Hi, I want to provide in my cluster which is objectstore only momentarily rbd as well but not on the disks which I’m currently using for objectstore. I’d add more nvme/ssd drives for the existing servers, however I’d name the devices like nvrbd/ssrbd and put the rbd pools with the crush rules on those. Will ceph knows that these drives are ssd/nvme drives and knows that alloc size should be 4k …? Another option create a different root and leave it as nvme or ssd but this one I don’t want to do. Thanks _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |