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。