Shredder ec2 leaves open socket in case startup was already run

### Actual behavior

Shredder does not close startup open socket in all startup cases

### Expected Behavior

Shredder should close socket in all startup cases

### Steps to reproduce

Run shredder once, get successful startup, run it again and see that nobody answers on that socket

### Checklist

- [x] Add a descriptive title
- [x] Include explicit steps on how to reproduce the issue
- [] Link to a repository that reproduces the issue (**optional**, but **greatly appreciated**)

### Tips

* If possible, create a small repo that reproduces the issue and provide a link
* Don't include full logging output -- link to a gist of the full logging output instead
* Don't paste screenshots of error messages -- copy and paste instead

Start by reproducing the issue by running Shredder twice and observing the startup socket behavior. Trace the startup path described in the issue to find where the socket is opened and where shutdown is handled. Done means every startup case closes the socket, including the second run, so nobody can connect to the leftover socket.

Written by the indexing model from the issue text.