Hadoop运维笔记 之 Snappy创建libhadoop.so导致datanode报错


为了解决上一篇文章中提到的Bug,我们将线上的CDH5升级到了目前最新的CDH5.2.0,但升级之后,有一部分服务器的datanode不能正常启动,报错如下:

2014-11-20 19:54:52,071 WARN org.apache.hadoop.hdfs.server.datanode.DataNode: Unexpected exception in block pool Block pool <registering> (Datanode Uuid unassigned) service to idc1-server1/10.100.1.100:8020
com.google.common.util.concurrent.ExecutionError: java.lang.UnsatisfiedLinkError: org.apache.hadoop.io.nativeio.NativeIO.link0(Ljava/lang/String;Ljava/lang/String;)V
	at com.google.common.util.concurrent.Futures.wrapAndThrowExceptionOrError(Futures.java:1126)
	at com.google.common.util.concurrent.Futures.get(Futures.java:1048)
	at org.apache.hadoop.hdfs.server.datanode.DataStorage.linkBlocks(DataStorage.java:870)
	at org.apache.hadoop.hdfs.server.datanode.BlockPoolSliceStorage.linkAllBlocks (BlockPoolSliceStorage.java:570)
	at org.apache.hadoop.hdfs.server.datanode.BlockPoolSliceStorage.doUpgrade (BlockPoolSliceStorage.java:379)
	at org.apache.hadoop.hdfs.server.datanode.BlockPoolSliceStorage.doTransition (BlockPoolSliceStorage.java:313)
	at org.apache.hadoop.hdfs.server.datanode.BlockPoolSliceStorage.recoverTransitionRead (BlockPoolSliceStorage.java:187)
	at org.apache.hadoop.hdfs.server.datanode.DataStorage.recoverTransitionRead (DataStorage.java:309)
	at org.apache.hadoop.hdfs.server.datanode.DataNode.initStorage(DataNode.java:1109)
	at org.apache.hadoop.hdfs.server.datanode.DataNode.initBlockPool(DataNode.java:1080)
	at org.apache.hadoop.hdfs.server.datanode.BPOfferService.verifyAndSetNamespaceInfo (BPOfferService.java:320)
	at org.apache.hadoop.hdfs.server.datanode.BPServiceActor.connectToNNAndHandshake (BPServiceActor.java:220)
	at org.apache.hadoop.hdfs.server.datanode.BPServiceActor.run(BPServiceActor.java:824)
	at java.lang.Thread.run(Thread.java:744)
Caused by: java.lang.UnsatisfiedLinkError: org.apache.hadoop.io.nativeio.NativeIO.link0 (Ljava/lang/String;Ljava/lang/String;)V
	at org.apache.hadoop.io.nativeio.NativeIO.link0(Native Method)
	at org.apache.hadoop.io.nativeio.NativeIO.link(NativeIO.java:838)
	at org.apache.hadoop.hdfs.server.datanode.DataStorage$2.call(DataStorage.java:862)
	at org.apache.hadoop.hdfs.server.datanode.DataStorage$2.call(DataStorage.java:855)
	at java.util.concurrent.FutureTask.run(FutureTask.java:262)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
	... 1 more
2014-11-20 19:54:52,073 WARN org.apache.hadoop.hdfs.server.datanode.DataNode: Ending block pool service for: Block pool <registering> (Datanode Uuid unassigned) service to idc1-server1/10.100.1.100:8020

但搜遍了Google也未能找到匹配的信息,唯一沾点边的都是一些在Windows平台上因为缺少lib导致的问题。
而在我们的环境中,只有一部分的服务器有以上问题,对比了所有Hadoop相关的软件包之后都没法发现有什么不同,这给我们分析问题带来了很大的干扰。

最后,我们尝试通过strace来跟踪datanode的进程。
yum install strace
strace -f -F -o /tmp/strace.output.txt /etc/init.d/hadoop-hdfs-datanode start
lsof | grep libhadoop.so

java 18527 hdfs mem REG 253,0 122832 270200 /usr/java/jdk1.7.0_45/jre/lib/amd64/libhadoop.so

发现它读取了一个lib文件:/usr/java/jdk1.7.0_45/jre/lib/amd64/libhadoop.so,而其它正常的服务器的datanode进程则是读取的/usr/lib/hadoop/lib/native/libhadoop.so。
经过验证发现/usr/java/jdk1.7.0_45/jre/lib/amd64/libhadoop.so是在安装Snappy软件包时创建的,在移走了它之后,datanode终于正常启动了。

看来,虽然datanode在启动时指定了 -Djava.library.path=/usr/lib/hadoop/lib/native,但jre中的lib被载入的优先级还是要高一些。

,

  1. No comments yet.
(will not be published)
*