资讯

精准传达 • 有效沟通

从品牌网站建设到网络营销策划，从策略到执行的一站式服务

网站建设 >

查看其它板块

由于OOM导致锁没有释放的定位问题是怎么就觉得

由于OOM导致锁没有释放的定位问题是怎么就觉得，针对这个问题，这篇文章详细介绍了相对应的分析和解答，希望可以帮助更多想解决这个问题的小伙伴找到更简单易行的方法。

创新互联是一家专注于做网站、成都做网站与策划设计,九原网站建设哪家好?创新互联做网站,专注于网站建设10余年,网设计领域的专业建站公司;建站业务涵盖:九原等地区。九原做网站价格咨询:13518219792

最近发现业务上返回慢，并且feign的fallback被触发了。查看日志，发现触发了重试，是什么触发的重试呢，通过异常堆栈发现：

Caused by: feign.RetryableException: connect timed out executing GET http://test-service/test-api
        at feign.FeignException.errorExecuting(FeignException.java:65) ~[feign-core-9.7.0.jar!/:?]
        at feign.SynchronousMethodHandler.executeAndDecode(SynchronousMethodHandler.java:105) ~[feign-core-9.7.0.jar!/:?]
        at feign.SynchronousMethodHandler.invoke(SynchronousMethodHandler.java:77) ~[feign-core-9.7.0.jar!/:?]
        at feign.hystrix.HystrixInvocationHandler$1.run(HystrixInvocationHandler.java:107) ~[feign-hystrix-9.7.0.jar!/:?]
        at com.netflix.hystrix.HystrixCommand$2.call(HystrixCommand.java:302) ~[hystrix-core-1.5.18.jar!/:1.5.18]
        at com.netflix.hystrix.HystrixCommand$2.call(HystrixCommand.java:298) ~[hystrix-core-1.5.18.jar!/:1.5.18]
        at rx.internal.operators.OnSubscribeDefer.call(OnSubscribeDefer.java:46) ~[rxjava-1.3.8.jar!/:1.3.8]
        ... 27 more
Caused by: java.net.SocketTimeoutException: connect timed out
        at java.net.PlainSocketImpl.socketConnect(Native Method) ~[?:?]
        at java.net.AbstractPlainSocketImpl.doConnect(AbstractPlainSocketImpl.java:399) ~[?:?]
        at java.net.AbstractPlainSocketImpl.connectToAddress(AbstractPlainSocketImpl.java:242) ~[?:?]
        at java.net.AbstractPlainSocketImpl.connect(AbstractPlainSocketImpl.java:224) ~[?:?]
        at java.net.Socket.connect(Socket.java:591) ~[?:?]
        at sun.net.NetworkClient.doConnect(NetworkClient.java:177) ~[?:?]
        at sun.net.www.http.HttpClient.openServer(HttpClient.java:474) ~[?:?]
        at sun.net.www.http.HttpClient.openServer(HttpClient.java:569) ~[?:?]
        at sun.net.www.http.HttpClient.(HttpClient.java:242) ~[?:?]
        at sun.net.www.http.HttpClient.New(HttpClient.java:341) ~[?:?]
        at sun.net.www.http.HttpClient.New(HttpClient.java:362) ~[?:?]
        at sun.net.www.protocol.http.HttpURLConnection.getNewHttpClient(HttpURLConnection.java:1248) ~[?:?]
        at sun.net.www.protocol.http.HttpURLConnection.plainConnect0(HttpURLConnection.java:1187) ~[?:?]
        at sun.net.www.protocol.http.HttpURLConnection.plainConnect(HttpURLConnection.java:1081) ~[?:?]
        at sun.net.www.protocol.http.HttpURLConnection.connect(HttpURLConnection.java:1015) ~[?:?]
