Fix bug causing nodes to terminate prematurely

Sometimes, a node stops immediately after it is started.

Test failure example:

Error: root: failed to run scenario: root: failed to run scenario: oasis: keymanager-0 node terminated: env: sub-process exited early

The problem occurs when a service (e.g., common worker, registry worker, storage worker, etc.) exits prematurely. In such cases, the service manager shuts down all services, even if the exiting service completed its job gracefully.

To address this, we can:

• Add an Err function to BackgroundService and check for errors before the service manager stops other services.
• Modify services to never exit, even if they are disabled.
• Rename te BackgroundService interface to Service and redesign to either:
  • Serve(ctx context.Context) error, or
  • Serve() error along with a separate Shutdown() error.

The third option is the most effective but also the most complex. Changing the interface ensures that services can notify the caller if an error occurs during execution.