Running Shark Locally 及可能出现的问题

cl_303303303

Shark本地安装

1.下载scala
wget http://www.scala-lang.org/files/archive/scala-2.9.3.tgz
最新有2.10.2.tgz文件
tar xvfz scala-2.9.3.tgz

2.下载shark and hive压缩包
wget http://spark-project.org/download/shark-0.7.0-hadoop1-bin.tgz (cdh3)
tar xvfz shark-0.7.0-*-bin.tgz

3. 配置环境变量
cd shark-0.7.0/conf
cp shark-env.sh.template shark-env.sh
vi shark-env.sh
export HIVE_HOME=/path/to/hive-0.9.0-bin
export SCALA_HOME=/path/to/scala-2.9.3

4.测试数据
CREATE TABLE src(key INT, value STRING);
LOAD DATA LOCAL INPATH '${env:HIVE_HOME}/examples/files/kv1.txt' INTO TABLE src;
SELECT COUNT(1) FROM src;
OK
500
Time taken: 2.149 seconds
没有了hive中的mr，速度快了不少
CREATE TABLE src_cached AS SELECT * FROM SRC;
SELECT COUNT(1) FROM src_cached;

安装过程中可能出现的问题及解决

1.CREATE TABLE src(key INT, value STRING);
FAILED: Error in metadata: MetaException(message:Got exception: org.apache.hadoop.ipc.RPC$VersionMismatch Protocol org.apache.hadoop.hdfs.protocol.ClientProtocol version

mismatch. (client = 61, server = 63))
FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.DDLTask

ERROR exec.Task: FAILED: Error in metadata: MetaException(message:Got exception: org.apache.hadoop.ipc.RPC$VersionMismatch Protocol

org.apache.hadoop.hdfs.protocol.ClientProtocol version mismatch. (client = 61, server = 63))
org.apache.hadoop.hive.ql.metadata.HiveException: MetaException(message:Got exception: org.apache.hadoop.ipc.RPC$VersionMismatch Protocol

org.apache.hadoop.hdfs.protocol.ClientProtocol version mismatch. (client = 61, server = 63))
        at org.apache.hadoop.hive.ql.metadata.Hive.createTable(Hive.java:544)
        at org.apache.hadoop.hive.ql.exec.DDLTask.createTable(DDLTask.java:3313)
        at org.apache.hadoop.hive.ql.exec.DDLTask.execute(DDLTask.java:242)
        at org.apache.hadoop.hive.ql.exec.Task.executeTask(Task.java:134)
        at org.apache.hadoop.hive.ql.exec.TaskRunner.runSequential(TaskRunner.java:57)
        at org.apache.hadoop.hive.ql.Driver.launchTask(Driver.java:1312)
        at org.apache.hadoop.hive.ql.Driver.execute(Driver.java:1104)
        at org.apache.hadoop.hive.ql.Driver.run(Driver.java:937)
        at shark.SharkCliDriver.processCmd(SharkCliDriver.scala:288)
        at org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:406)
        at org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:341)
        at shark.SharkCliDriver$.main(SharkCliDriver.scala:203)
        at shark.SharkCliDriver.main(SharkCliDriver.scala)
reason：Hadoop版本与SHARK的Hadoop core jar包版本不一致引起的。
解决：将${HADOOP_HOME}/hadoop-core-*.jar copy 到${SHARK_HOME}/lib_managed/jars/org.apache.hadoop/hadoop-core/目录下面，rm原来的hadoop-core-*.jar
重新进入Shark

