RPC参数校验框架实战：Dubbo + HibernateValidator

 

1.前言

用过SpringMVC的参数校验框架的同学都知道，通过注解的方式，可以优雅的实现Restful接口的参数校验。例如：

@NotBlank(message = "parameter appId required") String appId;

具体可以参考这篇文章：springmvc的数据校验的实现。

那么基于Dubbo框架的RPC接口是否也能使用类似的方式进行参数校验？下文是一种实现方案，供诸君参考。

2.准备工作

1）引入参数校验框架依赖的包

javax.el javax.el-api 3.0.0 org.glassfish javax.el 3.0.0 org.hibernate.validator hibernate-validator 6.0.12.Final

2）封装几个common类，用于生产者和消费者之间的返回值约定。

2-1）返回值通用类

@Data public class CommonResponse implements Serializable { private T data; private Integer code; private String message; }

2-2）返回code枚举

public enum ResponseCode { OK(0, "OK"), ... private Integer code; private String message; ResponseCode(Integer code, String message) { this.code = code; this.message = message; } }

2-3）通用异常类

public class RpcBizException extends RuntimeException{ private int code; private String message; public RpcBizException() { super(); } ... }

3. 自定义Validation

参考文档Dubbo参数验证扩展点

3.1 继承Dubbo的AbstractValidation类

import org.apache.dubbo.common.URL; import org.apache.dubbo.validation.Validator; import org.apache.dubbo.validation.support.AbstractValidation; /** * RPC接口自定义业务Validation（用于dubbo参数校验过滤） * */ public class DubboBizValidation extends AbstractValidation{ @Override protected Validator createValidator(URL url) { return new DubboBizValidator(url); } }

3.2 实现Dubbo的Validator接口

import java.util.Set; import javax.validation.ConstraintViolation; import javax.validation.ConstraintViolationException; import org.apache.dubbo.common.URL; import org.apache.dubbo.validation.Validator; import org.apache.dubbo.validation.support.jvalidation.JValidator; import xxx.ResponseCode; import xxx.RpcBizException; import lombok.extern.slf4j.Slf4j; /** * RPC接口自定义业务Validator（用于dubbo参数校验过滤） * */ @Slf4j public class DubboBizValidator implements Validator{ private JValidator jValidator; public DubboBizValidator(URL url) { jValidator = new JValidator(url); } @Override public void validate(String methodName, Class[] parameterTypes, Object[] arguments) throws Exception { try { jValidator.validate(methodName, parameterTypes, arguments); } catch (ConstraintViolationException e){ Set> violations = e.getConstraintViolations(); RpcBizException bizException = new RpcBizException(); if( null != violations && !violations.isEmpty() ) { ConstraintViolation violation = violations.iterator().next(); bizException.setCode(ResponseCode.PARAM_VALID_ERROR.getCode()); bizException.setMessage(violation.getMessage()); log.info("ConstraintViolationException occured : " + violation.getMessage()); } else { bizException.setCode(ResponseCode.INTERNAL_ERROR.getCode()); log.error("ConstraintViolationException occured but no message returned.", e); } throw bizException; } } }

要点在于将ConstraintViolationException异常捕获，封装为自定义的RpcBizException，用于后续过滤器处理。

4. 自定义Filter

参考文档：Dubbo调用拦截扩展

实现Dubbo的Filter接口，用于过滤RPC请求，并将自定义Validation抛出的异常封装为返回值返回给消费者。

import static org.apache.dubbo.common.constants.CommonConstants.CONSUMER; import static org.apache.dubbo.common.constants.CommonConstants.PROVIDER; import static org.apache.dubbo.common.constants.FilterConstants.VALIDATION_KEY; import org.apache.dubbo.common.extension.Activate; import org.apache.dubbo.common.utils.ConfigUtils; import org.apache.dubbo.rpc.AsyncRpcResult; import org.apache.dubbo.rpc.Filter; import org.apache.dubbo.rpc.Invocation; import org.apache.dubbo.rpc.Invoker; import org.apache.dubbo.rpc.Result; import org.apache.dubbo.rpc.RpcException; import org.apache.dubbo.validation.Validation; import org.apache.dubbo.validation.Validator; import xxx.CommonResponse; import xxx.RpcBizException; @Activate(group = {CONSUMER, PROVIDER}, value = VALIDATION_KEY, order = 9999) public class DubboBizValidationFilter implements Filter { private Validation validation; /** * Sets the validation instance for ValidationFilter * @param validation Validation instance injected by dubbo framework based on "validation" attribute value. */ public void setValidation(Validation validation) { this.validation = validation; } /** * Perform the validation of before invoking the actual method based on validation attribute value. * @param invoker service * @param invocation invocation. * @return Method invocation result * @throws RpcException Throws RpcException if validation failed or any other runtime exception occurred. */ @Override public Result invoke(Invoker invoker, Invocation invocation) throws RpcException { if (validation != null && !invocation.getMethodName().startsWith("$") && ConfigUtils.isNotEmpty(invoker.getUrl().getMethodParameter(invocation.getMethodName(), VALIDATION_KEY))) { try { Validator validator = validation.getValidator(invoker.getUrl()); if (validator != null) { validator.validate(invocation.getMethodName(), invocation.getParameterTypes(), invocation.getArguments()); } } catch (RpcBizException e) { // 处理DubboBizValidator返回的异常 Object retValue = CommonResponse.fail(e.getCode(), e.getMessage()); return AsyncRpcResult.newDefaultAsyncResult(retValue, invocation); } catch (RpcException e) { throw e; } catch (Throwable t) { return AsyncRpcResult.newDefaultAsyncResult(t, invocation); } } return invoker.invoke(invocation); } }

5. 配置

5.1 filter配置

在源码目录下创建文件：/resources/META-INF/dubbo/org.apache.dubbo.rpc.Filter

文件内容：

1）key：自定义filter名

2）value：自定义filter类全路径

dubboBizValidationFilter=xxx.DubboBizValidationFilter

5.2 validation配置

在源码目录下创建文件：/resources/META-INF/dubbo/org.apache.dubbo.validation.Validation

文件内容：

1）key：自定义validation名

2）value：自定义validation类全路径

dubboBizValidation=xxx.DubboBizValidation

5.3 dubbo配置

指定dubbo的validation和filter属性

其中"-validation"表示禁用Dubbo默认的filter，防止因优先级问题导致自定义filter无法执行。

建议在消费者和生产者两端都进行配置。消费者端配置参数校验可减少发生参数校验异常时的网络传输开销。

6. 使用参数校验框架

1）API定义

CommonResponse myMethod(@Valid XxxDTO dto)

返回值采用上文中的CommonResponse进行包装，确保异常消息序列化、反序列化能正确处理。

参数和SpringMVC中的用法一样，用javax.validation中的各种注解标注。

2）消费端处理示例

CommonResponse ret = xxxService.myMethod(dto); log.info("ret code : " + ret.getCode()); log.info("ret msg : " + ret.getMessage());