> Also in our favour is that the users of the cluster we are currently intending for this have established a practice of storing large objects. That definitely is in your favor. > but it remains to be seen how 60x 22TB behaves in practice. Be sure you don't get SMR drives. > and it's hard for it to rebalance. ^ This. > What is OLC? QLC SSDs store 33% more data than TLC, 4 voltage levels per cell vs 3. > Fascinating to hear about destroy-redeploy being safer than a simple restart-recover! This was Luminous, that dynamic may be different now, esp. with Nautilus async recovery. > Agreed. I guess I wanted to add the data point that these kinds of clusters can and do make full sense in certain contexts, and push a little away from "friends don't let friends use HDDs" dogma. Understood. Some deployments aren't squeezed for DC space -- today. But since many HDD deployments are using LFF chassis, the form factor and interface limitations down the road still complicate expansion and SSD utilization. > For now, we limit individual cloud volumes to 300 IOPs, doubled for those who need it. I'm curious how many clients / volumes you have vs. number of HDD OSDs and if you're using replication or EC. If you have relatively few clients per HDD that would definitely improve the dynamic. _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |