I'm having an issue where the existing Ceph cluster consists mostly of 10Tb 3.5" HDD's, and some smaller 1Tb SSD, in different pools.
This works perfectly. Storage pools have assigned a crush rule having set either ssd or hdd as storage class.
Now I wan't to assign a bunch of smaller 2.5" disks to the cluster.
The plan is to have those disks assigned to a crush rule to have one pool utilize only those disks (and keep existing pools from putting data on the smaller disks).
As I see it I can achieve this in two ways?
- Manually assign all OSD's to a specific crush rule. This would be a huge administrative burden.
- Assign a "custom" device class to the new OSD's. This would align pretty with how the existing OSD's are assigned to different crush rules, but I'm not sure if this is the correct way to do it, or if custom device classes is actually supported?
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
