配置shell脚本定期清理集群日志和监控数据，并对清理完Cloudera Management Service的监控数据后的问题解决
清理日志和监控数据的shell脚本
#配置shell自动删除
#!/bin/bash
# 清理日志
cd /var/log/hadoop-mapreduce
rm -rf *.out.*
rm -rf *.log.*
cd /var/log/hadoop-hdfs
rm -rf *.out.*
rm -rf *.log.*
cd /var/log/zookeeper
rm -rf *.out.*
rm -rf *.log.*
cd /var/log/hive
rm -rf *.out.*
rm -rf *.log.*
cd /var/log/cloudera-scm-agent
rm -rf *.out.*
rm -rf *.log.*
cd /var/log/audit
rm -rf *.out.*
rm -rf *.log.*
cd /var/log/hadoop-yarn
rm -rf *.out.*
rm -rf *.log.*
cd /var/log/hue-httpd
rm -rf *.out.*
rm -rf *.log.*
# 清理监控日志
cd /data01/var/lib/cloudera-host-monitor/ts/type/partitions
rm -rf type*
cd /data01/var/lib/cloudera-host-monitor/ts/stream/partitions
rm -rf stream*
cd /data01/var/lib/cloudera-host-monitor/ts/ts_stream_rollup_PT600S/partitions/
rm -rf ts_stream*
cd /data01/var/lib/cloudera-host-monitor/ts/ts_type_rollup_PT600S/partitions/
rm -rf ts_type*
cd /data01/var/lib/cloudera-service-monitor/ts/stream/partitions/
rm -rf stream*
cd /data01/var/lib/cloudera-service-monitor/ts/type/partitions/
rm -rf type*
cd /data01/var/lib/cloudera-service-monitor/ts/ts_stream_rollup_PT600S/partitions/
rm -rf ts_stream*
cd /data01/var/lib/cloudera-service-monitor/ts/ts_type_rollup_PT600S/partitions/
rm -rf ts_type*
清理完监控数据后的问题描述：
清理完监控数据后，发现，集群网络IO，磁盘IO，HDFS IO，集群CPU等监控状态会是无数据的状态
问题解决：
1、到Cloudera Manager首页界面，Cloudera页面最下面Cloudera Management Service->停止；
2、打开集群
在主节点输入：
[root@lk-bd01 ~]#service cloudera-scm-server stop
把SCM服务关闭
进入scm-agent安装的目录在每台节点输入：
[root@lk-bd01 ~]#/data01/opt/cm/etc/init.d/cloudera-scm-agent stop
把每台的agent都关闭
3、进入cloudera-service-monitor的安装目录
[root@lk-bd01 ~]#mv /var/lib/cloudera-service-monitor /var/lib/cloudera-service-monitor.moved
[root@lk-bd01 ~]#mv /var/lib/cloudera-host-monitor /var/lib/cloudera-host-monitor.moved
对cloudera-service-monitor和cloudera-host-monitor改一个名字。目的是为了做个备份，因为重启cloudera-scm-server服务的时候会重新生成新的目录，生成完新的目录之后cloudera-service-monitor.moved和cloudera-host-monitor.moved就可以进行删除了。
4、启动每一台的agent
[root@lk-bd01 ~]#/data01/opt/cm/etc/init.d/cloudera-scm-agent start
5、启动主节点的cloudera-scm-server服务
[root@lk-bd01 ~]#service cloudera-scm-server start
6、再次登入SM的管理界面之后，Cloudera页面最下面Cloudera Management Service->启动
监控状态这时就会有数据显示了
