ceph分布式存储在生产或者实验环境,经常涉及到物理节点加入或者删除,本文仅对移除物理节点的相关步骤做了操作记录,以方便需要时查阅。
操作之前通过ceph -s确保整个集群状态是OK的,然后通过ceph osd tree查看需要移除的节点下面所有含有的osd,然后进行out操作,等待数据迁移完成后进行下一步
ceph osd out 124
ceph osd out 107
ceph osd out 90
systemctl stop ceph-osd.target
或者逐个停掉
systemctl stop ceph-osd@124
systemctl stop ceph-osd@107
systemctl stop ceph-osd@90
ceph osd rm 124
ceph osd rm 107
ceph osd rm 90
ceph osd crush rm osd.124
ceph osd crush rm osd.107
ceph osd crush rm osd.90
ceph auth del osd.124
ceph auth del osd.107
ceph auth del osd.90
ceph osd crush rm compute37
https://blog.csdn.net/nasooo/article/details/117947537
https://www.cnblogs.com/schangech/p/8036191.html
https://blog.csdn.net/weixin_43634280/article/details/88954840