我已经将Zeppelin 0.5.5二进制文件下载到redhat 6.3实例中。我还成功地构建了spark 1.5.2。现在，我想让齐柏林飞船使用火花版本，而不是齐柏林飞船附带的预编译版本。
为此，我编辑了conf/，使其只包含以下未注释的行：
export SPARK_HOME=/home/eron/spark-1.5.2/
我还编辑了齐柏林飞船GUI，所以现在我有了主集：
master = spark://master:7077
然后我点击重新启动，并尝试再次浏览教程：
val bankText = sc.textFile("/mnt/zeppelin-0.5.5-incubating-bin-all/bank.csv")
org.apache.fract.transport.TransportExceptionorg.apache.srift.transport.TIOStreamTransport.read（TIOStreamTransport.java:132）网址：org.apache.srift.transport.readAll（TTransport.java:86）在org.apache.srift.protocol.TBinaryProtocol.readAll（TBinaryProtocol.java:429）在org.apache.srift.protocol.TBinaryProtocol.readI32（TBinaryProtocol.java:318）在org.apache.srift.protocol.TBinaryProtocol.readMessageBegin（TBinaryProtocol.java:219）在org.apache.srift.TServiceClient.rereceiveBase（TServiceClient.java:69）在org.apache.zeppelin.pinterpreter.frift.RemoteInterpreterService$Client.recv_enterpret（RemoteInterpreter Service.java:220）在org.apache.zeppelin.expreter.frift.RemoteInterpreterService$Client.depreter（RemoteInterpreterService.java:205）在org.apache.zeppelin.printerpreter.remote.RemoteInterpreter.printerpret（RemoteInterpreter.java:211）在org.apache.zeppelin.printerpret.LazyOpenInterpreter.printerpret（LazyOpenInterpreter.java:93）网址：org.apache.zeppelin.notebook。Paragraph。jobRun（Paragraph。java:207）在org.apache.zeppelinscheduler.Job.run（Job.java:170）org.apache.zeppelinscheduler.RemoteScheduler$JobRunner.run（RemoteScheduler.java:304）在java.util.concurrent.Executors$RunnableAdapter.call（Executitors.java:471）位于java.util.concurrent.FFutureTask.run（FutureTask.java:262）java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201（ScheduledThreadPoolExecutor.java:178）在java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run（ScheduledThreadPoolExecutor.java:292）在java.util.concurrent.ThreadPoolExecutiator.runWorker（ThreadPoolExecutiator.java:1145）在java.util.concurrent.ThreadPoolExecutor$Worker.run（ThreadPoolExecutor.java:615）在java.lang.Thread.run（Thread.java:745）
另外，只是做
sc
收益率：
java.net.ConnectException:在拒绝连接java.net.PlainSocketImpl.socketConnect（本机方法）java.net.AbstractPlainSocketImpl.doConnect（AbstractPlasinSocketimpl.java:339）在java.net.AbstractPlainSocketImpl.connectToAddress（AbstractPlasinSocketimpl.java:200）在java.net.AbstractPlainSocketImpl.connect（AbstractPlasinSocketimpl.java:182）位于java.net.SocksSocketImpl.connect（SocksSocketImpl.java:392）java.net.Socket.connect（Socket.java:579）位于org.apache.srift.transport.TSocket.open（TSocket.java:182）org.apache.zeppelin.pinterpreter.remote.ClientFactory.create（ClientFactory.java:51）在org.apache.zeppelin.pinterpreter.remote.ClientFactory.create（ClientFactory.java:37）在org.apache.commons.pool2.BasePooledObjectFactory.makeObject（BasePooledObjectFactory.java:60）在org.apache.commons.pool2.impl.GenericObjectPool.create（GenericObjectPool.java:861）在org.apache.commons.pool2.impl.GenericObjectPool.boroObject（GenericObjectPool.java:435）在org.apache.commons.pool2.impl.GenericObjectPool.boroObject（GenericObjectPool.java:363）在org.apache.zeppelin.pinterpreter.remote.RemoteInterpreterProcess.getClient（RemoteInterpreter Process.java:139）在org.apache.zeppelin.printerpreter.remote.RemoteInterpreter.printerpret（RemoteInterpreter.java：192）在org.apache.zeppelin.printerpret.LazyOpenInterpreter.printerpret（LazyOpenInterpreter.java:93）网址：org.apache.zeppelin.notebook。Paragraph。jobRun（Paragraph。java:207）在org.apache.zeppelinscheduler.Job.run（Job.java:170）org.apache.zeppelinscheduler.RemoteScheduler$JobRunner.run（RemoteScheduler.java:304）在java.util.concurrent.Executors$RunnableAdapter.call（Executitors.java:471）位于java.util.concurrent.FFutureTask.run（FutureTask.java:262）java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201（ScheduledThreadPoolExecutor.java:178）在java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run（ScheduledThreadPoolExecutor.java:292）在java.util.concurrent.ThreadPoolExecutiator.runWorker（ThreadPoolExecutiator.java:1145）在java.util.concurrent.ThreadPoolExecutor$Worker.run（ThreadPoolExecutor.java:615）在java.lang.Thread.run（Thread.java:745）上花费了0秒。
你知道我做错了什么吗？
版本的齐柏林飞船0.5.5将无法与高于1.5.0的任何Apache spark版本一起使用，因此您需要从github下载齐柏林飞船0.6.0，并使用命令"mvn-clean package-Pspark-1.5-DskipTests"重新编译。默认情况下，齐柏林飞船0.5构建spark 1.4。