一、场景:
服务端是微服务,服务A
有个上传图片的功能,需求是限制大小为10M,用户通过客户端把图片上传到服务A
,A
再调用服务B
把文件上传至阿里oss。
功能实际上很简单,实现的代码如下:
服务A
的Controller层逻辑
@PostMapping("/upload") public Result<String> upload(@RequestParam MultipartFile file) { Long userId = SecurityContextHolder.getUserId(); return Result.success(feedbackService.uploadAttachment(userId, file)); }
服务A
的Service层逻辑
@DubboReference(version = "1.0.0") private RemoteFileService fileService; @Override public String uploadAttachment(Long userId, MultipartFile file) { // 文件大小限制为10M if (file.getSize() > 10 * 1024 * 1024) { throw new ClientErrorException("文件大小超过10M限制"); } // 文件上传OSS return fileService.storage(file.getBytes(), file.getOriginalFilename(), userId); }
服务B
的业务逻辑,文件保存到阿里云OSS
@DubboService(version = "1.0.0") public class DubboFileService implements RemoteFileService { @Override public String storage(byte[] bytes, String fileName, Long userId) { String fileId = String.format("file/%d-%d%s", userId, System.currentTimeMillis(), fileName.substring(fileName.lastIndexOf('.'))); //上传到oss if (!ossService.upload(fileId, bytes)) { throw new ServerErrorException("上传失败"); } //文件信息 FileInfo info = new FileInfo(); info.setUserId(userId); info.setMd5(Md5Utils.getMD5(bytes)); info.setSize(bytes.length); info.setName(fileName); //完整路径 String filePath = String.format("%s/%s", ExternalUrl.CDN_HOST, fileId); info.setDownload(filePath); //保存到数据库 fileMapper.addOne(info); return filePath; } }
两个服务调用是采用的dubbo协议,配置如下:
公共配置:
dubbo: application: name: ${spring.application.name} protocol: # dubbo 协议 name: dubbo # dubbo 协议端口( -1 表示自增端口,从 20880 开始) port: -1 threads: 2000 registry: address: nacos://${spring.cloud.nacos.discovery.server-addr}/?namespace=dubbo check: false #生产者 provider: timeout: 25000 #消费者 consumer: timeout: 25000 check: false
服务A
配置
# dubbo dubbo: scan: # dubbo 服务扫描基准包 base-packages: com.xxx.feedback.dubbo cloud: subscribed-services: feedback
服务B
配置
# dubbo dubbo: scan: # dubbo 服务扫描基准包 base-packages: com.xxx.file.dubbo cloud: subscribed-services: file
二、测试过程
测试上传9.83M
文件,报500错误,显示后台异常,查看日志,异常日志如下
[ERROR] 2023-12-13 10:42:31.918 o.a.d.r.transport.AbstractCodec [52]: Data length too large: 12111020, max payload: 8388608, channel: NettyChannel [channel=[id: 0x25031a14, L:/192.168.0.2:53556 - R:/192.168.0.2:20881]] org.apache.dubbo.remoting.transport.ExceedPayloadLimitException: Data length too large: 10307502, max payload: 8388608, channel: NettyChannel [channel=[id: 0x25031a14, L:/192.168.0.2:53556 - R:/192.168.0.2:20881]] at org.apache.dubbo.remoting.transport.AbstractCodec.checkPayload(AbstractCodec.java:50) at org.apache.dubbo.remoting.exchange.codec.ExchangeCodec.encodeRequest(ExchangeCodec.java:270) at org.apache.dubbo.remoting.exchange.codec.ExchangeCodec.encode(ExchangeCodec.java:71) at org.apache.dubbo.rpc.protocol.dubbo.DubboCountCodec.encode(DubboCountCodec.java:40) at org.apache.dubbo.remoting.transport.netty4.NettyCodecAdapter$InternalEncoder.encode(NettyCodecAdapter.java:69) ... ...
