大数据日常运维命令

1、HDFS NameNode

/usr/local/fqlhadoop/hadoop/sbin/hadoop-daemon.sh start namenode
/usr/local/fqlhadoop/hadoop/sbin/hadoop-daemon.sh stop namenode

bin/hdfs haadmin -DFSHAAdmin -getServiceState n1

2、HDFS DataNode

/usr/local/fqlhadoop/hadoop/sbin/hadoop-daemon.sh start datanode
/usr/local/fqlhadoop/hadoop/sbin/hadoop-daemon.sh stop datanode

bin/yarn rmadmin -getServiceState rm1

3、Yarn ResourceManager

/usr/local/fqlhadoop/hadoop/sbin/yarn-daemon.sh start resourcemanager
/usr/local/fqlhadoop/hadoop/sbin/yarn-daemon.sh stop resourcemanager

4、Yarn NodeManager

/usr/local/fqlhadoop/hadoop/sbin/yarn-daemon.sh start nodemanager
/usr/local/fqlhadoop/hadoop/sbin/yarn-daemon.sh stop nodemanager

5、Yarn WebAppProxyServer

/usr/local/fqlhadoop/hadoop/sbin/yarn-daemon.sh stop proxyserver
/usr/local/fqlhadoop/hadoop/sbin/yarn-daemon.sh start proxyserver

X、Yarn 运维命令

yarn application -list
yarn application -queue root.lx_fdw
yarn application -list -queue lx_fdw -appStates ACCEPTED
yarn application -list -queue lx_fdw -appStates ACCEPTED | grep thrift_XXX.hadoop.com_10017

yarn application -kill application_1652347896268_18628697

6、Hive

服务端启动

6.1 hive metastore

cd /usr/local/fqlhadoop/hive/
bin/hive --service metastore -p 3316 &
ps -ef | grep ‘org.apache.hadoop.hive.metastore.HiveMetaStore’ | grep -v grep | awk ‘{print $2}’ | xargs kill -9

6.2 hive service

bin/hive --service hiveserver2 &

关联的配置，vim conf/hive-site.xml
hive.server2.thrift.port、hive.server2.thrift.bind.host

扫描二维码关注公众号，回复： 16749027 查看本文章

客户端连接
连接指定 Hive metastore 服务端
hive --hiveconf hive.metastore.uris=thrift://XXX.hadoop.com:3316

7、Spark

spark-sql --master=yarn --queue lx_etl --driver-memory 1g --driver-java-options '-XX:MetaspaceSize=1g -XX:MaxMetaspaceSize=1g' --num-executors 1  --executor-memory 1g --executor-cores 1   --conf spark.yarn.am.memory=2048m --hiveconf hive.cli.print.header=false 


## Spark HistoryServer
/usr/local/fqlhadoop/spark/sbin/start-history-server.sh

/usr/local/fqlhadoop/spark/sbin/start-history-server.sh hdfs://hacluster/sparklog
/usr/local/fqlhadoop/spark/sbin/stop-history-server.sh hdfs://hacluster/sparklog

进程名称：
org.apache.spark.deploy.history.HistoryServer



## Yarn JobHistoryServer
    /usr/local/fqlhadoop/hadoop/sbin/mr-jobhistory-daemon.sh start historyserver

8、HBase

/usr/local/fqlhadoop/hbase/bin/hbase-daemon.sh stop master
/usr/local/fqlhadoop/hbase/bin/hbase-daemon.sh start master

/usr/local/fqlhadoop/hbase/bin/hbase-daemon.sh stop regionserver
/usr/local/fqlhadoop/hbase/bin/hbase-daemon.sh start regionserver