One more newbie question )
I have a minimal working ceph cluster: 3 nodes, 6 OSDs.
I understand that I can scale up cluster adding one or more OSDs or
even nodes without stopping clients i/o and loosing consistency.
What if I decide to add a cache pool backed by SSDs or move a journal to
SSD? Should I stop clients i/o before performing operations of this kind?
Is there a rule of thumb to determine permitted operations from "online"
client's point of view?
ceph-users mailing list