fix(manager): restart nodes with the same safenode port

[RolandSherwin]

Description

This PR tries to set the safenode port to the same value when we run upgrade.

When we initially start a service based network, the --port field of the safenode is used when we manually add one node at a time. But when using the deployer, we do not do that, but instead set the node_count. This would mean that port is automatically assigned by the safenode.

Now during the upgrade process, we would want to re-use the same port, else we'd be causing a massive churn. So here is the high level overview on how to achieve this:

1. Get the node port via the node's listen addr.
2. Supply this port to the service definition. I.e, add in the --port arg.

The 2nd point would mean that we will be editing the service definition file to add the args. But I could not find any functions inside the service manager crate that would allow me to do that.
Thus, I resorted to overwriting the entire service definition file with all the same args that we used during the initial startup. To acheive this, I had to edit the NodeRegistry to add in extra fields like bootstrap_peers, local etc., for us to provide the --peers and --local args in this overwritten service definition file.