spark集群配置细则总结

修改目录与目录组：

sudo chown -R hadoop:hadoop spark-1.6.1-bin-hadoop2.6

sudo chown -R hadoop:hadoop jdk1.8.0_101


sudo chown -R hadoop:hadoop scala2.11.6


1./etc目录下

vi hosts

192.168.xxx.xxx data6(master节点)

192.168.xxx.xxx data2(worker节点)

192.168.xxx.xxx data3(worker节点)

2.spark/conf/目录下

vi slaves

data6

data2

data3

vi spark-env

export JAVA_HOME=/app/jdk1.7

export SPARK_MASTER_IP=data6

export SPARK_WORKER_INSTANCES=1

export SPARK_WORKER_MEMORY=30g

export SPARK_WORKER_CORES=6

export SPARK_LOG_DIR=/data/tmp

export SPARK_PID_DIR=/data/tmp

export SPARK_DAEMON_JAVA_OPTS="-Djava.io.tmpdir=/home/tmp"

export PYSPARK_PYTHON=/opt/anaconda3/bin/python3

export PYSPARK_DRIVER_PYTHON=/opt/anaconda3/bin/ipython3

export PYSPARK_DRIVER_PYTHON_OPTS="notebook --ip 0.0.0.0 --port 9999"

export PATH=$PATH:/usr/local/bin

export SPARK_CLASSPATH=/app/spark-1.6.1/lib/spark-examples-1.6.1-hadoop2.4.0.jar:/app/spark-1.6.1/lib/spark-assembly-1.6.1-hadoop2.4.0.jar:/app/spark-1.6.1/lib/spark-1.6.1-yarn-shuffle.jar:/app/spark-1.6.1/lib/nlp-lang-1.5.jar:/app/spark-1.6.1/lib/mysql-connector-java-5.1.26-bin.jar:/app/spark-1.6.1/lib/datanucleus-rdbms-3.2.9.jar:/app/spark-1.6.1/lib/datanucleus-core-3.2.10.jar:/app/spark-1.6.1/lib/datanucleus-api-jdo-3.2.6.jar:/app/spark-1.6.1/lib/ansj_seg-3.7.3-all-in-one.jar

vi hive-site.xml

<configuration>

<property>

<name>hive.metastore.uris</name>

<value>thirft://data6:9083</value>

<description>Thrift URI for the remote metastore. Used by metastore client to connect to remote metastore.</description>

</property>

<property>

<name>hive.server2.thrift.min.worker.threads</name>

<value>5</value>

<description>maximum number of Thrift worker threads</description>

</property>

<property>

<name>hive.server2.thrift.port</name>

<value>500</value>

<description>Port number of HiveSercer2 Thrift interfaace.</description>

</property>

<property>

<name>hive.server2.thrift.min.worker.threads</name>

<value>11000</value>

<description>Thrift URI for the remote metastore. Used by metastore client to connect to remote metastore.</description>

</property>

<property>

<name>hive.server2.thrift.bind.host</name>

<value>data6</value>

<description>bind host on which to run the HiveSercer2 Thrift interface</description>

</property>

<property>

<name>mapred.reduce.tasks</name>

<value>40</value>

</property>

vi log4j.properties

#Setting to quiet third party logs that are too verbose

log4j.logger.org.spark-project.jetty=WARN

log4j.logger.org.spark-project.jetty.util.component.AbstractLifeCycle=ERROR

log4j.logger.org.apache.spark.repl.SparkIMain$exprTyper=INFO

log4j.logger.org.apache.spark.repl.SparkILoop$SparkLoopInterpreter=INFO

log4j.logger.parquet=ERROR

#SPARK-9183:Setting to avoid annoying messages when looking up nonexitent UDFs in SparkSQL with Hive support

log4j.logger.org.apachce.hadoop.hive.metastore.RetryingHMSHandler=FATAL

log4j.logger.org.apachce.hadoop.hive.ql.exec.FunctionRegistry=ERROR

Spark集群搭建——SSH免密码验证登陆

机器准备

笔者有三台机器，左侧栏为ip，右侧为hostname，三台机器都有一个名为spark的用户。通过ping验证三台是可以通信的。

192.168.248.150 spark-master 192.168.248.153 ubuntu-worker 192.168.248.155 spark-worker1

根据机器信息加入到三台机器的/etc/hosts中。

配置

我们需要设置spark-master 能够免密码验证登陆ubuntu-worker、spark-worker1两台机器。

1. 安装ssh

   sudo apt-get install openssh-server

2. 生成秘钥

   执行命令ssh-keygen -t rsa，然后一直按回车键即可。

3. 复制spark-master结点的id_rsa.pub文件到另外两个结点：

   scp id_rsa.pub spark@ubuntu-worker:~/.ssh/

4. 到另外两个结点，将公钥加到用于认证的公钥文件中:

   cat id_rsa.pub >> authorized_keys

5. 修改两个worker的authorized_keys权限为600或者644、将.ssh文件权限改为700

   chmod 700 .ssh

   chmod 600  authorized_keys

6. 验证:

   登陆spark-master，在终端输入ssh ubuntu-worker,登陆成功则说明配置成功。

HDFS下载不成功问题

用windows上的IE来访问namenode节点的监控web页下载不了，需要修改了windows机器的C:WINDOWSsystem32driversetchosts文件，把hadoop集群中的几台机的主机名和IP地址加进去（一般在目录下/etc/hosts），让IE能解析就OK了。

Namenode没有启动问题，进行格式化

hadoop namenode -format

再启动HDFS