I’ve parsed a lot of docs trying to figure out how to extend a one node Opennebula config (sunstone/worker on the same server) with another physical host.
So far adding the node let met see the CPU/MEM ressources correctly but I can’t find how to add the datastore
located on the new server.
The remote storage is LVM, if I add it (using LVM config) I see the Datastore but can’t get the storage capacity and so on.
Trying to create a VM on the new host just doesn’t work, same for testing migration.
Any clue/howto available ? As the docs only talk about separate front-end + several nodes.
when you add LVM datastore you have to define BRIDGE_LIST, which is list of server or one server address where LVM can be managed.
So if you have new node with address 192.168.2.2, then you should add this IP to BRIDGE_LIST.
(I use dns names, so in my config is “node1 node2 node3” and frontend has this names with particular IP addresses in /etc/hosts)
Also, you need NFS server on host where opennebula engine is running
You need to mount this nfs storage to new node to /var/lib/opennebula
Configure no password ssh access
Does this mean that from the AllinOne I share via NFS DS 0 an 1 and from the new added server I share my storage via NFS to the AllinOne server ?
And then I declare the storage as “shared” or just LVM ?
Where your LVM storage is located? On new compute node? If so then just add new LVM storage and don’t forget add BRIDGE_LIST param with address of that node.
In this case, “AllInOne” server connect via SSH to new compute node to manage LVM volumes and also to instantiate new VMs. New VMs also need system storage, so you need to mount NFS storage, which is exported from “AllInOne” server.