Today I connected the ceph storage to the lxd storage pool.
Testing shows that using ceph storage does not seem to allow multiple lxd nodes to share the storage pool.
Since the same ceph pool name was specified when the storage pool was created, a message was prompted to configure the storage pool by using the reuse option.
When using ceph storage, can multiple lxd hosts use the same storage pool?
This is only possible in a cluster environment, standalone nodes cannot work together to avoid clashes so that’s why this won’t work outside of a LXD cluster.
Just set it up on the first node and any joining nodes will get it too.
It may ask you to specify the source on each joining nodes, just keep passing the same value and it will work.
We haven’t had much request for gluster and even ceph only has a tiny user base right now (less than a thousand users). If someone wanted to contribute support for it, we’d be fine with it though. I don’t know if it’s suitable to host a container/vm but even if it’s just suitable for storage volumes, a driver like the cephfs one could work.
Personally, I think lxd is more attractive than vmware if containers and VMs have online migration and shared storage pools are supported across multiple storage and filesystems. I think it will be a more attractive solution than the HCI solution.