ceph集群删除一个包含mon、osd和mds的节点

1、移除mon

[root@ceph04 ~]# ceph mon remove ceph04
removed mon.ceph04 at 192.168.11.130:6789/0, there are now 2 monitors

2、移除此节点上所有的osd

1)、查看此节点的osd

[root@ceph04 ~]# ceph osd tree

-4      1.08                    host ceph04
8       0.27                            osd.8   up      1
9       0.27                            osd.9   up      1
10      0.27                           osd.10  up      1

2)、把节点上的osd服务停掉

[root@ceph04 ~]# systemctl stop osd

3)、再次查看ceph osd状态

[root@ceph04 ~]# ceph osd tree

-4      1.08                    host ceph04
8       0.27                            osd.8   down    1
9       0.27                            osd.9   down    1
10      0.27                           osd.10  down    1

4)、删除所有的osd

[root@ceph04 ~]# ceph osd rm 8
removed osd.8

[root@ceph04 ~]# ceph osd rm 9
removed osd.9

[root@ceph04 ~]# ceph osd rm 10
removed osd.10

5)、删除所有osd的crush map

[root@ceph04 ~]# ceph osd crush rm osd.8
removed item id 8 name 'osd.8' from crush map

[root@ceph04 ~]# ceph osd crush rm osd.9
removed item id 9 name 'osd.9' from crush map

[root@ceph04 ~]# ceph osd crush rm osd.10
removed item id 10 name 'osd.10' from crush map

6)、删除所有osd的认证

[root@ceph04 ~]# ceph auth del osd.8
updated

[root@ceph04 ~]# ceph auth del osd.9
updated

[root@ceph04 ~]# ceph auth del osd.10
updated

7)、在ceph osd tree中删除此节点的crush map

[root@ceph04 ~]# ceph osd crush rm  ceph04
removed item id -4 name 'ceph04' from crush map

8）、卸载所有挂载在osd的硬盘

[root@ceph04 ~]# umount /var/lib/ceph/osd/ceph-8
[root@ceph04 ~]# umount /var/lib/ceph/osd/ceph-9
[root@ceph04 ~]# umount /var/lib/ceph/osd/ceph-10

3、移除mds

a、关闭此节点mds服务

[root@ceph04 ~]# systemctl stop mds

b、删除mds认证

[root@ceph04 ~]# ceph auth del mds.ceph04
updated

来源：oschina

链接：https://my.oschina.net/u/3021599/blog/4463696