集群服务器重启后恢复操作
1. master无需处理
2. client:
// 每台节点需要配置永久挂载,vi /etc/fstab
mount -a // 挂载生效
systemctl restart munge
systemctl restart ypbind
systemctl restart slurmd
3. master
sinfo 查询状态,如果是unkonown或者别的状态需要update
scontrol update Node=client[01-05] STATE=RESUME