大数据运维_大数据运维命令

一、运维操作

1、HDFS

HDFS 下线节点

https://blog.csdn.net/weixin_44758876/article/details/122535840
https://blog.csdn.net/moyefeng198436/article/details/113652813

2、HBase

2.1、HBase hbck

hbase org.apache.hbase.HBCK2 bypass -or 111887 111888
hbase org.apache.hbase.HBCK2
https://developer.aliyun.com/article/683107

2.2、快速拉起一个HBase集群

更换HBase使用的HDFS和ZK路径

<property>
    <name>hbase.rootdir</name>
    <value>hdfs://hacluster3/hbase_20231204</value>
</property>

<property>
    <name>zookeeper.znode.parent</name>
    <value>/hbase_20231204</value>
</property>
1
2
3
4
5
6
7
8
9

二、运维操作

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
进程名称：org.apache.hadoop.yarn.server.webproxy.WebAppProxyServer

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

主备切换
yarn rmadmin -transitionToStandby rm1

yarn rmadmin -getServiceState rm1
yarn rmadmin -getServiceState rm2
1
2
3
4
5
6
7
8
9
10
11
12

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

参考 https://blog.csdn.net/qq_37171694/article/details/128070374

客户端连接

连接指定 Hive metastore 服务端
hive --hiveconf hive.metastore.uris=thrift://XXX.hadoop.com:3316
1
2

7、Spark

## spark-sql
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
IP：XX.10.29.98
进程名称：org.apache.spark.deploy.history.HistoryServer
端口：18080

## Yarn JobHistoryServer
    /usr/local/fqlhadoop/hadoop/sbin/mr-jobhistory-daemon.sh start historyserver
IP：XX.10.29.98
进程名称：org.apache.hadoop.mapreduce.v2.hs.JobHistoryServer
端口：19888
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

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