How to correct uneven usage of all OSD disks?
Issue Description
Multiple OSD disks have uneven usage in Automation Suite 23.10.
Troubleshooting Steps
To check the ceph balance module which is enabled and active by running commands
- kubectl -n rook-ceph exec -it $(kubectl -n rook-ceph get pods | grep rook-ceph-tools | cut -d ' ' -f1) -- ceph mgr module ls
- kubectl -n rook-ceph exec -it $(kubectl -n rook-ceph get pods | grep rook-ceph-tools | cut -d ' ' -f1) -- ceph balancer status
Resolution
According to this document, manually adjust OSD weight to average usage, assigning the weight value to 1.0 for all 150GB disks, and 1.67 for all 250GB disks (1.67=250/150). Execute the following commands:
- ceph osd crush reweight osd.0 1.67
- ceph osd crush reweight osd.1 1.67
- ceph osd crush reweight osd.2 1.67
- ceph osd crush reweight osd.3 1.0
- ceph osd crush reweight osd.4 1.0
- ceph osd crush reweight osd.5 1.0
OSD will gradually show average usage after some time.