18Spark on Yarn配置日志Web UI(HistoryServer服务)

Veröffentlicht am | Bearbeitet am | in | Aufrufe:

1.进入spark目录和配置文件

1
2
[root@sht-sgmhadoopnn-01 ~]# cd /root/learnproject/app/spark/conf
[root@sht-sgmhadoopnn-01 conf]# cp spark-defaults.conf.template spark-defaults.conf

2.创建spark-history的存储日志路径为hdfs上(当然也可以在linux文件系统上)

1
2
3
4
5
6
7
8
9
10
11
 [root@sht-sgmhadoopnn-01 conf]# hdfs dfs -ls /
 Found 3 items
 drwxr-xr-x   - root root          0 2017-02-14 12:43 /spark
 drwxrwx---   - root root          0 2017-02-14 12:58 /tmp
 drwxr-xr-x   - root root          0 2017-02-14 12:58 /user
 You have new mail in /var/spool/mail/root
 [root@sht-sgmhadoopnn-01 conf]# hdfs dfs -ls /spark
 Found 1 items
 drwxrwxrwx   - root root          0 2017-02-15 21:44 /spark/checkpointdata
[root@sht-sgmhadoopnn-01 conf]# hdfs dfs -mkdir /spark/historylog
#在HDFS中创建一个目录，用于保存Spark运行日志信息。Spark History Server从此目录中读取日志信息

3.配置

1
2
3
4
5
6
7
8
[root@sht-sgmhadoopnn-01 conf]# vi spark-defaults.conf
spark.eventLog.enabled           true
spark.eventLog.compress          true
spark.eventLog.dir               hdfs://nameservice1/spark/historylog
spark.yarn.historyServer.address 172.16.101.55:18080
#spark.eventLog.dir保存日志相关信息的路径，可以是hdfs://开头的HDFS路径，也可以是file://开头的本地路径，都需要提前创建
#spark.yarn.historyServer.address : Spark history server的地址(不加http://).
这个地址会在Spark应用程序完成后提交给YARN RM，然后可以在RM UI上点击链接跳转到history server UI上.

4.添加SPARK_HISTORY_OPTS参数

1
2
3
4
5
6
7
8
9
10
11
12
 [root@sht-sgmhadoopnn-01 conf]# vi spark-env.sh
 #!/usr/bin/env bash
 
 export SCALA_HOME=/root/learnproject/app/scala
 export JAVA_HOME=/usr/java/jdk1.8.0_111
 export SPARK_MASTER_IP=172.16.101.55
 export SPARK_WORKER_MEMORY=1g
 export SPARK_PID_DIR=/root/learnproject/app/pid
 export HADOOP_CONF_DIR=/root/learnproject/app/hadoop/etc/hadoop
export SPARK_HISTORY_OPTS="-Dspark.history.fs.logDirectory=hdfs://mycluster/spark/historylog \
-Dspark.history.ui.port=18080 \
-Dspark.history.retainedApplications=20"

5.启动服务和查看

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
 [root@sht-sgmhadoopnn-01 spark]# ./sbin/start-history-server.sh 
 starting org.apache.spark.deploy.history.HistoryServer, logging to /root/learnproject/app/spark/logs/spark-root-org.apache.spark.deploy.history.HistoryServer-1-sht-sgmhadoopnn-01.out
 [root@sht-sgmhadoopnn-01 ~]# jps
 28905 HistoryServer
 30407 ProdServerStart
 30373 ResourceManager
 30957 NameNode
 16949 Jps
 30280 DFSZKFailoverController
31445 JobHistoryServer
[root@sht-sgmhadoopnn-01 ~]# ps -ef|grep spark
root     17283 16928  0 21:42 pts/2    00:00:00 grep spark
root     28905     1  0 Feb16 ?        00:09:11 /usr/java/jdk1.8.0_111/bin/java -cp /root/learnproject/app/spark/conf/:/root/learnproject/app/spark/jars/*:/root/learnproject/app/hadoop/etc/hadoop/ -Dspark.history.fs.logDirectory=hdfs://mycluster/spark/historylog -Dspark.history.ui.port=18080 -Dspark.history.retainedApplications=20 -Xmx1g org.apache.spark.deploy.history.HistoryServer
You have new mail in /var/spool/mail/root
[root@sht-sgmhadoopnn-01 ~]# netstat -nlp|grep 28905
tcp        0      0 0.0.0.0:18080               0.0.0.0:*                   LISTEN      28905/java          
[root@sht-sgmhadoopnn-01 ~]#