为解决DataBufferLimitException: Exceeded limit on max bytes to buffer : 262144 引发OOM事故

之前做项目有一个需求就是上传文件，用到的是spring cloud框架，统一走的网关，但是这网关默认允许文件上传大小为256KB，超过的话就会爆DataBufferLimitException: Exceeded limit on max bytes to buffer : 262144的错误。

spring是提供了这个参数的配置如下截图

 本来以为改一下这个参数就可以了，但是还是会报错，定位到了路由工厂配置类（RequestBodyRoutePredicateFactory继承AbstractRoutePredicateFactory）会重新创建ServerRequest对象，主要参考DataBufferLimitException: Exceeded limit on max bytes to buffer : 262144_lizz666的博客-CSDN博客

那么spring帮我们初始化的内存值就不生效了

所以就按解决方法引入的这个Bean，去调用getReaders()方法，让重新加载nacos的值

试了一下确实没有问题了，然后就上线了，

然后悲剧就开始了，在运行的期间出现OOM了，大概2个小时出现一次。

后面定位到这行代码有问题

 那么就翻一下源码看一下，发现在new DefaultPartHttpMessageReader()对象的时候会创建一个调度任务，当时就感觉不对劲了

 就再点进去看看发现会new 一个BoundedElasticScheduler对象

 

 于是去查了一下资料

这就尴尬了，每个请求都会创建单线程池，这并发一大不就gg了么。

最后没有想到什么好办法，只能重写源码了

/*
 * Copyright 2002-2019 the original author or authors.
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *      https://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package org.springframework.core.codec;

import org.reactivestreams.Publisher;
import org.springframework.core.ResolvableType;
import org.springframework.core.io.buffer.DataBuffer;
import org.springframework.core.io.buffer.DataBufferUtils;
import org.springframework.lang.Nullable;
import org.springframework.util.MimeType;
import reactor.core.publisher.Flux;
import reactor.core.publisher.Mono;

import java.util.Map;

/**
 * Abstract base class for {@code Decoder} implementations that can decode
 * a {@code DataBuffer} directly to the target element type.
 *
 * <p>Sub-classes must implement {@link #decodeDataBuffer} to provide a way to
 * transform a {@code DataBuffer} to the target data type. The default
 * {@link #decode} implementation transforms each individual data buffer while
 * {@link #decodeToMono} applies "reduce" and transforms the aggregated buffer.
 *
 * <p>Sub-classes can override {@link #decode} in order to split the input stream
 * along different boundaries (e.g. on new line characters for {@code String})
 * or always reduce to a single data buffer (e.g. {@code Resource}).
 *
 * @author Rossen Stoyanchev
 * @since 5.0
 * @param <T> the element type
 */
@SuppressWarnings("deprecation")
public abstract class AbstractDataBufferDecoder<T> extends AbstractDecoder<T> {

	private int maxInMemorySize = 100 * 1024 * 1024;


	protected AbstractDataBufferDecoder(MimeType... supportedMimeTypes) {
		super(supportedMimeTypes);
	}


	/**
	 * Configure a limit on the number of bytes that can be buffered whenever
	 * the input stream needs to be aggregated. This can be a result of
	 * decoding to a single {@code DataBuffer},
	 * {@link java.nio.ByteBuffer ByteBuffer}, {@code byte[]},
	 * {@link org.springframework.core.io.Resource Resource}, {@code String}, etc.
	 * It can also occur when splitting the input stream, e.g. delimited text,
	 * in which case the limit applies to data buffered between delimiters.
	 * <p>By default this is set to 256K.
	 * @param byteCount the max number of bytes to buffer, or -1 for unlimited
	 * @since 5.1.11
	 */
	public void setMaxInMemorySize(int byteCount) {
		this.maxInMemorySize = byteCount;
	}

	/**
	 * Return the {@link #setMaxInMemorySize configured} byte count limit.
	 * @since 5.1.11
	 */
	public int getMaxInMemorySize() {
		return this.maxInMemorySize;
	}


	@Override
	public Flux<T> decode(Publisher<DataBuffer> input, ResolvableType elementType,
			@Nullable MimeType mimeType, @Nullable Map<String, Object> hints) {

		return Flux.from(input).map(buffer -> decodeDataBuffer(buffer, elementType, mimeType, hints));
	}

	@Override
	public Mono<T> decodeToMono(Publisher<DataBuffer> input, ResolvableType elementType,
			@Nullable MimeType mimeType, @Nullable Map<String, Object> hints) {

		return DataBufferUtils.join(input, this.maxInMemorySize)
				.map(buffer -> decodeDataBuffer(buffer, elementType, mimeType, hints));
	}

	/**
	 * How to decode a {@code DataBuffer} to the target element type.
	 * @deprecated as of 5.2, please implement
	 * {@link #decode(DataBuffer, ResolvableType, MimeType, Map)} instead
	 */
	@Deprecated
	@Nullable
	protected T decodeDataBuffer(DataBuffer buffer, ResolvableType elementType,
			@Nullable MimeType mimeType, @Nullable Map<String, Object> hints) {

		return decode(buffer, elementType, mimeType, hints);
	}

}

只能先这么处理了，后面再看看怎么把配置值放到nacos上。

总结：明明只改了一行代码却引发了OOM的事故，平常解决问题，不能看到代码就直接抄，要分析多思考，希望能够帮助到其他人避免踩坑。