My ceph cluster has 3 3TB and 3 1TB drives with SSD WAL and DB. The write speeds are kinda meh on my VMs, from what I understand the 3 TB drives will get 3x the write request of the 1TB drives. Is my understanding correct ? and would it be better if I swapped my 3TB with 1TB drives making it 2 1TB drive per host instead of 1 3TB and 1 1TB.
Replace 3TB drive with 1 TB drives (Ceph)
- Thread starter maomaocake
- Start date
-
- Tags
- ceph cluster ceph scaling
You should try to distribute the drives equally among the hosts. The 3TB drives will still get 3x the IO requests than the 1TB drives. Both drive types will have the same IOPS and bandwidth, so you will always see a non-optimal performance.
