Ceph集群动态调整
1.新增/删除MONITORS
1.1新增MONITOR
命令：
ceph-deploy mon create {host-name [host-name]...}
示例：
ceph-deploy mon create osnode17:172.17.23.4
1.2删除MONITOR
命令：
ceph-deploy mon destroy {host-name [host-name]...}
示例：
ceph-deploy mon destroy osnode17
2.新增/删除OSDS
2.1列举磁盘
命令：
ceph-deploy disk list {node-name [node-name]...}
示例：
ceph-deploy disk list osnode19
2.2清理磁盘
命令：
ceph-deploy disk zap {osd-server-name}:{disk-name} #注意：该命令会删除相应磁盘数据
示例：
ceph-deploy disk zap osnode19: /dev/sdh osnode19: /dev/sdi
2.3准备OSDS
命令：
ceph-deploy osd prepare {node-name}:{data-disk}[:{journal-disk}]
示例：
ceph-deploy osd prepare osnode19:/dev/sdh osnode19: /dev/sdi
2.4激活OSDS
命令：
ceph-deploy osd activate {node-name}:{data-disk-partition}[:{journal-disk-partition}]
示例：
ceph-deploy osd activate osnode19:/dev/sdh1 osnode19:/dev/sdi1
2.5创建OSDS
命令：
ceph-deploy osd create {node-name}:{disk}[:{path/to/journal}] #注意：该命令为ceph-deploy osd prepare和ceph-deploy osd activate命令的集合
示例：
ceph-deploy osd create osnode19:/dev/sdj osnode19:/dev/sdk
2.6删除OSDS
命令：
#注意：ceph-deploy暂未提供删除OSDS的命令，因而只能手动执行
ceph osd out {osd-num} #将OSD移除ceph集群
ceph stop osd.{osd-num} #停止OSD服务
ceph osd crush remove {osd-num} #移除crush map中的信息
ceph auth del {osd-num} #移除OSD认证
ceph osd rm {osd-num} #移除OSD
umont /var/lib/ceph/osd/ceph-{osd-num} #移除OSD挂载
rm -rf /var/lib/ceph/osd/ceph-{osd-num} #删除OSD的数据目录
示例：
ceph osd out 18
service ceph stop osd.18
ceph osd crush remove osd.18
ceph auth del osd.18
ceph osd rm 18
umont /var/lib/ceph/osd/ceph-18
rm -rf /var/lib/ceph/osd/ceph-18
脚本：
vi osd-remove.sh
#!/bin/sh
for arg in "$@"
do
ceph osd out $arg
service ceph stop osd.$arg
ceph osd crush remove osd.$arg
ceph auth del osd.$arg
ceph osd rm $arg
umount /var/lib/ceph/osd/ceph-$arg
rm -rf /var/lib/ceph/osd/ceph-$arg
done
chmod +x osd-remove.sh
alias osd-remove=./osd-remove.sh
3.参考
http://ceph.com/docs/master/rados/deployment/ceph-deploy-mon/
http://ceph.com/docs/master/rados/deployment/ceph-deploy-osd/
http://docs.ceph.com/docs/master/rados/operations/add-or-rm-osds/#removing-osds-manual