        at sun.net.www.protocol.http.HttpURLConnection.getInputStream0(HttpURLConnection.java:1587) ~[?:?]
        at sun.net.www.protocol.http.HttpURLConnection.getInputStream(HttpURLConnection.java:1515) ~[?:?]
        at java.net.HttpURLConnection.getResponseCode(HttpURLConnection.java:527) ~[?:?]
        at feign.Client$Default.convertResponse(Client.java:150) ~[feign-core-9.7.0.jar!/:?]
        at feign.Client$Default.execute(Client.java:72) ~[feign-core-9.7.0.jar!/:?]
        at org.springframework.cloud.sleuth.instrument.web.client.feign.TracingFeignClient.execute(TracingFeignClient.java:91) ~[spring-cloud-sleuth-core-2.0.3.RELEASE.jar!/:2.0.3.RELEASE]
        at org.springframework.cloud.sleuth.instrument.web.client.feign.LazyTracingFeignClient.execute(LazyTracingFeignClient.java:55) ~[spring-cloud-sleuth-core-2.0.3.RELEASE.jar!/:2.0.3.RELEASE]
        at org.springframework.cloud.openfeign.ribbon.RetryableFeignLoadBalancer$1.doWithRetry(RetryableFeignLoadBalancer.java:103) ~[spring-cloud-openfeign-core-2.0.3.RELEASE.jar!/:2.0.3.RELEASE]
        at org.springframework.cloud.openfeign.ribbon.RetryableFeignLoadBalancer$1.doWithRetry(RetryableFeignLoadBalancer.java:88) ~[spring-cloud-openfeign-core-2.0.3.RELEASE.jar!/:2.0.3.RELEASE]
        at org.springframework.retry.support.RetryTemplate.doExecute(RetryTemplate.java:287) ~[spring-retry-1.2.4.RELEASE.jar!/:?]
        at org.springframework.retry.support.RetryTemplate.execute(RetryTemplate.java:180) ~[spring-retry-1.2.4.RELEASE.jar!/:?]
        at org.springframework.cloud.openfeign.ribbon.RetryableFeignLoadBalancer.execute(RetryableFeignLoadBalancer.java:88) ~[spring-cloud-openfeign-core-2.0.3.RELEASE.jar!/:2.0.3.RELEASE]
        at org.springframework.cloud.openfeign.ribbon.RetryableFeignLoadBalancer.execute(RetryableFeignLoadBalancer.java:54) ~[spring-cloud-openfeign-core-2.0.3.RELEASE.jar!/:2.0.3.RELEASE]
        at com.netflix.client.AbstractLoadBalancerAwareClient$1.call(AbstractLoadBalancerAwareClient.java:104) ~[ribbon-loadbalancer-2.2.5.jar!/:2.2.5]
        at com.netflix.loadbalancer.reactive.LoadBalancerCommand$3$1.call(LoadBalancerCommand.java:303) ~[ribbon-loadbalancer-2.2.5.jar!/:2.2.5]
        at com.netflix.loadbalancer.reactive.LoadBalancerCommand$3$1.call(LoadBalancerCommand.java:287) ~[ribbon-loadbalancer-2.2.5.jar!/:2.2.5]
        at rx.internal.util.ScalarSynchronousObservable$3.call(ScalarSynchronousObservable.java:231) ~[rxjava-1.3.8.jar!/:1.3.8]
        at rx.internal.util.ScalarSynchronousObservable$3.call(ScalarSynchronousObservable.java:228) ~[rxjava-1.3.8.jar!/:1.3.8]
        at rx.Observable.unsafeSubscribe(Observable.java:10327) ~[rxjava-1.3.8.jar!/:1.3.8]

