RAC集群状态异常

某日巡检发现，4节点RAC的3个节点集群状态异常
查看节点的集群状态
grid@p720hi4:/home/grid$ crs_stat -t
CRS-0184: Cannot communicate with the CRS daemon.

集群资源正常，数据库正常

解决：
grid@p720hi4:crsctl stop has -f
等个5分钟

grid@p720hi4:/u01/app/11.2.0/grid/bin$ ps -ef | grep has
grid 9371778 40304832 0 00:23:29 pts/1 0:00 tail -f ohasd.log
root 9764934 1 0 Apr 09 - 0:00 /bin/sh /etc/init.ohasd run

等has进程没有后，启动has

先启动has，再启动cluster
crsctl start has
crsctl start cluster

集群节点报错： [client(37027852)]CRS-10001:08-Sep-18 01:25 ACFS-9253: Failed
to unmount mount point ‘/ogg’. Mount point likely in use.
[client(37027854)]CRS-10001:08-Sep-18 01:25 ACFS-9254: Manual
intervention is required. [client(37027870)]CRS-10001:08-Sep-18 01:25
ACFS-9252: The following process IDs have open references on mount
point ‘/ogg’:

解决思路：
1、查看是否有人占用ogg目录，比如有会话在GGSCI>中；
2、查看acfs资源是否启动
crsctl stat res -t