}
});
还可以下面这样指定熔断方法:
Boolean execute = retryTemplate.execute(
//RetryCallback
retryContext -> {
String hello = helloService.hello();
log.info("调用的结果:{}", hello);
return true;
},
// RecoverCallBack
retryContext -> {
//RecoveryCallback
log.info("已达到最大重试次数");
return false;
}
);
guava retry
依赖:
<dependency>
<groupId>com.github.rholder</groupId>
<artifactId>guava-retrying</artifactId>
<version>2.0.0</version>
</dependency>
代码:
public boolean googleRetry(String source, String dest) throws ExecutionException, RetryException {
//定义重试机制
Retryer<Boolean> retryer = RetryerBuilder.<Boolean>newBuilder()
//重试条件
.retryIfException()
.retryIfRuntimeException()
.retryIfExceptionOfType(Exception.class)
.retryIfException(Predicates.equalTo(new Exception()))
.retryIfResult(Predicates.equalTo(false))
//等待策略 每次请求时间间隔
.withWaitStrategy(WaitStrategies.fixedWait(myConfig.getRetrySleep(), TimeUnit.MILLISECONDS))
//停止策略 : 尝试请求6次
.withStopStrategy(StopStrategies.stopAfterAttempt(myConfig.getRetries()))
//时间限制 : 某次请求不得超过2s , 类似: TimeLimiter timeLimiter = new SimpleTimeLimiter();
//.withAttemptTimeLimiter(AttemptTimeLimiters.fixedTimeLimit(2, TimeUnit.SECONDS))
.build();
//定义请求实现
Callable<Boolean> callable = new Callable<Boolean>() {
@Override
public Boolean call() throws Exception {
process(source, dest);
return true;
}
};
return retryer.call(callable);
}
guava retry更多高级用法
参考: https://www.jianshu.com/p/557eb67bb3d8