HashSet/HashMap在併發下的奇怪行爲

最近服務器 一直不太正常，經常cpu佔到200%,load到2，隨時都有可能報警

 

通過 top -H -p xx看到
 PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
27987 admin     25   0 1559m 542m  13m R 100.2 27.0  15995:57 java                                                                                                              
27988 admin     25   0 1559m 542m  13m R 100.2 27.0  16027:01 java

 

看來cpu完全消耗到了2個線程上，通過jstack threaddump,發現

 

"pool-4-thread-1" prio=10 tid=0x0000000059fa8000 nid=0x6d53 runnable [0x0000000043ce0000..0x0000000043ce0a90]
   java.lang.Thread.State: RUNNABLE
        at java.util.HashMap.put(HashMap.java:374)
        at java.util.HashSet.add(HashSet.java:200)
        at com.xx.smsgw.domain.MessagePool.put(MessagePool.java:29)
        at com.xx.smsgw.domain.SmsMTQueue.putSmsMT(SmsMTQueue.java:43)
        at com.xx.smsgw.smc.mt.MTChannelImpl.put(MTChannelImpl.java:45)
        at com.xx.smsgw.smc.route.MTRouteTask.run(MTRouteTask.java:49)
        at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
        at java.lang.Thread.run(Thread.java:619)

"pool-4-thread-2" prio=10 tid=0x000000005a621000 nid=0x6d54 runnable [0x00000000455f9000..0x00000000455f9a10]
   java.lang.Thread.State: RUNNABLE
        at java.util.HashMap.transfer(HashMap.java:484)
        at java.util.HashMap.resize(HashMap.java:463)
        at java.util.HashMap.addEntry(HashMap.java:755)
        at java.util.HashMap.put(HashMap.java:385)
        at java.util.HashSet.add(HashSet.java:200)
        at com.xx.smsgw.domain.MessagePool.put(MessagePool.java:29)
        at com.xx.smsgw.domain.SmsMTQueue.putSmsMT(SmsMTQueue.java:43)
        at com.xx.smsgw.smc.mt.MTChannelImpl.put(MTChannelImpl.java:45)
        at com.xx.smsgw.smc.route.MTRouteTask.run(MTRouteTask.java:49)
        at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
        at java.lang.Thread.run(Thread.java:619)

 

這就奇怪了，耗在HashMap上了，更奇怪的是我多次threaddump，這個信息完全保持不變，就好像thread僵在那一樣

 

從下面的這篇日誌找到答案：

https://jira.jboss.org/jira/browse/JGRP-525

 

說是併發的HashSet訪問導致，解決方案是

• 同步HashSet/HashMap

• HashSet/HashMap-->ConcurrentHashMap

 

我個人認爲如果每個線程一個單獨的HashSet/HashMap實例也可以解決，關鍵是要避免在線程不安全的環境中使用HashSet/HashMap