2.出现java.lang.NoClassDefFoundError
/app/hadoop/shark/shark-0.7.0/lib_managed/jars/org.apache.hadoop/hadoop-core/
java.lang.NoClassDefFoundError: org/apache/hadoop/thirdparty/guava/common/collect/LinkedListMultimap
        at org.apache.hadoop.hdfs.SocketCache.<init>(SocketCache.java:48)
        at org.apache.hadoop.hdfs.DFSClient.<init>(DFSClient.java:253)
        at org.apache.hadoop.hdfs.DFSClient.<init>(DFSClient.java:220)
        at org.apache.hadoop.hdfs.DistributedFileSystem.initialize(DistributedFileSystem.java:89)
        at org.apache.hadoop.fs.FileSystem.createFileSystem(FileSystem.java:1611)
        at org.apache.hadoop.fs.FileSystem.access$300(FileSystem.java:68)
        at org.apache.hadoop.fs.FileSystem$Cache.getInternal(FileSystem.java:1645)
        at org.apache.hadoop.fs.FileSystem$Cache.get(FileSystem.java:1627)
        at org.apache.hadoop.fs.FileSystem.get(FileSystem.java:254)
        at org.apache.hadoop.fs.FileSystem.get(FileSystem.java:123)
        at org.apache.hadoop.fs.FileSystem.get(FileSystem.java:238)
        at org.apache.hadoop.fs.Path.getFileSystem(Path.java:183)
        at org.apache.hadoop.hive.metastore.Warehouse.getFs(Warehouse.java:104)
        at org.apache.hadoop.hive.metastore.Warehouse.getDnsPath(Warehouse.java:136)
        at org.apache.hadoop.hive.metastore.Warehouse.getWhRoot(Warehouse.java:151)
        at org.apache.hadoop.hive.metastore.HiveMetaStore$HMSHandler.getDefaultDatabasePath(HiveMetaStore.java:475)
        at org.apache.hadoop.hive.metastore.HiveMetaStore$HMSHandler.createDefaultDB_core(HiveMetaStore.java:353)
        at org.apache.hadoop.hive.metastore.HiveMetaStore$HMSHandler.createDefaultDB(HiveMetaStore.java:371)
        at org.apache.hadoop.hive.metastore.HiveMetaStore$HMSHandler.init(HiveMetaStore.java:278)
        at org.apache.hadoop.hive.metastore.HiveMetaStore$HMSHandler.<init>(HiveMetaStore.java:248)
        at org.apache.hadoop.hive.metastore.HiveMetaStoreClient.<init>(HiveMetaStoreClient.java:114)
        at org.apache.hadoop.hive.ql.metadata.Hive.createMetaStoreClient(Hive.java:2092)
        at org.apache.hadoop.hive.ql.metadata.Hive.getMSC(Hive.java:2102)
        at org.apache.hadoop.hive.ql.metadata.Hive.createTable(Hive.java:538)
        at org.apache.hadoop.hive.ql.exec.DDLTask.createTable(DDLTask.java:3313)
        at org.apache.hadoop.hive.ql.exec.DDLTask.execute(DDLTask.java:242)
        at org.apache.hadoop.hive.ql.exec.Task.executeTask(Task.java:134)
        at org.apache.hadoop.hive.ql.exec.TaskRunner.runSequential(TaskRunner.java:57)
        at org.apache.hadoop.hive.ql.Driver.launchTask(Driver.java:1312)
        at org.apache.hadoop.hive.ql.Driver.execute(Driver.java:1104)
        at org.apache.hadoop.hive.ql.Driver.run(Driver.java:937)
        at shark.SharkCliDriver.processCmd(SharkCliDriver.scala:288)
        at org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:406)
        at org.apache.hadoop.hive.cli.CliDriver.processLine(CliDriver.java:341)
        at shark.SharkCliDriver$.main(SharkCliDriver.scala:203)
        at shark.SharkCliDriver.main(SharkCliDriver.scala)
Caused by: java.lang.ClassNotFoundException: org.apache.hadoop.thirdparty.guava.common.collect.LinkedListMultimap
        at java.net.URLClassLoader$1.run(URLClassLoader.java:202)
        at java.security.AccessController.doPrivileged(Native Method)
        at java.net.URLClassLoader.findClass(URLClassLoader.java:190)
        at java.lang.ClassLoader.loadClass(ClassLoader.java:307)
        at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:301)
        at java.lang.ClassLoader.loadClass(ClassLoader.java:248)
        ... 36 more
reason：CDH版本的缺少一个第三方包guava-*.jar
解决：建一个目录${SHARK_HOME}/lib_managed/jars/org.apache.hadoop/thirdparty，拷贝${HADOOP_HOME}/lib/guava-r09-jarjar.jar到这个目录
重新进入Shark

3.show tables出现问题
Failed with exception java.io.IOException:java.io.IOException: Cannot create an instance of InputFormat class org.apache.hadoop.mapred.TextInputFormat as specified in

mapredWork!
reason：缺少hadoop-lzo-*.jar引起的
解决：建一个目录${SHARK_HOME}/lib_managed/jars/org.apache.hadoop/lib, 拷贝${HADOOP_HOME}/lib/hadoop-lzo-*.jar到这个目录
重新进入Shark

4.SELECT count(1) FROM src_cached出现问题
spark.SparkException: Job failed: ShuffleMapTask(6, 0) failed: ExceptionFailure(java.lang.NoSuchMethodError: sun.misc.Unsafe.copyMemory(Ljava/lang/Object;JLjava/lang/Object;JJ)
V)at spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:642)
        at spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:640)
        at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:60)
        at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:47)
        at spark.scheduler.DAGScheduler.abortStage(DAGScheduler.scala:640)
        at spark.scheduler.DAGScheduler.handleTaskCompletion(DAGScheduler.scala:601)
        at spark.scheduler.DAGScheduler.processEvent(DAGScheduler.scala:300)
        at spark.scheduler.DAGScheduler.spark$scheduler$DAGScheduler$$run(DAGScheduler.scala:364)
        at spark.scheduler.DAGScheduler$$anon$1.run(DAGScheduler.scala:107)
FAILED: Execution Error, return code -101 from shark.execution.SparkTask

reason：java1.6版本低，需要安装jdk7.
解决：安装jdk7, JAVA_HOME指向新的JDK7,问题解决
tar xvfz jdk-7u25-linux-x64.tar.gz -C /usr/java/
export JAVA_HOME=/usr/java/jdk1.7.0_25
export CLASSPATH=/usr/java/jdk1.7.0_25/lib
重新进入Shark