Implement node refresh command

[leninmehedy]

Once a network is deployed, if we need to refresh (setup & start) a single node using solo node setup -i node1, it fails because it generates a new config.txt assuming there is only one node.

Also node start fails because it tries to setup mirror-node with one node.

We should implement solo node refresh command to refresh a single/multiple node with the below functionalities:

• It shouldn't generate config.txt

• It shouldn't generate new keys

• It should not restart the mirror node

• It should dump all saved state

• It should allow destroying the whole pod....

• It should only setup and start the specified nodes.

• required by Enhance node refresh command to update config.txt when required #186

[leninmehedy]

While solo node refresh is useful, but we need to allow adding a new node (e.g. support solo node add).
solo state should maintain a list of all members so that we can generate config.txt when provisioning a new node or refreshing a node.

solo node add flow:

1. Use helm chart to deploy a new pod
2. Setup the new node (with correct config.txt)
3. Send config.txt and public keys to all nodes
4. Send freeze transaction to all nodes across all clusters
5. Wait for 2/3 majority of nodes to FREEZE.
6. Restart/Recycle all node (systemctl restart network-node, for deterministic image it would be to kill the pod).

This functionality is captured in: #115