Fix an OSD that has become associated with two block devices

Problem

A user replaced some OSDs in a cluster. After the replacement, one of the OSDs in the cluster had two block devices associated with it. The user wanted only one block device to be associated with the OSD.

Solution

Confirm that more than one block device is associated with the OSD under consideration by running the command cephadm ceph-volume lvm list xxx, where xxx is the number associated with the OSD:

$ cephadm ceph-volume lvm list 001 

Using recent ceph image registry.suse.com/ses/7.1/ceph/ceph@sha256:000aaaaaa00aa0aa00000aa00a0000000000000aa0a000aa0a00000a0a000a0a

ceph-00000aa0-000a-0a00-a0a0-aa0a0aaa0000/osd-block-a0a00a00-a000-0a0b-aa00-a00000a000a0

 [block]       /dev/ceph-AAAAAaa0-000a-0a00-a0a0-aa0a0aaa0000/osd-block-a0a00a00-a000-0a0a-aa00-a00000a000a0

 block device              /dev/ceph-AAAAAaa0-000a-0a00-a0a0-aa0a0aaa0000/osd-block-a0a00a00-a000-0a0a-aa00-a00000a000a0
 block uuid                xxxxxx-xxxx-xxx0-0xxx-xxx-xxxx-xxx0xx
 cephx lockbox secret    
 cluster fsid              xxxxxxxx-xxxx-xxxx-xx00-xxxxxxxxxxxx
 cluster name              ceph
 crush device class      
 encrypted                 0
 osd fsid                  xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
 osd id                    209
 osdspec affinity          drive_group_ssd-backed
 type                      block
 vdo                       0
 devices                   /dev/sdq

 [block]       /dev/ceph-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/osd-block-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx

 block device              /dev/ceph-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/osd-block-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
 block uuid                xxxxxx-xxxx-xxxx-xxxx-xxxx-xxxx-xxxxxx
 cephx lockbox secret    
 cluster fsid              xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
 cluster name              ceph
 crush device class      
 db device                 /dev/ceph-block-dbs-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/osd-db-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
 db uuid                   xxxxxx-xxmx-xxxx-xxxx-xxxx-xxxx-xxxxxx
 encrypted                 0
 osd fsid                  xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
 osd id                    209
 osdspec affinity          drive_group_ssd-backed
 type                      block
 vdo                       0
 devices                   /dev/sdag

Use the command dmsetup remove to remove the target logical volume:

dmsetup remove ceph-AAAAAaa0-000a-0a00-a0a0-aa0a0aaa0000/osd-block-a0a00a00-a000-0a0b-aa00-a00000a000a0

Use the command wipefs -a to remove /dev/sdq.
```
wipefs -a /dev/sdq
```
Warning: Using wipefs deletes the filesystem on the disk (in this example /dev/sdq). This command erases all filesystem, RAID, and partition-table signatures on the storage device (in this example /dev/sdq).
/dev/sdq is now empty and can be removed from the system.

Problem​

Solution​

Problem

Solution