发现调用微服务test-service连接超时，注意是连接超时，而不是读取超时。接下来，首先查看ribbon连接超时配置，看是否配置的连接超时过短，发现是：

#ribbon默认连接超时
ribbon.ConnectTimeout=500

500ms建立tcp连接已经不短了，所以配置没有问题。从当前实例手动调用微服务实例的接口，看是否是通的：

curl http://test-service的ip:test-service的端口

发现访问成功，没有阻塞。证明网络连接没有问题，我们来看下netstat网络连接状态：

Active Internet connections (w/o servers)
Proto Recv-Q Send-Q Local Address           Foreign Address         State       
tcp        0      0 test-5c665cc496-6jkx5:54644 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54666 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54680 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54088 ip-10-238-8-68.eu-central-1.compute.internal:8211 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54674 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54682 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54652 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp       32      0 test-5c665cc496-6jkx5:47092 244-213-201-91.test.com:https CLOSE_WAIT  
tcp        0      0 test-5c665cc496-6jkx5:54662 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54624 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp       32      0 test-5c665cc496-6jkx5:47098 244-213-201-91.test.com:https CLOSE_WAIT  
tcp        0      0 test-5c665cc496-6jkx5:54672 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54630 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:58758 ip-10-238-9-71.eu-central-1.compute.internal:MySQL ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:8251 10-238-8-117.api-gateway.test1.svc.cluster.local:43132 FIN_WAIT2   
tcp        0      0 test-5c665cc496-6jkx5:54648 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:58778 ip-10-238-9-71.eu-central-1.compute.internal:mysql ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54646 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54628 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54650 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54632 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:54638 ip-10-238-8-119.eu-central-1.compute.internal:6379 ESTABLISHED 
tcp        0      0 test-5c665cc496-6jkx5:59434 ip-10-238-9-71.eu-central-1.compute.internal:mysql ESTABLISHED 
tcp       32      0 test-5c665cc496-6jkx5:47104 244-213-201-91.test.com:https CLOSE_WAIT  
tcp        1      0 test-5c665cc496-6jkx5:54146 ip-10-238-8-68.eu-central-1.compute.internal:8211 CLOSE_WAIT  
tcp       32      0 test-5c665cc496-6jkx5:47100 244-213-201-91.test.com:https CLOSE_WAIT  
tcp       32      0 test-5c665cc496-6jkx5:47106 244-213-201-91.test.com:https CLOSE_WAIT

发现：

  1. 并没有建立与现在service-test这个微服务任何实例ip的链接

  2. 并没有很多网络连接，也并没有很多Timed_waiting还有close_wait

猜想，可能是调用的service-test的ip地址不是最新的。

现在该Athas登场了，启动Arthas，我们查看下feign调用的真实ip： 由于我们使用了sleuth，所以这里用sleuth监控的feigh client去看调用ip：

watch org.springframework.cloud.sleuth.instrument.web.client.feign.TracingFeignClient execute params[0].url()

对应的代码是：

package org.springframework.cloud.sleuth.instrument.web.client.feign;

final class TracingFeignClient implements Client {


	@Override public Response execute(Request request, Request.Options options)
			throws IOException {
			    //.....
			}
}

通过观察，我们发现，确实是以前的ip。 那么为什么没有更新呢，我们来查看下EurekaClient相关的代码，参考我另一篇文章：Spring Cloud Eureka 全解 （4） - 核心流程-服务与实例列表获取详解. 我们来看下PollingServerListUpdater这个类，负责更新的线程池是：

int coreSize = poolSizeProp.get();
ThreadFactory factory = (new ThreadFactoryBuilder())
        .setNameFormat("PollingServerListUpdater-%d")
        .setDaemon(true)
        .build();
_serverListRefreshExecutor = new ScheduledThreadPoolExecutor(coreSize, factory);

