Refactor the flags for create_instance and run_systemtest

All of the current flags these scripts take are single letters. At this point, they bear very little resemblance to what they actually do and it is impossible to guess without first looking at the script. This makes using these scripts and understanding what the CI does very confusing. This is made even more confusing by the fact that the same flag can do different things in both of these scripts. Adding more flags is also problematic as you basically have to choose random letters at this point.

These flags should be cleaned up so that a long version is always present (and optionally a short version). So instead of -s to specify which test to do in run_systemtest.sh, we use something like --test-script <test script> (and optionally a shorter version like -s or -t.

This will make things much more readable and easier to extend in the future. As a reference, the existing build scripts already have functionality for handling these types of flags.

For now, I will see this issue as a blocker for some other issues related to these scripts as we just make the problem worse if we start introducing new features or changes to the scripts above.