So I got ceph running, got OSDs on each node (three nodes), but I'm totally lost on how to properly pool them together so that I can spin up a virtual machine and have high availability. Can anyone shine some light on how to go about just spinning up VMs and having them mirrored/checksummed across my nodes?
Totally lost with Ceph
- Thread starter 100percentjake
- Start date
Within the browser for one of your nodes, select a node on the left-hand list. Under the CEPH option, find the Pool button. Once you create a pool, you choose the Datacenter button. Under storage, select RBD to set up a shared storage connection.
I have the default pool, called 'rbd', on all of my nodes. I went to Datacenter and then created a new RBD called "ceph" with the IPs of the Monitors (my three cluster nodes each have a monitor daemon), the default 'admin' user name, and "Disk Image" and "Container", "enable", etc. but when I go to create a VM and choose the "Ceph" RDB disk as my destination it returns a Capacity of zero bytes.
Yes, I have been following that wiki page to set this cluster up. Keyring is copied as "ceph.keyring" in accordance with the name of my RBD as dictated in /etc/pve/storage.cfg. On any node I can go to Ceph->OSD in the GUI and see the OSD on each node and how big it is, but the RBD is showing 0 bytes still.
I figured it out. I had not capitalized "Ceph.keyring" as it is in /etc/pve/storage.cfg . I suppose coming from the lawless capitalization-ignorant land of Windows has made me sloppy. Thanks for your time everyone. I just spun up a Debian test server and it looks to be working brilliantly but I'm sure I'll be back on the forums with more questions eventually, haha.