谢天谢地，这个线程是有名字的。我们通过Arthas看下线程列表：

[arthas@24]$ thread
thread
Threads Total: 736, NEW: 0, RUNNABLE: 81, BLOCKED: 0, WAITING: 634, TIMED_WAITIN
G: 21, TERMINATED: 0                                                            
ID     NAME                GROUP        PRIORI STATE %CPU   TIME   INTER DAEMON 
11     Log4j2-TF-2-AsyncLo main         5      TIMED 39     138:46 false true   
17     Log4j2-TF-5-AsyncLo main         5      TIMED 37     137:27 false true   
1045   as-command-execute- system       10     RUNNA 19     0:0    false true   
1027   AsyncAppender-Worke system       9      WAITI 0      0:0    false true   
68     AsyncResolver-boots main         5      TIMED 0      0:0    false true   
264    AsyncResolver-boots main         5      WAITI 0      0:0    false true   
1025   Attach Listener     system       9      RUNNA 0      0:0    false true   
9      Common-Cleaner      InnocuousThr 8      TIMED 0      0:0    false true   
151    DataPublisher       main         5      TIMED 0      0:2    false true   
107    DestroyJavaVM       main         5      RUNNA 0      0:37   false false  
69     DiscoveryClient-0   main         5      TIMED 0      0:1    false true   
70     DiscoveryClient-1   main         5      WAITI 0      0:1    false true   
748    DiscoveryClient-Cac main         5      WAITI 0      0:9    false true   
751    DiscoveryClient-Hea main         5      WAITI 0      0:7    false true   
71     DiscoveryClient-Ins main         5      TIMED 0      0:0    false true   
24     Druid-ConnectionPoo main         5      WAITI 0      0:0    false true   
21     Druid-ConnectionPoo main         5      WAITI 0      0:0    false true   
25     Druid-ConnectionPoo main         5      TIMED 0      0:0    false true   
22     Druid-ConnectionPoo main         5      TIMED 0      0:0    false true   
20     Druid-ConnectionPoo main         5      TIMED 0      0:0    false true   
67     Eureka-JerseyClient main         5      WAITI 0      0:0    false true   
3      Finalizer           system       8      WAITI 0      0:0    false true   
119    ForkJoinPool.common main         5      WAITI 0      0:9    false true   
30     I/O dispatcher 1    main         5      RUNNA 0      0:2    false false  
155    NFLoadBalancer-Ping main         5      TIMED 0      0:0    false true   
150    PollingServerListUp main         5      WAITI 0      0:0    false true   
157    PollingServerListUp main         5      WAITI 0      0:0    false true   
2      Reference Handler   system       10     RUNNA 0      0:0    false true   
146    RibbonApacheHttpCli main         5      TIMED 0      0:0    false true   
135    RxComputationSchedu main         5      TIMED 0      1:44   false true   
132    RxIoScheduler-1 (Ev main         5      TIMED 0      0:0    false true   
4      Signal Dispatcher   system       9      RUNNA 0      0:0    false true   
114    SimplePauseDetector main         5      TIMED 0      2:10   false true   
115    SimplePauseDetector main         5      TIMED 0      2:10   false true   
116    SimplePauseDetector main         5      TIMED 0      2:12   false true

发现关键字PollingServerListUp有两个线程，150和157，分别看看他们在干什么：

[arthas@24]$ thread 150
thread 150
"PollingServerListUpdater-0" Id=150 WAITING on java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject@778af81b
    at java.base@11.0.4/jdk.internal.misc.Unsafe.park(Native Method)
    -  waiting on java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject@778af81b
    at java.base@11.0.4/java.util.concurrent.locks.LockSupport.park(LockSupport.java:194)
    at java.base@11.0.4/java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2081)
    at java.base@11.0.4/java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(ScheduledThreadPoolExecutor.java:1170)
    at java.base@11.0.4/java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(ScheduledThreadPoolExecutor.java:899)
    at java.base@11.0.4/java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1054)
    at java.base@11.0.4/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1114)
    at java.base@11.0.4/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
    at java.base@11.0.4/java.lang.Thread.run(Thread.java:834)