查了资料,dubbo默认的传输大小是8388608B(8M)。修改配置,公共配置增加了dubbo.protocol.dubbo.payload
一项。
# dubbo配置,其他略 dubbo: protocol: # dubbo 协议 name: dubbo # dubbo 协议端口( -1 表示自增端口,从 20880 开始) port: -1 threads: 2000 dubbo: # 传输大数据容量 10MB payload: 10485760
继续测试,问题依旧,配置完全没有生效。继续查资料,又看到另外一种配置方式
# dubbo dubbo: application: name: ${spring.application.name} protocol: # dubbo 协议 name: dubbo # dubbo 协议端口( -1 表示自增端口,从 20880 开始) port: -1 threads: 2000 # 传输大数据容量 10M payload: 10485760 registry: address: nacos://${spring.cloud.nacos.discovery.server-addr}/?namespace=dubbo check: false #生产者 provider: payload: 10485760 timeout: 25000 #消费者 consumer: parameters: payload: 10485760 timeout: 25000 check: false
测试成功,说明配置已经生效了。更新到生产环境,再次测试,上传界面一直无响应,等了许久显示超时。日志打印的错误如下:
[ERROR] 2023-12-13 11:23:44.471 c.i.c.s.h.GlobalExceptionHandler [96]: 请求地址'/upload'发生未知异常. org.apache.dubbo.rpc.RpcException: Failed to invoke the method storage in the service com.xxx.file.api.RemoteFileService. Tried 3 times of the providers [192.168.0.2:20881] (1/1) from the registry 192.168.0.234:8848 on the consumer 192.168.0.2 using the dubbo version 2.7.15. Last error is: Invoke remote method timeout. method: storage, provider: dubbo://192.168.0.2:20881/com.xxx.file.api.RemoteFileService?anyhost=true&application=feedback&category=providers&check=false&deprecated=false&dubbo=2.0.2&dynamic=true&generic=false&init=false&interface=com.xxx.file.api.RemoteFileService&metadata-type=remote&methods=download,storage,allFiles&path=com.xxx.file.api.RemoteFileService&payload=11557050&pid=24099&protocol=dubbo&qos.enable=false®ister.ip=192.168.0.2&release=2.7.15&remote.application=file&revision=1.0.0&service.name=ServiceBean:/com.xxx.file.api.RemoteFileService:1.0.0&side=consumer&sticky=false&timeout=25000×tamp=1699945248523&version=1.0.0, cause: org.apache.dubbo.remoting.TimeoutException: Waiting server-side response timeout by scan timer. start time: 2023-12-13 11:23:19.423, end time: 2023-12-13 11:23:44.443, client elapsed: 86 ms, server elapsed: 24934 ms, timeout: 25000 ms, request: Request [id=591, version=2.0.2, twoway=true, event=false, broken=false, data=null], channel: /192.168.0.2:40518 -> /192.168.0.2:20881 at org.apache.dubbo.rpc.cluster.support.FailoverClusterInvoker.doInvoke(FailoverClusterInvoker.java:110) at org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invoke(AbstractClusterInvoker.java:265) at org.apache.dubbo.rpc.cluster.interceptor.ClusterInterceptor.intercept(ClusterInterceptor.java:47) at org.apache.dubbo.rpc.cluster.support.wrapper.AbstractCluster$InterceptorInvokerNode.invoke(AbstractCluster.java:92) at org.apache.dubbo.rpc.cluster.support.wrapper.MockClusterInvoker.invoke(MockClusterInvoker.java:98) at org.apache.dubbo.registry.client.migration.MigrationInvoker.invoke(MigrationInvoker.java:170) at org.apache.dubbo.rpc.cluster.support.registry.ZoneAwareClusterInvoker.doInvoke(ZoneAwareClusterInvoker.java:100) at org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invoke(AbstractClusterInvoker.java:265) at org.apache.dubbo.rpc.cluster.interceptor.ClusterInterceptor.intercept(ClusterInterceptor.java:47) at org.apache.dubbo.rpc.cluster.support.wrapper.AbstractCluster$InterceptorInvokerNode.invoke(AbstractCluster.java:92) at org.apache.dubbo.rpc.proxy.InvokerInvocationHandler.invoke(InvokerInvocationHandler.java:96) at org.apache.dubbo.common.bytecode.proxy2.storage(proxy2.java) at com.xx.feedback.service.impl.FeedbackServiceImpl.uploadAttachment(FeedbackServiceImpl.java:387) at com.xxx.feedback.service.impl.FeedbackServiceImpl$$FastClassBySpringCGLIB$$f1a53452.invoke(<generated>) ... ...
尝试修改超时配置,把timeout
时长改长1分钟。
# dubbo配置 dubbo: application: name: ${spring.application.name} protocol: # dubbo 协议 name: dubbo # dubbo 协议端口( -1 表示自增端口,从 20880 开始) port: -1 threads: 2000 # 传输大数据容量 10M payload: 10485760 registry: address: nacos://${spring.cloud.nacos.discovery.server-addr}/?namespace=dubbo timeout: 25000 check: false #生产者 provider: payload: 10485760 timeout: 65000 #消费者 consumer: parameters: payload: 10485760 timeout: 60000 check: false
再次测试依旧报错,难道说超时配置也不生效?没办法了,只能修改代码了。
服务A
(服务消费者)增加timeout
参数,但这个时间比消费者略大一点
@DubboReference(version = "1.0.0", timeout = 60 * 1000)
服务B
(服务生产者)增加timeout
参数,但这个时间比消费者略大一点
@DubboService(version = "1.0.0", timeout = 65 * 1000)
重启生效。
三、总结
相同的代码在不同的环境可能执行效果不一样,一定要认真检查和测试,保证功能没有任何问题才能正式使用。
另外,payload
最好设置大一点,我们正常的业务可能是文件+表单一起提交的,大小可能超过10M,防止这种情况就把payload
设置大一点。
还有人说可以使用dubbo的hessian
协议来解决大文件传输,这个我还没研究,后面有时间再看看。