CEpH detaches OSD nodes

1. Disable the CEpH OSD process.

Service CEpH stop OSD

2. Balance the data in the CEpH Cluster

3. Delete the OSD node when all PG balances are active + clean.

CEpH cluster status before deletion

[[Email protected] ~] # CEpH OSD tree

# ID weight type name up/down reweight

-1 4 root default

-2 1 Host os-node3

0 1 osd.0 down 0

-3 1 Host os-node4

1 1 osd.1 up 1

-4 1 Host os-node5

2 1 osd.2 up 1

-5 1 Host os-node6

3 1 osd.3 up 1

1) Delete An OSD hard disk from the cluster

[[Email protected] ~] # CEpH osd rm 0

Removed osd.0

[[Email protected] ~] # CEpH OSD tree

# ID weight type name up/down reweight

-1 4 root default

-2 1 Host os-node3

0 1 osd.0 dne

-3 1 Host os-node4

1 1 osd.1 up 1

-4 1 Host os-node5

2 1 osd.2 up 1

-5 1 Host os-node6

3 1 osd.3 up 1

2) Delete An OSD hard disk crush map in the Cluster

[[Email protected] ~] # CEpH OSD crush RM osd.0

Removed item ID 0 name 'oss. 0' from crush Map

[[Email protected] ~] # CEpH OSD tree

# ID weight type name up/down reweight

-1 3 root default

-2 0 host os-node3

-3 1 Host os-node4

1 1 osd.1 up 1

-4 1 Host os-node5

2 1 osd.2 up 1

-5 1 Host os-node6

3 1 osd.3 up 1

3) Delete An OSD node from the cluster

[[Email protected] ~] # CEpH osdcrush RM os-node3

Removed item ID-2 Name 'OS-node3' from crush Map

[[Email protected] ~] #

[[Email protected] ~] # CEpH OSD tree

# ID weight type name up/down reweight

-1 3 root default

-3 1 Host os-node4

1 1 osd.1 up 1

-4 1 Host os-node5

2 1 osd.2 up 1

-5 1 Host os-node6

3 1 osd.3 up

This article is from the "zhanguo1110" blog, please be sure to keep this source http://zhanguo1110.blog.51cto.com/5750817/1535781