网站开发立项报告新浪体育世界杯

目录

1、问题背景

2、分析源码过程

3、解决办法

---

最近在测试环境spring cloud gateway突然出现了异常，在这里记录一下，直接上干货

1、问题背景

测试环境spring cloud gateway遇到以下异常

DataBufferLimitException: Exceeded limit on max bytes to buffer : 262144（超出了缓冲区的最大字节数限制）

乍一看，问题很简单啊，通过配置加大缓存区不就行了啊，于是就在application.yml加了以下配置

#将缓存区设置为2m
spring:codec:max-in-memory-size: 2MB

可是问题又出现了，通过调试发现配置的max-in-memory-size在程序启动初始化确实是生效的。但是有业务调用的时候，此参数的接收值为null,maxInMemorySize还是读取的默认值（256K）。

那咋整，只能从源码入手了。

2、分析源码过程

通过异常日志，可以定位到异常位置

后来发现我们自定义的拦截器获取body的信息是获取方式，代码如下

因为HandlerStrategies.withDefaults() 是每次都需要重新创建对象，并非是spring注入的对象，所以每次获取的都是默认值，导致配置不生效。

3、解决办法

在我们自定的拦截器中注入ServerCodecConfigurer类，通过该类获取配置。这样获取到的就是我们在application.yml中配置的缓存区配置的字节数限制了。

具体代码：

@Component
@Slf4j
public class RequestFilter implements GlobalFilter, Ordered {@Overridepublic int getOrder() {return OrderedConstant.HIGHEST_PRECEDENCE;}//手动注入ServerCodecConfigurer@AutowiredServerCodecConfigurer codecConfigurer;@Overridepublic Mono<Void> filter(ServerWebExchange exchange, GatewayFilterChain chain) {StopWatch stopWatch = new StopWatch();stopWatch.start();long startTime = System.currentTimeMillis();try {final Log logDTO = new Log();ServerHttpRequest request = exchange.getRequest();// 设置X-Request-IdAtomicReference<String> requestId = new AtomicReference<>(GenerateIdUtils.requestIdWithUUID());Consumer<HttpHeaders> httpHeadersConsumer = httpHeaders -> {String headerRequestId = request.getHeaders().getFirst(HeaderConstant.REQUEST_ID);if (!Strings.isNullOrEmpty(headerRequestId)) {requestId.set(headerRequestId);}logDTO.setRequestId(requestId.get());httpHeaders.set(HeaderConstant.REQUEST_ID, requestId.get());httpHeaders.set(HeaderConstant.START_TIME_KEY, String.valueOf(startTime));};// codecConfigurer.getReaders()获取pplication.yml中配置的缓存区配置的字节数ServerRequest serverRequest = ServerRequest.create(exchange,codecConfigurer.getReaders());URI requestUri = request.getURI();String uriQuery = requestUri.getQuery();String url = requestUri.getPath() + (!Strings.isNullOrEmpty(uriQuery) ? "?" + uriQuery : "");HttpHeaders headers = request.getHeaders();MediaType mediaType = headers.getContentType();String method = request.getMethodValue().toUpperCase();// 原始请求体final AtomicReference<String> requestBody = new AtomicReference<>();final AtomicBoolean newBody = new AtomicBoolean(false);if (mediaType != null && LogHelper.isUploadFile(mediaType)) {requestBody.set("上传文件");} else {if (method.equals("GET")) {if (!Strings.isNullOrEmpty(uriQuery)) {requestBody.set(uriQuery);}} else {newBody.set(true);}}logDTO.setLevel(Log.LEVEL.INFO);logDTO.setRequestUrl(url);logDTO.setRequestBody(requestBody.get());logDTO.setRequestMethod(method);logDTO.setIp(IpUtils.getClientIp(request));ServerHttpRequest serverHttpRequest = exchange.getRequest().mutate().headers(httpHeadersConsumer).build();ServerWebExchange build = exchange.mutate().request(serverHttpRequest).build();return build.getSession().flatMap(webSession -> {logDTO.setSessionId(webSession.getId());if (newBody.get() && headers.getContentLength() > 0) {Mono<String> bodyToMono = serverRequest.bodyToMono(String.class);return bodyToMono.flatMap(reqBody -> {logDTO.setRequestBody(reqBody);// 重写原始请求ServerHttpRequestDecorator requestDecorator = new ServerHttpRequestDecorator(exchange.getRequest()) {@Overridepublic Flux<DataBuffer> getBody() {NettyDataBufferFactory nettyDataBufferFactory = new NettyDataBufferFactory(new UnpooledByteBufAllocator(false));DataBuffer bodyDataBuffer = nettyDataBufferFactory.wrap(reqBody.getBytes());return Flux.just(bodyDataBuffer);}};return chain.filter(exchange.mutate().request(requestDecorator).build()).then(LogHelper.doRecord(logDTO));});} else {return chain.filter(exchange).then(LogHelper.doRecord(logDTO));}});} catch (Exception e) {log.error("请求日志打印出现异常", e);return chain.filter(exchange);}}}