Affect(row-cnt:0) cost in 322 ms.
[arthas@24]$ thread 157
thread 157
"PollingServerListUpdater-1" Id=157 WAITING on java.util.concurrent.locks.ReentrantReadWriteLock$NonfairSync@3adcda2
    at java.base@11.0.4/jdk.internal.misc.Unsafe.park(Native Method)
    -  waiting on java.util.concurrent.locks.ReentrantReadWriteLock$NonfairSync@3adcda2
    at java.base@11.0.4/java.util.concurrent.locks.LockSupport.park(LockSupport.java:194)
    at java.base@11.0.4/java.util.concurrent.locks.AbstractQueuedSynchronizer.parkAndCheckInterrupt(AbstractQueuedSynchronizer.java:885)
    at java.base@11.0.4/java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireQueued(AbstractQueuedSynchronizer.java:917)
    at java.base@11.0.4/java.util.concurrent.locks.AbstractQueuedSynchronizer.acquire(AbstractQueuedSynchronizer.java:1240)
    at java.base@11.0.4/java.util.concurrent.locks.ReentrantReadWriteLock$WriteLock.lock(ReentrantReadWriteLock.java:959)
    at com.netflix.loadbalancer.BaseLoadBalancer.setServersList(BaseLoadBalancer.java:475)
    at com.netflix.loadbalancer.DynamicServerListLoadBalancer.setServersList(DynamicServerListLoadBalancer.java:156)
    at com.netflix.loadbalancer.DynamicServerListLoadBalancer.updateAllServerList(DynamicServerListLoadBalancer.java:267)
    at com.netflix.loadbalancer.DynamicServerListLoadBalancer.updateListOfServers(DynamicServerListLoadBalancer.java:250)
    at com.netflix.loadbalancer.DynamicServerListLoadBalancer$1.doUpdate(DynamicServerListLoadBalancer.java:62)
    at com.netflix.loadbalancer.PollingServerListUpdater$1.run(PollingServerListUpdater.java:116)
    at java.base@11.0.4/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:515)
    at java.base@11.0.4/java.util.concurrent.FutureTask.runAndReset(FutureTask.java:305)
    at java.base@11.0.4/java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:305)
    at java.base@11.0.4/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
    at java.base@11.0.4/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
    at java.base@11.0.4/java.lang.Thread.run(Thread.java:834)

    Number of locked synchronizers = 1
    - java.util.concurrent.ThreadPoolExecutor$Worker@6ab1c5f9

发现PollingServerListUpdater-1一直等待获取锁，看来问题就在这里了，查看对应的Ribbon代码，发现：

  1. PollingServerListUpdater-1需要获取allServerLock的写锁

  2. allServerLock的读锁，只有runPinger（对于每个实例ping，需要读取实例列表）获取。

那么是不是Ping出了问题，我们看下对应的获取锁的代码BaseLoadBalancer：

public void runPinger() throws Exception {
    //省略无用代码
    allLock = allServerLock.readLock();
    allLock.lock();
    allServers = allServerList.toArray(new Server[allServerList.size()]);
    allLock.unlock();

我们看到，这里的代码，并没有用try{lock} finally {unlock}的套路，如果中间代码异常，则会锁不能释放。 并且这里是重入锁，lock多少次就要unlock多少次，少一次，其他线程都获取不到。

allServers = allServerList.toArray(new Server[allServerList.size()]);并没有涉及任何IO，只是数据变换，也没有空指针异常。猜想是发生了OOM异常，导致内存没有分配。检查日志，果然发现了OOM。

这件事告诉我们，对于锁，一定要try{lock} finally {unlock}。就算代码不会抛出任何异常，发生OOM时，也有可能导致锁不能释放

关于由于OOM导致锁没有释放的定位问题是怎么就觉得问题的解答就分享到这里了，希望以上内容可以对大家有一定的帮助，如果你还有很多疑惑没有解开，可以关注创新互联行业资讯频道了解更多相关知识。


本文名称：由于OOM导致锁没有释放的定位问题是怎么就觉得
当前路径：http://cdkjz.cn/article/gigcoi.html
返回首页 了解更多建站资讯

相关资讯

多年建站经验

多一份参考，总有益处

联系快上网，免费获得专属《策划方案》及报价

咨询相关问题或预约面谈，可以通过以下方式与我们联系

大客户专线   成都：13518219792   座机：028-86922220

在线咨询 提交需求
友情链接 交换友情链接
成都网站运维成都机柜租用合江网站建设成都网站排名小程序开发公司雅安机房托管重庆网站建设成都装修公司成都租用服务器四川成都托管服务器

成都网站建设公司地址：成都市青羊区太升南路288号锦天国际A座10层 建设咨询028-86922220

成都快上网科技有限公司-四川网站建设设计公司 | 蜀ICP备19037934号 Copyright 2020,ALL Rights Reserved cdkjz.cn | 成都网站建设 | © Copyright 2020版权所有.

专家团队为您提供成都网站建设,成都网站设计,成都品牌网站设计,成都营销型网站制作等服务,成都建网站就找快上网！ | 成都网站建设哪家好？ | 网站建设地图