先备份 HDFS/Hive 等业务数据卸载操作会永久删除数据生产环境优先走 CM 界面正常退役不要直接暴力 rm。所有节点都要执行 agent 清理CM 主节点额外处理 scm-server 和元数据库适用CDH5 / CDH6一、CM Web UI 先操作推荐优先做Parcel 集群必走1.停止所有集群服务2.停止Cloudera Management Service3.进入【Parcels】页面对 CDH parcel →停用4.删除集群、删除 Cloudera Management Service 集群二、所有节点停止CM Agent 清理目录# 1. 停止agent systemctl stop cloudera-scm-agent # 杀残留大数据进程可选防止占用 for u in cloudera-scm flume hadoop hdfs hbase hive httpfs hue impala mapred oozie solr spark sqoop yarn zookeeper; do pids$(ps -u $u -o pid 2/dev/null) [ -n $pids ] kill $pids done # 2. 卸载挂载点非常关键不umount删不掉 umount /var/run/cloudera-scm-agent/process # 3. yum删除CM包 yum -y remove cloudera-manager-* yum clean all # 4. 清理CM/Agent配置、日志、缓存 rm -rf /usr/share/cmf \ /var/lib/cloudera* \ /var/cache/yum/cloudera* \ /var/log/cloudera* \ /var/run/cloudera* \ /etc/cloudera* \ /tmp/.scm_prepare_node.lock \ /tmp/cmf* \ /tmp/scm* # 5. 清理CDH组件数据目录⚠️ 业务数据在这里确认备份再执行 rm -rf /var/lib/flume-ng \ /var/lib/hadoop* \ /var/lib/hue \ /var/lib/oozie \ /var/lib/solr \ /var/lib/sqoop* \ /var/lib/zookeeper \ /var/lib/impala \ /var/lib/hbase \ /var/lib/hive \ /var/lib/kudu \ /var/lib/spark \ /var/lib/yarn* \ /var/lib/sentry \ /var/lib/kafka \ /etc/hadoop* \ /etc/hbase* \ /etc/hive* \ /etc/hue \ /etc/impala \ /etc/kafka \ /etc/kudu \ /etc/flume-ng \ /etc/zookeeper \ /etc/spark \ /etc/solr \ /etc/sqoop \ /etc/sentry # 6. 删除Parcel安装目录/opt/cloudera是核心 rm -rf /opt/cloudera三、仅主节点停止CM Server# 停止CM Server systemctl stop cloudera-scm-server四、如果是自建MySQL数据库清理CM 元数据库DROP DATABASE amon; DROP DATABASE hue; DROP DATABASE nav; DROP DATABASE navms; DROP DATABASE oozie; DROP DATABASE rman; DROP DATABASE scm; DROP DATABASE sentry; DROP DATABASE metastore;